We’re not SEO experts. Nor do we want to be. We’re user experience experts specializing in one aspect of UX: speed. Gaming SEO will not save us. Being kind to people is the sure win. Speed is web hospitality.
Countless people are making a business of search engine optimization (SEO).
Skyscraper strategy for backlinks
SEO services such as ahrefs, Semrush, and Moz.
Each year, companies spend more than $65 billion on search engine optimization (SEO). Yet, no one guarantees anything. Some guarantee the first page on the SERPs – but it’s a false promise.
SERP: Search Engine Results Pages refer to Google listings after searching.
Is any SEO service in any way useful or are they all useless?
REFERENCES from our PagePipe terms of service:
14. TESTING We make reasonable efforts to standardize the site. There is no way we can test or guarantee for every single past-and-future browser or device. We do speed tests and reports on Pingdom.com and WebPagetest.org. We do not use Google PageSpeed Insight scores for benchmarking.
16. RANKING DISCLAIMER By signing this agreement, you acknowledge that PagePipe.com neither owns nor governs the actions of any search engine. You also acknowledge that due to fluctuations in the relative competitiveness of some search terms, recurring changes in the search engine algorithms, and other competitive factors, it is impossible to guarantee number one rankings or consistent top ten rankings, or any other specific rankings for any particular search term.
That explains our position. Selling SEO is a ripoff. We sell user experience. And only one aspect of UX – speed.
So besides relevant content what makes a difference in internet traffic?
Titles are still important.
While titles are important for findability, they are critical for click-through. You have to hook people. They’re choosing between you and 10 other listings on a Google results page. They’ll read “the one” most specific to the decision they’re trying to make. And if it sounds easy to read. And interesting – not boring.
Findability is the ease of finding information. Both using search engines and by users already on the website.
So here’s one example of how to make titles more appealing:
Parent Category is: Shoes Sub Category: Nike Product Name: Air Max
Now plugin this phrase into Google search: Air Max Nike Shoes
We see popular questions pop up suggested by Google. They will be different for each user and geography. They will not be the same results everywhere in the world based on user search intent. They want to hook you into a search – and see advertising, so they make money (clicks).
This is our results:
What are Nike Air Max shoes for?
What is the best Nike Air Max shoe?
Do Air Maxes make you taller?
What are the newest Air Max shoes?
Four questions. All arouse human curiosity. Use something similar for the title on your blog page. Don’t write for machines. Write for humans. Tell them the conclusion in the title. It’s a headline. Study how to write eye-catching headlines for people.
For the permalink use something different: Air Max Nike Shoe review (or something similar).
The title is the implied (connotative) meaning. The permalink is the denotative or dictionary definition of what it is. This appeals to both sides of the human mind. Emotion and logic.
Both will appear in the listing.
Nothing says they should match. You have the opportunity to convey more meaning by not being redundant. This will improve your traffic. How do we know? It’s worked for our profit. It’s not provable – like all SEO.
Where is “above-the-fold” on a phone or desktop? Repairing render-blocking is wasteful ivory-tower foolishness. Not a real-world problem. But the anxiety sells lots of speed services like CDN and paid plugins.
Removing render-blocking elements like Font Awesome frequently breaks the hamburger menu and search magnifying glass and social icons, etc.
Are we saying that trying to eliminate render-blocking resources as identified by GTMetrix (and similar speed metric software) isn’t productive, useful, or a ‘real-world problem’?
That is correct. It is a waste of time and energy.
Anything that reduces resources should be considered for removal – but how good is good enough? Load times under 2 seconds are good enough.
PagePipe doesn’t chase render-blocking. It is sold by speed scammers often.
Render-blocking resources are only considered because there is a theoretical concept called: above the fold. Where is that exactly on a mobile phone? Abstraction!
Render-blocking is usually not important.
Who are these people creating wasteful test criteria? Steve Souders coined the term “web performance optimization” in 2004. He started the criteria for web speed for Yahoo Performance group and later Google PageSpeed insights. He no longer works for either company. He now works at SpeedCurve.com.
Who else are these speed experts making the test requirements?
No one ever knows. “Invisible people” get to dictate what is good and bad. Does it solve or cause real-world problems? We have seen no documentation that proves fixing render-blocking is a worthwhile real-world practice for small- to- medium-size businesses. It doesn’t change UX, SEO, or profitability. It’s like bragging, “My music system reproduces music flawlessly beyond the range of human hearing.” If you can’t hear a change – so what? Who cares?
Perfect is not practical.
PagePipe practices the Pareto principle and industrial value analysis. 80 percent of the speed effort only yields 20 percent improvement. We look for the 20 percent that will give 80 percent improvement. Render-blocking removal falls in the esoteric.
Font Awesome render-blocking should be avoided during website construction – but if it gets inadvertently built-in. Do not remove it. Doing so causes problems for most sites. We try to avoid themes that use Font Awesome – always.
A freelancer on Upwork offers speed optimization. He has excellent reviews. He also has multiple recommendations from Facebook groups. He has an impressive profile and offers a guarantee of 95+ on the Google PageSpeed insight test.
He claims the following techniques increase the speed of websites:
Time to First Byte
Image optimization and webP implementation
Render-blocking script and above the fold optimization
CDN and Cloudflare implementation
How can he guarantee 95+ speed optimization only US$30/hour?
Simple. He’s a scammer. We don’t recommend any of those methods.
The first clue of a scam: TTFB is hardwired into the server’s physical properties. It can’t be easily fixed. But few know that.
All these wonderful scam specs make no difference in SEO, UX, or profit.
Measuring with GT Metrix or Google PageSpeed Insight test scores are not an indicator of a speed professional. Those are tools used most by pretenders and wannabees. Sorry.
How is PagePipe different from WP Faster WordPress optimization services?
Glad you asked. WP Faster optimization services sound great. While the market stats they quote are true for big companies – they’re irrelevant for our audience. We work with small-to-medium size International web businesses.
People visiting PagePipe need a speed doctor. They know it already. We don’t have to do any convincing of speed benefits. WP Faster wants speed converts. Our audience already is made of speed evangelists. We suspect WP Faster produces similar performance results in the end … but you pay double to triple the cost. The sky’s the limit with their pricing.
So the price is a big difference. But our speed philosophy is different, too. We use the Pareto principle – 80/20 rule – as our guide. This is value analysis. Or cost-benefit analysis. We discover unrealized speed improvement opportunities.
Not everything is important or makes a significant difference in speed.
Anything you want that isn’t The WP-Faster way voids their refund policy. They, also, specifically say they “don’t teach how to do what they do.” We teach. Our goal is technical obsolescence. We want you to understand how to fix speed yourself and avoid future problems. WP Faster won’t even talk to you on the phone unless you pay a $6,000 deposit.
What WP Faster says confuses. They have abnormalities and contradictions in their FAQ section. For example:
WP Faster requires a mandatory installation of the $200 per month Cloudflare plan. They say not using Cloudflare voids their warranty. What warranty? And what host do they recommend? The speed-despised, $35-per-month SiteGround hosting!
PagePipe makes your site fast on your existing shared hosting without CDN. We don’t add costs using extra affiliate services for kickbacks. We reduce your annual overhead.
The flimsy WPFaster.com pseudo-guarantee:
“The sale is final, even if we were, through no fault of our own, unable to install, in full or in part, the performance architecture we have devised for your site.” – WP Faster
They repeat over and over: no refunds. With legal ferocity and verbosity.
Do better UX and speed improve profits?
It’s documented proof for BIG companies. But will it help you? That depends. Are you selling something perceived as good? If the answer is, “No, I sell rubbishuseless unwanted stuff.” Then don’t waste your time tweaking speed.
But, if your offer and content are good, speed helps. Faster load times make visitors happy. No waiting.
Speed Band-Aids and Quick Fixes
Many WordPress speed optimization services charge too much for applying band-aid fixes to slow sites. They promise improved test scores. But not guarantees of improved load time gains in milliseconds. Test scores are esoteric tweaks that make no significant difference. The test results recommend score improvement with technical services such as:
Leverage browser caching.*
Use a CDN (content distribution network).
Plugin prioritization and load order.
Remove above-the-fold render-blocking.
Lazy loading images.*
Improve server response time (TTFB).*
This list above is one of minor improvements. We do some of them* but only if it doesn’t break your site – or cost you an arm-and-a-leg in custom coding.
Installing WP Rocket caching plugin – and charging suckers $100 for that is not a service. WP Rocket is a plugin you can buy for $49 annually. But speed tricksters don’t pay that price.
They buy all-you-can-eat “unlimited installations” of WP Rocket plugin. That has a special $199 price tag. Meaning: it costs them almost nothing per-site installation. The client-gotcha is this: a scammer doesn’t pay for future annual renewals. This leaves the client holding the bag the next year.
These frauds list each automated feature of WP Rocket plugin as a separate ala carte service item. This deception is snake oil.
These overpriced practices don’t produce minimal load time improvement. They give real speed professionals a bad reputation. If you think marking up a mere plugin installation is a good business strategy, have mercy! Don’t do this to site owners. It’s a ripoff.
Competitor WordPress Speed Optimization Services
Eight common observations of speed-service competitors:
1Prices start low. Typically around $100. The high end is around $1,000. But many insist on bidding services and won’t give a fixed price. For some fees, the sky is the limit. Or they insist on expensive add-ons increasing the annual site break even. They usually get an affiliate reward for pushing clients toward these extra services. These are things like:
paid services like CDN
2All 35 homepages have a title saying, “WordPress speed optimization services.” But one said, “Turn your slow WordPress site into a supercharged powerhouse.” Better promise but still not quantifiable. Unprovable claim. “Optimization” sounds like a jargon buzzword. They don’t have a motivating benefits-oriented headline.
3 Tons of rocket logos. And the *second-place* cliche logo is a timer or gauge. With flames, of course.
4The slow irony. The page-design quality of these speed optimization service’s websites is fluffy and low credibility. Lots of popups and chats. Most sites advertising speed services – actually load very slow. We list the initial load time of each one below using a browser timer. Not “walking the talk” is evidence of inferior services.
5 They don’t teach speed strategy. They just do site tune-ups. Hands – but not shared brains. You don’t get consultation or education.
6Competitor services focus on Google PageSpeed Insight test scores. Improving scores doesn’t necessarily improve speed. We’ve seen pages that get all-green 100 scores that load in 12 seconds. Crazy slow!
The only performance metric that counts is improving speed in milliseconds. Not a cached speed. We’re talking about unprimed cache. That’s a first-time visitor’s experience. That’s what counts – first impressions of site quality. Some speed competitors focus on reducing the number of requests a page calls. They consider that a good-enough metric indicating speed improvement. Again, request reduction doesn’t guarantee improved load time in milliseconds. Vanity scores and requests are secondary compared to actual load time in milliseconds. The goal (performance budget) is loading a real page in under 2000 milliseconds.
7Bait-and-Switch Tactics They attract potential customers by advertised low-priced items. But then encourage buying a higher-priced speed option. Nasty!
8Bogus SEO promises. Speed service competitors in error claim speeding up a website improves SEO. There’s only a 0.5 percent benefit in page ranking algorithms from speed. Google speed mythology is not a valid reason to spend money on speed. The better reason is providing an excellent user experience. Speed is the primary and principal component of UX. Everyone hates a slow website. Speed is about the halo bias caused by a visitor’s first impressions.
There are at least 8 noteworthy WordPress speed service competitors:
Today, PagePipe uses 70 plugins. About 30 of those not updated for over 1 year. Some for many years. We’re not embarrassed about that. It’s not a mistake.
Plugins listed in our ebooks are currently used on PagePipe. And also on client sites.
So the question is “Outdated? By what definition?”
Some think outdated plugins produce a warning like:
“This plugin hasn’t been tested with the latest 3 major releases of WordPress. It may no longer be maintained or supported and may have compatibility issues when used with more recent versions of WordPress.”
Being orphaned or abandoned doesn’t mean “bad or rotten.”
These lonely plugins still work. And often for over a decade without complaints. That isn’t brokenness.
“Does 8 months since an update concern us? Not in the least. There are plugins that are 8-years old in the directory that work fine. Those “best if used by” freshness dates are silly. They throw people off with their arbitrary “expiration-date” warnings.”
WordPress places warnings when a plugin isn’t tested with recent versions. Does that mean it won’t work any more with new versions of WordPress? Nope.
WordPress’ motive is their legal protection against liability and lawsuits. C.Y.A. If a plugin doesn’t work any more or presents security hazards, it’s removed fast. And some are. In particular, malicious plugins. They call those “take downs.” Plugin authors remove some because they didn’t get the market results they wanted. But generally plugins stay as long as there isn’t any noise about them. Retired or dead author’s plugins stay in the WordPress free directory.
No plugin is safe. Not paid (premium) plugins. Not obsolete plugins. And not recently updated plugins. A common plugin problem is automatic updates loading onto managed WordPress sites. Bugs in the new version mangle the site or causes conflicts.
There’s no such thing as a risk-free plugin or theme. Even reckless WordPress messes up with their own Automattic-authored plugins.
Good-old “Plugin Logic” is our secret, speed-weapon plugin. It’s used on every site we touch. SELECT.ME issue #11 talks about it. It’s an amazing plugin.
Want to keep a specific plugin from updating? We recommend “Block Specific Plugin Updates” plugin. There are times this is handy.
There’s plugin churning in the 55,000+ plugin database. Don’t let silly warnings discourage you. They aren’t for your protection. They’re protecting WordPress.
Don’t fear old plugins.
How many plugins is too many?
PagePipe is hosted on cheap GoDaddy magnetic servers with no CDN. GoDaddy hosting is the second most hated provider in the world. The first is BlueHost. We’re out to prove even “bad” hosting can get fast page speed. (We host our store on Bluehost! Our blog on GoDaddy!) PagePipe.com is living proof these recommendations for speed actually work.
PagePipe now use 70 plugins on the blog (GoDaddy) and 24 plugins on the secure store (BlueHost). Even with this many plugins, load time is under 2 seconds on cheap, shared hosting. It’s not plugin quantity, it’s the quality that makes a difference. Web designers can’t be arbitrary in loading and activating plugins. The result is slow pages. And all our plugins are freebies from the plugin directory.
It’s a myth using many plugins slows down your website. Being sloppy in judging plugin quality or necessity is the culprit. That’s within a web designer’s control. It calls for wisdom and speed testing. The best plugins add no page weight at all – weightlessness! (In reality, about 1 millisecond – or less – per plugin to the initial page load.)
The WP Speed Guru promise: Pay $69 and learn how to make your site lightning fast. Baloney.
So we decided to learn how deep these hidden secrets are. We suckered and paid. Here’s what you get in this vacuous series of videos:
1Intro and Case Study This is 3-minutes demonstrating WP Speed Guru’s homepage speed. That includes a screencast of GT Metrix download history. WP Speed Guru claims the speed improvements came from dropping from 60 to 15 requests. The emphasis throughout the videos is on reducing HTTP requests. This doesn’t always result in speed improvement as WP Speed Guru demonstrates later.
And WP Speed Guru considers scores important. Scores aren’t significant to us in speed testing. Only milliseconds of load time is important. We know the number of requests isn’t indicative or proportional to speed improvement. There are various reasons for this such as Time To First Byte (TTFB) and HTTPS/SSL handshaking. These are server overhead measurements. The only way to overcome a bad TTFB is to move to a different server or host. What is bad? Any TTFB over 750 milliseconds is bad. So what is a good TTFB? Anything under 300 milliseconds.
Speed results reported in these videos is using GTmetrix online testing. To get TTFB numbers in milliseconds, test with ByteCheck.com or WebPagetest.org.
2Free Testing Tools This segment is 14 minutes long. GTmetrix online speed test show speed changes on a demonstration restaurant site. WP Speed Guru recommends a paid theme vendor. The site under test has wasteful features we’d never use on a speed site. These include a homepage slider and over-the-top animation. Other things are a YouTube video, accordion navigation, Google Maps, and a Blog section. WP Speed Guru homepage speed test is 2.6 seconds. 1.13-megabyte page weight, and 47 requests.
WP Speed Guru recommends 6 tests in a row. This is something we often recommend, too. WP Speed Guru talks about uncontrollable things like TTFB and Google Font load times. These fluctuate outside web developers or site owners’ control. That is often true. So no disagreement there. But have we gotten our $69 worth of valuable speed secrets? Not even close. This is all common knowledge shared on PagePipe and other blogs for free.
WP Speed Guru then demonstrates a speed test on Pingdom.com with the same test page. A two-year-old test is evident since Pingdom changed its interface long ago. WP Speed Guru also cherry-picks a location close to the server. This, of course, will improve the results. WP Speed Guru says grades determine if things will fall into place for speed. Everything WP Speed Guru suggests is score-based troubleshooting. WP Speed Guru disclaims scoring practices by saying, “This may not apply to your site. Some score suggestions need to be fixed and some won’t.” WP Speed Guru doesn’t tell us which suggestions matter.
WP Speed Guru then shows a Pingdom waterfall for the test page. WP Speed Guru checks for red 404 errors indicating broken links. This is so elementary and basic common sense. Broken links are bad for many reasons – speed delays being one of them. WP Speed Guru looks at requests from server or external servers in the waterfall. But again doesn’t make any suggestions on what to do about these except “Repair them.”
WP Speed Guru describes how a CDN works. A basic explanation is free on Wikipedia.
WP Speed Guru then runs WebPagetest.org. WP Speed Guru shows test server locations and browser selections. And fiddles around with selectors. After showing WebPagetest.org options, WP Speed Guru says, “Pingdom is most accurate. But the best test is checking with your browser.” Implying but not saying, “use a browser timer extension or addon.”
WP Speed Guru talks about deviation in results depending upon the test chosen. But never quantifies what to expect. At PagePipe, we use Pingdom for best-case load-time scenarios. And WebPagetest.org for worst-case scenarios. The test results in milliseconds rarely match. Pingdom isn’t a better test. Only a faster one. It’s less comprehensive.
Repeatable testing is always on WebPagetest.org. But you’re comparing relative change – and not absolute change after making improvements.
Is any of this secret? No. This is common knowledge.
3How to Set Expires Headers The next section and 4 minutes is about How to Set Expires Headers The goal stated is to remove the red “F” failure flag on WebPagetest.org caching and turn it green. WP Speed Guru’s method is not recommended by us. WP Speed Guru shows this method to make things seem complex or frightening. The solution is simple.
WP Speed Guru demonstrates rewriting HTaccess files via FTP or Cpanel. What? We use a simple no-coding plugin. WP Speed Guru examines server code changes and edits via Apache Cpanel. And edits the htaccess file for expires header with copy and paste. WP Speed Guru downloads to edit and then reuploads. Very clunky. But the big surprise is WP Speed Guru doesn’t tell the “magic code” that is added. This is the hardest way to fix this minor speed problem. A plugin solution is easier and requires few settings and causes no speed drag. We sense this step is hand-waving. Is it secret? No.
WP Speed Guru retests on GTmetrix. 2.4s, 1.13M, 47 requests. Scores change from red to green – but speed improves by only 200 milliseconds. This amount is not uncommon for TTFB fluctuation on a host server like SiteGround or WP Engine. This secret is insignificant. Does PagePipe add a plugin anyway? Yes. But only as a matter of principle. It doesn’t improve UX or SEO in the slightest. It’s a vanity speed metric.
5Free CDN with Cloudflare This section is 21 minutes and is about how to set up Cloudflare CDN. Boo! We never use this third-party service. Ever. If it’s on a site, we remove it first. It’s a band-aid to mask real speed problems. WP Speed Guru even demonstrates later how CDN is unnecessary if you use caching plugins. Cloudflare is a complicated setup for a free account. This section is full of BS about the theoretical benefits of CDNs.
This section is a waste of 21 minutes on which you’ll never get a return.
You can duplicate minification and concatenation on Cloudflare with free no-setting plugins. WP Speed Guru tells us CDN can break pagebuilder plugins. And needs special settings for selective deactivation of minification.
If you want Cloudflare CDN, all this information is available on their website. It’s not secret.
But PagePipe doesn’t recommend free CDNs. They produce 503 server-not-found errors. A broken page or site results. Not worth it.
6Great Free Caching Plugins This section is 10 minutes of detrimental garbage.
He installs Wordfence Security since it has caching. This slow multi-function plugin causes delays of 250 milliseconds globally to pages. Slower speed overhead. It can cause unnecessary server resource overages. A bad plugin choice.
With caching, the initial page load remains unchanged. But repeat visits with primed cache yields homepage results of 0.9s, 1.12M, 47 requests. What percentage of your traffic are first-time visitors? We bet it’s the majority. They don’t benefit during visits when the first impression is most critical.
Our note: Minification often breaks sites. So don’t use it unless needed – and usually, it’s not needed.
WP Speed Guru then installs WP Fastest Cache. 1.4s, 1.18M. 39 requests. Even though this is slower. The author thinks it’s a speed win by reducing the number of requests by 8. Ridiculous conclusion.
7How To Optimize Images A 6-minutes tutorial of how to optimize images manually in Photoshop. WP Speed Guru recommends not using extra plugins for image optimization. While we do hand optimization, it’s bad advice to not use image compression plugins.
Is any of this knowledge secret? No. Free tutorials are all over the web.
How to do image optimization is available elsewhere on the Internet. WP Speed Guru shows how compression didn’t hurt the image quality. This is common knowledge. Not secret.
8Embedded Video For 8 minutes, WP Speed Guru demos an “About” page with an embedded YouTube video. Before test on GTmetrix: 2.9s, 945k, 31 requests WP Speed Guru recommends using a lightbox plugin. Video then opens in the lightbox and plays. WP Speed Guru used an unidentified paid X-theme plugin.
New results: 1.3s, 507k, 24 requests Achieve these same results with a free lazy load for video plugin. Secret? Nope.
9Google Maps 4-minute explanation of a map on a contact page. This shows how a developer used a still image of a static map. The still-image link opens a new window with the Google Maps page. Again, this clever offloading trick isn’t a secret.
WooCommerce WP Speed Guru examines requests added by WooCommerce – even without features. This is global loading or site drag. WP Speed Guru uses FTP access to edit the child theme’s functions.php file. WP Speed Guru copies and pastes some code into the file. But doesn’t give us the code. Crazy. Get our free PDF instead to tune up WooCommerce.
WP Speed Guru saves 7 requests but the WooCommerce page speed is slower. Lame demonstration.
Big deal. WP Speed Guru showed caching reduces requests. But still doesn’t get the page under the 2-second performance goal.
These load-time failures are never addressed.
Was this a secret? Not ever.
WP Speed Guru then installs instead Autoptimize plugin and glosses over settings.
Results are then: 2.9s, 1.04M, 23 requests. WP Speed Guru claims speed victory because the requests are fewer. But speed still isn’t under 2 seconds. Absurdity.
The Final Speed Test A final 3-minutes demo of Pingdom using different server locations. 1s, 1.32M, 29 requests with unprimed cache. 809m, 1.2M, 29 requests after caching. This demonstrates that speed is non-geographic influenced. It also shows how Pingdom always gives best-case results for cherry-picking speed. Really? Now he uses Pingdom to measure?
Inconclusive. These recommendations aren’t secret – all are available for free.
WP Speed Guru subtracts your $69 tutorial fee if you buy tuneup services. Services for the confused.
EXTRAS Useful Performance Plugins – links given. They are only a few and of minor consequence.
Knowledge Base – a link to WP Speed Guru’s blog.
WP Speed Guru evaluates by the number of requests. This is wrong. It’s load time in milliseconds that matters. Save your $69.
The point of diminishing speed returns is when you damage profitability. Then you’re into the realm of fanaticism. We’ve been there for a long time. That’s an unquenchable desire for excellence. It can be a monster. So let’s keep speed changes in perspective. In other words, do as we say – and not as we do.
For us, speed gets designed in. Speed repair is a frustrating, after-the-fact rescue. Sometimes, the patient loses an arm or another appendage. Cutting out cancer. Most site owners don’t want to suffer the painful amputation – and still, want to live forever. Nope. That isn’t how it works.
Anyway, what is reasonable for speed. For example:
Moving video files to another host like AWS. Will that save the day? That is unknown. Using an image placeholder and not loading the video until clicked is enough. That means using “video lazy load” plugins. That trick won’t work for header videos like some use.
Also, lazyload plugins don’t work with Elementor. Bad voodoo on Elementors part. It forces you to use their built-in video lazy load functions. But they won’t work for headers either. Trapped.
Let’s review some web speed realities:
1As far as we’re concerned, delayed loads produce the same user experience as offloading. Here we are talking about “perceived” load time versus “actual” load time. We build for human UX and not machine SEO. The fact your header video loads last is good enough for Google. Is it good enough for your human audience? Most likely. Has any user complained? Is the cost of offloading video justifiable? That is a business call. ROI. How much is video slowness affecting profits? A-B tests? If it’s so bad, why use video instead of still images?
2Font Awesome is a curse and laziness for developers who don’t want to produce image files. Those icons are faddish and will look dated. Clipart for websites. So we ask, “What enqueues (turns on) Font Awesome?” Themes like Tiny Hestia, GeneratePress, and Twenty-nineteen don’t load Font Awesome. Why? Because speed is an issue. Older and crummy designed themes use FA. If it’s not the theme, is it a plugin? Plugins can be selectively deactivated or stripped.
So if we install Asset Queue Manager (AQM), we then see the hidden scripts and styles used on a page. We can also site-wide dequeue them. This can be delicate surgery. But we use it all the time.
Below is a screengrab of the AQM plugin control panel. Circled in red are those things associated with Font Awesome.
The kicker is:
Does deactivating Font Awesome break critical icons? Those are the mobile navigational menu (hamburger) and the search looking-glass icon.
Only testing reveals that. But looking at Elementor code, it says it’s loading the lightweight Elementor Icon (eicon) set instead. (When disabling Gutenberg block editor, deactivate the block-library style circled in green.)
The truth is you can dequeue a ton of Elementor styles and scripts. But it requires trial and error to figure that out. Tedious. Now if you like, you can dequeue with the paid PerfMatters plugin. But we don’t use that because we only investigate free plugin alternatives. Sorry. Snobbery. We know.
Would we change the theme of a site and swap all the FA icons? Yes. We recently replaced 40 icons sprinkled throughout a client’s site. With what? PNG files. Why? Speed. Pure speed. One-color PNG icons only weigh 1k. Font Awesome loads 70 to 80k on every stinkin’ page.
There are even more serious ways to remove FA by editing the functions.php core file. We won’t go there anymore. We’ve nuked sites doing that. But if you pay someone enough, they’ll do almost anything. Not us. We can be taught.
3Hosting Google Analytics locally may gain you anywhere from 100 to 300 milliseconds. Nice but fanatic. Do we do it? Of course.
We turn conventional speed wisdom on its head. Sixty-four active plugins – on our site – load in under a second. Amazing. What’s our secret sauce?
When we do a speed audit – whether for PagePipe or for a client – we dial back PHP 7.x if we can to version PHP 5.6. Then we can run P3 plugin. The goal is getting some feedback of where 80 percent of plugin speed goes – and reduce that load. It’s Pareto’s principle in action: the 80/20 rule. This doesn’t work often. Sadness. Goodbye P3 plugin. We use other techniques now.
We use selective activation or deactivation of plugins on URLs. That’s the real-top-secret sauce. Most people don’t know plugins can add weight globally to websites. Not just to the page where they’re used. For example, Contact Form 7 adds about 43 milliseconds to every page and post on a site. It’s global loading and we call it “site drag.” It’s best to activate that plugin only on the contact page.
Next, we remove plugins whose features don’t add much value. You’ll see in this audit we removed 8 test plugins for a gain of 194 milliseconds. “Good work, Steve!” [breaks arm patting himself on back].
But it’s theoretical. P3 isn’t that accurate. You have to check real gains in milliseconds with Pingdom.com or WebPagetest.org.
We labeled all deactivated plugins used during maintenance as “offline” in the table below. Deactivated, mothballed plugins don’t slow down a website. That’s a common myth. No harm.
So the total estimated plugin load time is 435 milliseconds (±30 percent). But that’s never loaded at the same time on all pages and post. Because selective activation saves the day again! More tweaking is always possible – but go for the low hanging fruit first. We even use selective activation to rid us of nasty plugin conflicts.
A selective activation example.
Easy Digital Downloads is a heavy 158-millisecond ecommerce plugin. It’s not activated on our high traffic pages. This accomplishes our main goals for now. We’ll tweak more later on that one. We repeat: selective plugin activation is the secret. It keeps the Home page fast on shared magnetic hosting with no CDN.
The plugins causing the most trouble to configure are Watu Quiz, Easy Digital Downloads, and UpDraft Plus. They required serious thinking and a learning curve. What made the effort worth the trouble? These are complex functions and important site features. It’d be a nightmare and tedious to achieve any other way. We prefer plug-and-play plugins with no settings. But for these three that’s impossible. They need customizing and making many settings. Once you’re done, it feels great. Like when you stop pounding your head against a brick wall.
“Let me explain… No, there is too much. Let me sum up.” – Inigo Montoya in The Princess Bride.
Our original intent was writing about each of our 58+ plugins. That’s too much work. So we decided to summarize PagePipe’s speed strategy.
OUR SPEED STRATEGY
PagePipe uses the default Twenty-seventeen theme. We’ve studied it, tested it, experimented, torture-tested, and written a lot about it. It’s fast.
Deciding what not to do on your site is just as important for speed as what you do.
1. Add Search to Menu – necessary for this theme. No sidebars on pages. [We no longer use this plugin. That’s here but we do on other sites.]
2. Add Widget After Content and Simple Content Adder (for top of content). These two plugins allow two random rotators for images and testimonials on all posts. They are selectively deactivated on pages.
6. Every resource we use is free! WordPress, the theme, and the plugins.
PLUGIN AUDIT TABLE
Our PagePipe plugins below are ranked slowest to fastest. We then examine the 80/20 ratio and focus the cumulative cutoff at 80 percent. Offenders are shown in “red.” Their value must be justified. Notice we deleted some heavy and redundant ones.
plugin name,ms,percent,cumulative,activation/deactivation ,,,, Easy Digital Downloads ,158.2,36.35%,36.35%,Home & 2 conflicts WP Experiments Free ,,0.00%,36.35%,Deleted 109.9ms Worth The Read ,,0.00%,36.35%,Deleted 44.8ms Simple Content Adder ,40.3,9.26%,45.61%,Home & 2 pages WordPress Popular Posts ,31.7,7.28%,52.90%,Home & pages – posts only Broken Link Checker ,,0.00%,52.90%,Offline 28.3 WordPress 23 Related Posts Plugin ,25.4,5.84%,58.73%,Home & 4 pages WP Image Refresh ,23.4,5.38%,64.11%,Home & 3 pages WP Counter ,,0.00%,64.11%,Deleted 19.8ms Easy Forms for MailChimp ,18.8,4.32%,68.43%,one page UpdraftPlus Backup/Restore,17.3,3.98%,72.40%,run weekly WP HTaccess Editor,,0.00%,72.40%,Offline 16.6ms PDF Image Generator ,,0.00%,72.40%,Deleted 15.3ms Blog Manager Light ,9.7,2.23%,74.63%,blog listing directories Redirection ,9.4,2.16%,76.79%, Post Notif,9.4,2.16%,78.95%, WP Super Simple Speed ,8.9,2.05%,81.00%,ecommerce pages Cache Enabler ,8.6,1.98%,82.97%,ecommerce pages Limit Login Attempts Reloaded ,7.7,1.77%,84.74%, Watu Quiz ,6.1,1.40%,86.14%,mailchimp conflict 1pg WP Thor Heartbeat,6.0,1.38%,87.52%,mailchimp conflict 1pg Master Slider,4.9,1.13%,88.65%,on catalog page Block Bad Queries (BBQ) ,4.7,1.08%,89.73%, Lazy Load by WP Rocket ,4.7,1.08%,90.81%,video lazy load conflict Quotes ,4.0,0.92%,91.73%,two pages Optimize Database Delete Revisions ,,0.00%,91.73%,Offline 4.0ms Simple Drop Cap ,2.9,0.67%,92.39%, WP Editor Widget ,2.8,0.64%,83.62%, Title Remover ,2.7,0.62%,84.24%, Admin Post Navigation ,2.4,0.55%,84.79%, Add Search To Menu ,2.2,0.51%,85.29%, Disable Comments ,2.2,0.51%,85.80%, Asset Queue Manager ,,0.00%,85.80%,offline 2ms Imsanity ,,0.00%,85.80%,offline 2ms WEN Responsive Columns,,0.00%,85.80%,delete redundant 1.9ms Add Widget After Content ,1.8,0.41%,86.21%, Beacon Plugin ,,0.00%,86.21%,Offline 1.5ms Disable Emojis ,1.4,0.32%,86.53%, HW Image Widget ,1.3,0.30%,86.83%, Plugin Logic Rules ,1.2,0.28%,87.11%, Easy Table ,1.2,0.28%,87.39%,pages not posts Perfect Pullquotes ,1.1,0.25%,87.64%,pages not posts Pro Related Post Widget ,1.1,0.25%,87.89%,Home & 2 pages Shortcode For Current Date ,1.1,0.25%,88.14%,Home & 2 pages WP jQuery Plus ,,0.00%,88.14%,Deleted redundant 1.1ms Change Table Prefix ,1.0,0.23%,88.37%, Email Address Encoder ,1.0,0.23%,88.60%, Host Analytics js Local ,1.0,0.23%,88.83%, Deactivate XML-RPC Service,0.9,0.21%,89.04%, Lazy Load XT ,0.9,0.21%,89.25%,Video & conflict pages Tuxedo Big File Uploads,0.8,0.18%,89.43%, Query Strings Remover ,,0.00%,89.43%,Deleted redundant 0.8ms Remove Google Fonts References ,0.8,0.18%,89.61%, Far Future Expiry Header ,,0.00%,89.61%,Deleted redundant 0.6ms Hide Featured Image ,0.6,0.14%,89.75%, Simple WP Sitemap ,0.5,0.11%,89.87%, WP Author Date and Meta Remover ,0.5,0.11%,89.98%, Current Year & Copyright Shortcodes ,0.4,0.09%,90.07%, Plugin Logic ,0.4,0.09%,90.17%, Restore Image Title ,0.4,0.09%,90.26%,Home & 1 page Enable Media Replace,0.3,0.07%,90.33%, One-Click Child Theme ,0.3,0.07%,90.40%,on two pages (2017 theme) WENS Responsive Column Shortcodes ,0.3,0.07%,90.46%,where used Restore Image Title,0.2,0.05%,90.51%, Server-Mail-Check,0.2,0.05%,90.56%, Simple back to top,0.1,0.02%,90.58%,two conflict pages CAOS Analytics,0.0,0.00%,90.58%, ,,,, TOTAL milliseconds,435.2,,, active plugin count,53,,,
Popular premium WordPress themes have a big commonality: They claim they do anything.In fact, the Divi theme’s motto is “The sky is the limit.”
You’ll pay $89 annual rental fees to use Divi Theme
Divi, the most popular paid theme is the slowest we’ve encountered. It loads in about 1 second. That’s only the theme? Yep. That doesn’t include WordPress core or images or plugins or anything else. With a 2-second performance budget, our speed is half-way evaporated.
Speed tests are performed with Pingdom. Hardly scientific. But results give relative comparisons of performance value. The candidate themes are hosted on the same cheap, shared GoDaddy server without CDN or caching.
The potential to “Doing anything and everything” is seductive to site owners. Imagine. No limitations. The promise of no limitations is saying, “Go ahead. Add all the heavy extras your heart desires. Fill all the slots and widgets our expensive theme provides. Gorge yourself.” Is it any wonder websites with these premium themes end up being the slowest?
It’s not the poor theme’s fault.
It’s a design self-disciple problem.
Creativity requires limitations. It’s when resources are scarce – like time, energy, or money – that creativity is most desirable. Scarcity forces value analysis on extra additions. It reduces the temptation to bloat.
And other top selling WordPress themes are full of extras.
They all have a hefty purchase license. That’s annual recurring $60 to $70 rental fees.
Every year. Cha-ching. You pay again. That premium theme adds to site overhead.
A paid/premium theme must be the best. Right? Surely, they’re better than free themes. Better for what? Not speed. That’s for sure.
Multi-function, multipurpose, feature-rich, do-everything themes do not guarantee fast page loads. These are the themes most abused by site owners. Stripped themes force prudence before adding features. All-you-can-eat themes don’t. They encourage gluttony.
Feature-itis or creeping featurism is the ongoing expansion of website features. These extra features go beyond basic needs. They result in bloat and over-complication – rather than simple design. This is also called gold plating or over-engineering. Trying to make something better than required is also called: wasteful.
A better method is building on a simple foundation using a stripped or bare-bones theme.
How do you add features and functionality to a stripped theme? Simple. With plugins. But plugins slow things down, don’t they? No. It’s the quality of plugins, not quantity that makes a difference.
How much do stripped themes cost? Well, they’re free. Is free bad? You know: shoddy and fragile? Nope.
Q. Automattic authored how many themes? A.117 themes in the WordPress directory.
Q.How many themes contain updated code? A.Freshness? 34 themes. That’s it. Only 29 percent are garden-fresh.
We analyzed every one of the 117 themes written by Automatic. Inc. That’s WordPress’ parent company. 34 themes were recently updated. The rest are stale as dried toast. That’s right. Only 29 percent of Automattic-authored themes are fresh. Those good-fresh-ones are on 2.8 million websites.
Last Year of Update, Automattic, Themes 2018,34,29.06% 2017,27,23.08% 2016,11,9.40% 2015,30,25.64% 2014,3,2.56% 2013,11,9.40% 2011,1,0.85% ,, Total,117,
So what? Who cares?
Outdated and obsolete: that’s 71 percent of Automattic downloadable themes. Still hanging around in the WordPress Theme Directory. Without an update for over 1 year. Their shelf-life expired – molding – but still available for the unsuspecting. Why would WordPress keep old stuff laying around? Isn’t that unsanitary?
Well, it makes their big theme collection look BIGGER, doesn’t it? But finding a current theme that’s NOT stale is sheer gambling. And perhaps prohibited in some states and countries.
Let us save you some time and trouble. We’ll share some theme observations.
We deleted a few theme candidates failing to qualify for our test criteria. Those are:
Boutique and Deli are both child themes for Storefront theme. So we cross them off the list. Also, Twenty-eleven theme and Twenty-ten theme aren’t responsive. They never were. That’s no good for Google mobile-first page ranking. So they’re history, too, even though they are still available.
That culling left 30 responsive fresh themes.
Now, let’s reduce the remaining theme contenders another notch. We only consider theme’s with zip download files under 1 megabyte. That size has the potential for speed. 1MB is our cutoff. We don’t have time to mess around looking under the hood of every single theme. The theme author didn’t care enough to lighten the package! We won’t reward bad apathetic behavior with an active installation. No vote from us.
Most themes by other authors and companies in the Directory are not responsive. We tested that stuff. There are over 1,600 responsive themes out of 5,100 free themes. Anything authored before 2012 is most suspect. So watch out! Test to be sure.
That eliminated 10 more fluffy themes:
Are the 10 themes above our 1M-cutoff the slowest? Well, they’re on the high end. Storefront is the slowest Automattic-authored theme at 26.50 milliseconds load time. We admit. That’s pretty light. But it was the worst. It’s often used with the 280-millisecond WooCommerce plugin. Very heavy site drag. Much of our 2-second performance budget – Gone. Ouch!
We axed it. Cruelty.
What did Twenty-fourteen default theme ever do to offend PagePipe? Besides being over the 1MB weight limit? It activates 120k of Google fonts and enqueues Genericons. That’s all it took. Do we care that 300,000 sites are actively using it? Nope. We never even liked the look of that lame theme. Sorry.
So of the remaining candidates, who’s the fastest loading theme? Shocking. Twenty-twelve default loads in only 14.7 milliseconds. Dang. That’s fast.
Susty theme is the fastest loading theme on the planet – it also loads in 14.7 milliseconds. We use it for experiments.
So what’s the slowest theme we’ve ever encountered? Without question: Divi by Elegant Themes. Go figure. About 1-second of site drag – just the Divi theme. It’s a dog. Isn’t it popular? You bet. Don’t follow the herd.
Do we like the speedy Twenty-twelve default theme? We’ve seen worse. Customize it. Fix it. At least, it has an optional and popular one-column design. No sidebar if you desire. But in the end, it gets the ax, too. Read on:
Theme speed doesn’t translate into the fastest load time when combined with core.
The fastest theme under those conditions is Pictorico with 632 milliseconds. Twenty-twelve and AltoFocus are over 1 second. We’d ax those two from our list.
Is this a fair test? No. Of course not. Some themes come loaded with an image. The Twenty-seventeen theme is one of those. It has a large sample image header. Yet, it’s number 5. Twenty-twelve theme which was the fastest loading theme when installed is now a loser.
It’s a quick selection process when you have no time – and tons of prospects to test. The goal is building a fast site. There’s no perfect speed theme. Never will be.
Our point? If a theme has a zip package weight under 1 megabyte, it increases your chances of being fast loading. It’s a filter reducing theme-selection waste. The rule of thumb is free WordPress-authored themes load in under 20 milliseconds.
But they’re often snubbed by techno-geeks.
20 milliseconds – that’s fast. But it doesn’t translate into real speed once running on a real website with WordPress core. Why? Too many nuances to list here but they include things like Google Fonts and jQuery. Some even have sliders. Gads!
We strip all that junk with removal plugins.
Our biggest point: Free themes get a bad rap for speed. We see so many blogs telling how important it is choosing a fast premium theme. It’s implied only paid themes achieve long product life cycles. That produces a lot of unnecessary anxiety.
Can you say: Affiliate?
The main thing is avoiding extreme slow 1-second themes. Divi theme is almost 2 seconds with core added. Avoid using a paid theme. Never trust a vendor’s demo. They’ll be slower in the real world because they’re “feature rich”- aka heavy. Choose a stripped-down free theme. Then add as many plugins as needed for the functions and features you need. It’ll be faster loading than a prepackaged, off-the-shelf, paid theme. Avoid the most popular plugins. They’re the slowest.
That’s our speed formula – and it works.
More Theme Strategy
The twenty-ten default theme is now retired from WordPress.com. Themes retire on that host when they are no longer compatible with WordPress.com. Twenty Ten is still available for sites hosted elsewhere using WordPress core. Twenty-ten isn’t responsive for mobile devices.
Gutenberg won’t force change on any site theme. (Promises. Promises.) There are still sites on WordPress.com using classic original themes. WordPress no longer supports retired themes, but they don’t force anyone to change it.
Why stick with a trusty default theme?
Support and upgrades keep arriving for a long, long time. Twenty-thirteen got a refresh in 2018. That’s 5 years of free upgrades. But so did Twenty-ten theme. That’s seven or eight years!
That means Twenty-seventeen default theme will be thriving from 2022 to 2025. Think about it. Future-proof your site. Twenty-seventeen is still our go-to theme for building speed sites.
NOTE: The typical site shelf-life is 3 years. Then the owner or audience gets bored and the theme changes.
We admit we use themes like GeneratePress.Tom Usborne is the theme author and has a two-man company. If Tom gets run over by a bus, it won’t go well for the future of GeneratePress theme. But will a bus run over WordPress? They’d have to throw a hundred web developers under the same bus. Big bus.
WordPress claims a valuation of over 1 billion dollars. It’s not going away soon. Of course, Google could always buy WordPress with pocket change – and ruin every single theme. Stranger things happen.
Meanwhile, until Google Armageddon, default themes are the winning ponies to bet on. Customize with free plugins.
Is my site slow because of my theme selection? It’s the theme you should NOT pick that’s important. Don’t choose a paid, multi-functional theme especially Divi or The7. Chose a free, current, Automattic theme.
Be the fastest.
What we’ve described here is site origin optimization strategy.
1Start with a free theme authored by Automattic preferably a default theme with a download zip file size under 1 megabyte.
Why an Automattic theme?
It’s free. No annual renewal fees.
It has longevity of at least 5 years of free updates – maybe longer. A team of hundreds and a billion-dollar company support it. Not a few. Hundreds.
It’s faster than feature-rich paid themes. Always under 50 milliseconds load time.
2Add plugins that aren’t multipurpose. Avoid popular plugins. Use instead discrete plugins. Best case, discrete plugins only have one function and no settings. It takes several plugins to match the features of a multifunction integrated plugin. But many discrete plugins still load much faster than a single heavy plugin. How much faster? At least, 10 times faster.
Why use discrete plugins?
They’re not complex and simple to set up.
They’re faster than paid or popular plugins.
They’re prone to very same problems as paid plugins.
One feature can be selectively deactivated. Individual selection is rarely possible with a multifunction plugin.
3Finally, use selective deactivation and activation of features only where they are needed. The easiest plugin to use for selective activation is Plugin Logic. Remember most plugins cause site drag. That’s global loading on all pages and post whether the functions are used or not.
It takes special handling to keep GeneratePress fast. A simple plugin addition can cause Google font loading or jQuery enqueue and other blunders. Keep an eye on these things so the speed bonuses aren’t wasted.
And the test results are incredible:
We’ve looked at GeneratePress before. When did GeneratePress become so magical and wonderful for mobile speed? Officially, December 14, 2017 version 2.0.1 was released. That’s when all these speed features appeared.
★★★★★ GeneratePresstheme Active installs: 200,000+ Compressed file size: 882k.
Most commercial and free themes now include Font Awesome. It is a workaround to create social icons, responsive-menu-hamburger icons, and search-field icons. It causes site drag. Dequeuing Font Awesome for speed unfortunately ruins mobile screens.
Authors of free themes include Font Awesome 40 percent of the time. On paid themes, it’s almost always present. The selected GeneratePress v2.1 theme doesn’t enqueue Font Awesome. The “Lightweight Social Icons” plugin (also authored by GeneratePress) loads a faster social-icon subset produced using Fontello. But what we’ve found to be even faster is using PNG icons as social-media links. Those are about 300 bytes each. Then no plugin or fewer HTTP requests are needed.
GeneratePress theme has American, English-speaking authors. They have a fantastic reputation and respect in the WordPress community. GeneratePress is on 200,000 websites. They only make one theme so they can provide stellar customer service.
Tom Usborne is the author of GeneratePress. It’s his one-and-only, pet-theme. And it’s all he need ever produce. The theme only weighs 30k on the frontend. Fantastic achievement.
Some people complain the WordPress Twenty-sixteen theme is old-fashioned or plain. We think it’s fine for certain jobs. If the goal is to communicate and not impress with any animation, we think Twenty-sixteen fits the bill.
It’s not the year 2016 any more – but even today, we recommend Twenty-sixteen theme for speed. We’ve achieved page load times as low as a half second using this theme on cheap magnetic shared hosting.
Magnetic hard disk drives (HDD) are one of the most affordable ways to store large amounts of data. It’s old-school and used on cheap hosting like GoDaddy. The usual alternative is a Solid State Drive (SSD).
An SSD has no moving parts whatsoever. SSD storage is much faster than its HDD equivalent. HDD storage is made up of a magnetic spinning disk and has moving mechanical parts inside. HDD is physically larger than SSDs and much slower to read and write. In our experience, this still make insignificant difference in Time To First Byte (delays). There is no benefit for a website owner in reality. It’s all theoretical.
Many hosts brag about providing SSD servers. Yet, actually get worse TTFB than traditional magnetic servers.
So why do hosts offer SSD and claim it’s superior? It is superior but it doesn’t make a difference to the cyberspace “renter.” It enables drastic reduction in power consumption, maintenance, and lowers expense from air conditioning to cool overheated hard drives. SSD consumes a fraction of the rack space meaning lower square footage – less real estate is needed. These overhead cost savings are benefits not passed along to you the user. The benefit is long-term profitability for the hosting company. Not you.
SSD hosting is a cost effective way of differentiating a hosting service for marketing purposes. It’s pure specsmanship. It’s inappropriate use of component specifications to establish presumed competitive superiority. No such superiority exists in real-world speed tests of the whole system. The component specs are good. But the change didn’t remove the punitive delays caused by oddities. Thing such as Google-mandated HTTPS/SSL handshaking for pseudo-security or innate latency from distant geographic location.
Back to the real topic of Twenty-sixteen theme quality. We liked the look for what our goals were but … the load time was 3 seconds. Too long. We decided to track down the speed culprits and eliminate them. We got the load time down to 1.5 seconds with just a few simple adjustments. The examination tool we used was Webpagetest.org online website performance tester. Here are the things we discovered and some of the possible fixes:
1Watch out for automatic image resizing.
When we placed the custom header image, the size increased from a 30k PNG to 96k. The solution was simple. Just skip cropping. WordPress automatically saves files to it’s own preferred specifications. Cropping can work OK for speed with JPEG images – but rarely for GIF and PNG images. This sometimes makes the file size bigger – instead of smaller. That’s poor web engineering but as long as you double check image sizes for aberrations like this one, you’ll always have a fast site.
2Comment out Genericons in function.php or use a plugin to do it.
Genericons are a special font symbol. They aren’t really used by this theme. Genericons are a bloated bane we have written about before. Commenting out a single line of code in the functions.php file is one solution. Find the file using the Editor. Use the symbols /** in front and **/ on the other end of the single line that has the three words “genericons” in it. That gets rid of almost 40k of deadwood page weight.
Get rid of calling Google Fonts from their remote server and load local websafe or mobile-system-stack fonts instead. We confess Merriweather slab-face Google font has more class but there’s nothing wrong with using resident “Georgia” and “Helvetica Neue” – the fall back fonts in the theme stack. We used “Disable Google Fonts” plugin first. But it failed to remove all offending instances. We then installed “Remove Google Fonts References” plugin instead. It worked great. All font calls were removed.
4Install a caching plugin.
We used to install the free WP Super Cache plugin and enable all recommend settings. We don’t recommend this plugin now. It’s complicated and slow loading. We recommend using Cache Enabler plugin. It’s faster and simpler.
5Get rid of the funereal black border.
Under Customize, we changed the background color to #125faa blue for our border around the page. No change in page weight – but much easier on the eyes.
There you have it. We cut the page weight significantly and reduced the load time in half.
Speed is a component of User Experience (UX). Visitor’s check speed at a subconscious level. And even a conscious level for delayed page loads. Visitors judge site quality and web hospitality before assessing written content or images.
Site visitors hate slow loading pages. It’s frustrating and annoying waiting to see page content. Slow loading pages delay critical decision making. If a page is too slow, visitors leave. This abandonment reveals human intolerance and impatience. They want to have quick answers: “Is this the right place for me or not?”
Good user experience enhances source credibility. Credibility is trustworthiness, expertise, and enthusiasm. Speed cues us how much a site owner cares about our experience. It’s website body language.
Credibility is the intangible effect Google is evaluating. They examine 200 plus signals that generate their page ranking algorithm. Speed is one of those signals. How much does it “weigh” in the formula? Less than 1 percent. Focus on content quality. It’s the primary and biggest SEO signal. Content must be relevant and engaging.
User experience and credibility are intangible feelings.
Google determines relevancy and engagement based on measurable parameters. This is also called visitor intent. It’s not traffic volume (unique visitor counts). It’s affected by bounce rate, return visits, and dwell time.
Speed affects many metrics – including bounce rate.
Does data like bounce rate not make sense to you?
Bounce rate is the number of people who knock on your door and then leave before doing anything on your website. For example, perhaps 89 percent of your people leave immediately. They hit and “bounce” off the site. One look and they’re gone. Use the knowledge to improve. You will improve your site by measuring.
We’ve seen many different bounce rates. For example:
On a French life-style site, traffic is 100,000 visitors per month. 95 percent of those visitors bounce. They immediately click the browser back button or starting a new search.
In the French case study, 5,000 visitors had “intent” or “engaged” with the site. That is indicated with click-thru or long dwell time. Dwell time is how long they stay on a page consuming (viewing or reading) content. Click-thru is exploring further into the site.
On a medical site, there were 15,000 visitors per month and 85 percent bounce rate. This means 2,250 visitors engaged.
On our speed site, PagePipe, we have 30,000 visitors per month with a 18 percent bounce rate. This indicates 24,600 engaged.
So which site achieves the most successful engagement? We’d assume the French site with the most traffic is best. Is bigger better?
Let’s say all these sites are pretty equal in engagement. We could further quantify engagement quality by inspecting dwell time. How long did they stay?
Now “dwell time” is hypothetical, because people sometimes leave browser pages open all day long as they wander around the web.
In the beginning, where did all these blogs start at? Zero visitors. Often for months. It took time to determine what they were offering and how to make money. And build credibility.
There is no site credibility for first-time visitors. They’re suspicious and anxious. They assume all sites will attempt ripping them off. Speed helps overcome that distrust. Speed’s a reassuring feature.
Now, one NY site we evaluated had 1 million visitors per month. The bounce rate was 20 percent. That meant 800,000 people engaged every month. Wow! A gold standard. The site was about tech gadgetry for women. Female geeks. This is the most successful site we’ve investigated thus far.
The site owner is now a self-made millionaire. But her site became slow and clunky. The mobile experience wasn’t good and most of her visitors were on mobile devices. Her visitors tolerate that so far because they value her content. But her competitors are delivering similar content with better user experience. Speed is fundamental UX. That keeps her awake at night.
We couldn’t help her with site origin optimization. The site speed was swamped by the negative effect of ads. Advertising scripts from third parties were big culprits slowing down her site. We had no way to remove those without affecting her income. Speed was in the hands of third-party sources. Ads and affiliate links were her main revenue source. Sadness.
The quality of PagePipe’s visitors gets better and better. We don’t want more visitors. We want better ones. Quality not quantity. Will we ever have 1 million visitors per month? No. We’re OK with that. It’s not our goal. Interacting with people who want to improve is good enough.
Credibility translates into SEO ranking.
All three sites mentioned (but not NY) have about the same quantity of qualified visitors. How long they stay (dwell time) also indicates the quality of content. Again, when you engaged them, it indicates their investment of time in your offering.
How many monthly visitors you get is insignificant to efficiency. How many buy something is the metric that counts most. One visitor buying thousands of dollars of profitable services or products is worth as much as 5,000 who spend 50 cents. This is niche market thinking. It works for us.
One million visitors can produce no money. People must show interest in your offer.
The goal is helping a few people – or even one. Today perhaps you’re our one. We offer you a free education. This street knowledge cost us with learning-curve pain.
Bounce rate helps determine the quality of leads you get on your site. The lower the percentage, the better visitors are understanding you have what they’re looking for. It means you passed their mind filter. They understand what you do and why. They value it. “Staying” is their vote.
I’m glad to see traffic picking up. But out of all the people visiting my site over the last few days, I had one new subscriber.
Who cares? We get about a few subscriber per week. We collect subscribers to build a list. Our list is small – around 1,000 names.
On some sites, a signup is the most dominant site element. On PagePipe, selling ebooks is our emphasis – not list building. A list is cool – but it’s not top priority. What are you sending your list? If you never use it, is it valuable? Only if you start an email campaign. Then it may be an opportunity. Or an opportunity cost.
Is there a way to see visitor’s URL addresses from metrics and start marketing to them?
No. They belong to Google at best. But they don’t compile a list. List building is done with onsite email signups. You can own that.
The single best thing to enhance your SEO is write relevant content for your target audience. Make it your goal to get up over 50 posts. But it can’t be thin content. It has to be engaging and interesting. People won’t read or watch non-entertaining drivel. Solve problems. Save them from themselves and their foolish behaviors. Give away your knowledge. Then they’ll trust you.
PagePipe gives you free knowledge.
You can put too much energy into improving UX and speed. Then it’s an inefficient waste or obsession. It may not move the SEO needle even after two years of analysis. It may mean there’s deterioration or erosion of traffic to competitors. Sometimes keeping what engagement you’ve earned is good enough.
Metrics may reveal 70 to 80 percent of traffic is on mobile devices. Then future-proof with mobile-friendly changes.
We don’t regurgitate herd myths about speed. We want you to find the truth.
Aggregator websites collect and post syndicated material from around the Web, including news, specialized publishing, or the latest bargains and deals in Internet shopping.
Good aggregation helps readers find interesting news and information. It also gives linked sites added exposure. They are middle men, but they greatly benefit both sides.
Aggregator websites are all about monetization.Website monetization is the process of converting existing traffic being sent to a particular website into revenue. The most popular ways of monetizing a website are by implementing Pay per click (PPC) and Cost per impression (CPI/CPM) advertising.
There are many WordPress themes that cater to the needs of this niche market. As near as we can tell, there is no free theme that can do all the necessary features. But there are paid themes and one in particular is of interest: Bimber by ThemeForest.
Bimber has great Mobile UX. Banner ad appears underneath a top feature, thumb-swipable, horizontal scroll. No scrolling jank from the ad shoving things around as it loads on the page. Very nice.
Aggregator websites have resource intensive hogs like Disqus, Facebook, Twitter, Google Analytics, YouTube, Vimeo, Google News, etc. There is nothing wrong with using plugins that may be resource intensive, but you need to balance the trade-off of the plugins’ functionality with the speed and optimization of your website. This requires value analysis. Especially for mobile users.
Value analysis includes the following: combination, simplification, elimination, standardization, and substitution. This is strategic performance optimization. A value analysis audit examines every requested component and it’s contribution to site goals. This requires setting self-imposed limitations. Aggregator sites are seductive for adding too many features in the hopes of perfecting revenue.
Perfection has too high of a price. Metrics need to be evaluated. Only 20 percent of assets provide for 80 percent of the profit-producing features.
Bimber theme is touted as a lightweight aggregator theme. Their marketing says they are a “viral & buzz theme.” Those are unmeasurable, unregulated advertising claims. How fast is “lightweight”? Viral and Buzz are jargon, weasel words. This diminishes their credibility.
Test results for Bimber speed reports vary depending upon the date of the test, the location, and the testing tool. Times ranged from best-case 1 second to worst-case 5 seconds. that is quite a spread. And page weights varied from 830k to 1.7M. Again, a big spread. So testing conditions with changing parameters render objective benchmarking a flop. It’s hard to say if our performance goal of a 2-second load time is achievable especially internationally.
The average Internet page weight today is around 2.3M. None of these test reports used a demo page of that size. It also appears that aggregator sites are above average in page weight with above 4M of files. This diminishes the prospect of good speed (under 2 seconds) for mobile users.
Bimber is frequently featured on articles dedicated to fast WordPress themes. Fast, of course, is relative.
Pingdom Score: Load time: 981ms, Requests: 61, Page size: 830kB (Note: Aggregator sites are never this light – under 1M. These are unreliable results. A fake page.)
Google Page Speed Insights: Desktop: 89/100, Mobile: 85/100
Another recent 2016 test claims the Bimber theme loads within 1.5 seconds. (Pingdom score 84/100 to Amsterdam, Netherlands), Page size: 1.7MB, 83 requests. (Note: This is a more realistic page weight test. But still too small for any real-world aggregator situation. And only 83 requests is also not a real-world number. HTTP request would be in the 100s of calls.)
“On the (Bimber) homepage, you have a 728×90 ad banner, exactly below the header which will attract your users. A 300×250 is placed within the content so that it would look more like a part of your site rather than an advertisement. This is the most profitable ad placement in this theme.”
Bimber demo framebuster code won’t allow testing on browser-based iPhone simulators. Or on Yslow. So we couldn’t get results there. Yslow tests would have revealed how much lazy loading of assets were really occurring (if any).
About those HTTP calls: there were only 47 to 83 requests on these tests. That doesn’t represent the hundreds of usual HTTP requests that occur on a normal aggregator website. This demo is a best-case scenario. Or a mock-up dummy.
Bimber’s advertising page claims the theme is optimized for Google PageSpeed. We couldn’t care less about this test claim. But we ran the demo page anyway for verification. We got the usual frustrating, boiler-plate, error messages this lame test usually produces for WordPress sites. The scores are meaningless if the page loads in 5 seconds. Google doesn’t even use PageSpeed scores in it’s own ranking algorithm. Instead, Google uses Time To First Byte (TTFB). Bimber can get good time to first byte under the right conditions (322 milliseconds). That depends most on the hosting service.
This data shows the demo is “lazy-loading” about 400k of assets. But they are not using a lazy-load plugin. It may be part of the Google CDN feature or built-into the theme. We’re not sure. Most (half) images are stored in the media library.
The most important thing for SEO is Time To First Byte. Bimber TTFB is only 322 milliseconds (very fast) when hosted on a fast NGINX server (not Apache based).
But this speed number only influences less than 1 percent in the Google algorithm. Relevant content will always superseded speed for ranking. Speed is not a significant SEO factor. It’s a major UX factor.
The Bimber demo uses two free plugins from the WordPress plugin repository.
Sitelinks Search Box Free plugin so people can reach your content more quickly from search results. If you use WordPress SEO by Yoast plugin version 1.6 or newer, you don’t need to use this plugin, as this feature has been included in the version 1.6 update.
After activating the plugin you only have to “Wait for Google Search algorithms to identify your site as a candidate for the new sitelinks search box”.
2. W3 Total Cache plugin This is a popular plugin (over 1 million installs). Some people swear by it. But from our tests, we’ve never had improved speed results on a well-optimized site. The demo is successfully using WTO plugin for minification and content delivery network support. But other alternatives exist.
Note: W3T caching plugin wouldn’t be our recommendation. The most raved about caching plugin is WP Rocket. It’s a paid plugin we’ve never tested. And never will because all our tests of other available caching plugins reveal they only help improve speed if your site is grossly bloated. If the page is well optimized, there is no improvement whatsoever with caching plugins.
Elements that are significantly delaying or improving Bimber demo page load time:
At least, 1.5 to 2 second load time delay is for Facebook API JS libraries connections and calls. Out of 5 seconds total Facebook is a huge problem!
JS and CSS are minified and concatenated. This improves speed and is most likely being done with the W3T plugin.
The demo uses GStatic CDN (Google) to offload JS/images/CSS – and probably fonts. But mainly images. There are some blogger complaints that GStatic CDN does NOT improve performance and actually slows down page loading. It would require benchmark tests. We see CDNs as band-aids for sloppy web design.
About 1M of images on the Bimber demo page could have been compressed further by 379k (almost 40% reduction in file size — visually lossless). The speed gain would be less than a second – but worth it.
The demo page uses the Google viewport meta tag which means the content is optimized for mobile content. A viewport controls how a webpage is displayed on a mobile device. Without a viewport, mobile devices will render the page at a typical desktop screen width, scaled to fit the screen. Setting a viewport gives control over the page’s width and scaling on different devices.
The demo uses built-in Aggregation Functionality plus Pingbacks, RSS, Really Simple Discovery, and Window Live Writer Support. All of these things cause “drag” because they are offsite calls for assets or code.
We don’t know if the following are standard Bimber theme features or not: The demo also uses Open Graph Protocol (supported by Facebook). And Twitter Cards for linking content to Twitter. These cause drag.
Bimber uses “scrset” in their theme code for adapting images for high- and low-resolution displays. (Good for iPhones and iMacs).
Bimber Theme Conclusions:
Bimber theme has all the bells and whistles any aggregator business would love to have. It’s mobile UX is superior. It’s $49 price tag is cheap. To get load times below 5 seconds will take work. In the final speed analysis, every single HTTP request must be evaluated for contribution to site value.
So what’s a 10-second site? Extra slow? Man, that’s worse than extra slow. And how about a horrible 25-second page load? XXS (extra-extra slow)? This test is an embarrassment – a misleading sham.
How much money did Google spend creating this wonderful web gadget?
Face it. Very few sites are under 1 second. Even Google couldn’t do it on their model homepage. Perhaps only 1 or 2 percent of the web can make a less-than-1-second claim. This bogus test crushes the spirits of ordinary site owners. Do common WordPress sites have unlimited resources to throw at speed performance? Hardly. And if you are “1 second exactly” are you “fast” or “average”? They say both in their groupings. Confusing weirdness.
Wake up! Speed does NOT affect SEO directly. It affects user experience. UX indirectly affects metrics indicating user intent.
User intent is a major factor in search engine optimization and conversion optimization.
The speed test most professionals prefer is webpagetest.org for testing speed – and TTFB. That service is also owned by Google – but doesn’t wear their branding. It’s open source and free. Pingdom is valuable, too. Other tests we steer clear of for a variety of reasons. All tests give different results.
Tests like Google PageSpeed Insights are bogus, too. Making their suggested changes may change your score. They definitely won’t change your speed. And, never improves SEO page ranking. Wasted time and energy.
All Google’s suggestions for “improving” our PagePipe store ruins our site’s ecommerce features completely. It’s faster! Big deal. It’s broken. No sales – but dang it’s fast. Don’t trust the recommendations of any speed score.
Speed scores are irrelevant. Only load time in milliseconds and page weight in kilobytes or megabytes are what count for mobile. But isn’t perceived speed good enough on mobile? Nope. The weight loading behind the scenes consumes mobile data allowances. Feature bloat costs mobile users money.
What are the real averages not the ideal averages? We don’t live in a Utopian world using WordPress. How bad is bad?
So how good is good-enough speed?
MachMatrics has answers based on 2018 Internet data:
What is the average load time? 8.66 seconds.
What’s recommended? Under 3 seconds.
What’s the average webpage size? 1.88 megabytes (M)
What’s recommended? Under 500 kilobytes (k)
What’s the average number of resources? 115.6 requests
What’s recommended? Under 50.
What’s the average server delay? (TTFB – time to first byte) 2.11 seconds
What’s recommended? Under 1.3 seconds
“A one-second delay in webpage time equals a 7% reduction in conversions, 11% fewer page views and 16% reduction in customer satisfaction.”
“We found a clear correlation between a faster time to first byte (TTFB) and a higher search engine rank. While it could not be outright proven that decreasing TTFB directly caused an increasing search rank, there was enough of a correlation to at least warrant some further discussion of the topic.” – source
Those are NOT our recommendations for mobile devices. Ours are:
Load time 2 seconds or less
Webpage size Under 1 megabyte
Requests Under 25
Time to first byte (TTFB) Over 1 second is terrible. Less than 1 second is marginal. 500-millisecond TTFB is good. 100 to 200 milliseconds is great.
Speed is the unofficial gatekeeper to your content. – source
What were the Google-test official recommendations for PagePipe store?
Serve images in next-gen formats JPEG 2000, JPEG XR, and WebP are idealistic image formats. The difference they make is insignificant. And some browsers can’t even render the files.
Avoid an excessive DOM size A joke? Beyond the skills of most web designers.
Now this seems really bad doesn’t it. 2.4 seconds? Perhaps hypocritical. But we advocate Kinsta’s Bryan Jackson and his proposed hierarchy of user speed tolerance. It’s good speed strategy. Entry pages (our blog) need to be under 2 seconds. This is where first impression counts most. It’s why we didn’t use HTTPS/SSL on the blog. We need the 500 millisecond boost in speed.
When people are curious enough to enter our store, they’re more tolerant and wait the extra half second for SSL – but over on another host. They are engaged now. Brian says 3 seconds is the upper limit of speed tolerance. It depends upon the user’s perceived value of the expected content.
Fixing Think-with-Google’s speed suggestions are practically impossible on a WordPress site. If you pay to have these esoteric changes done, you’ve just wasted good money.
Do yourself a favor and don’t evaluate your site speed with this lame Think-with-Google test. Use WebPagetest.org or Pingdom.com instead. Be kind to yourself and your wallet.
If it ain’t broke, don’t fix it. Changing a site for mere redecoration is absurd. Aesthetic design alterations rarely make a difference in profitability. Graphic complication is often whimsical, not strategic.
Change can cause problems. Before now, it was foolish to update themes to newer versions without a child theme for protection.
Using a child theme was a prerequisite to preserving custom CSS coding. When we recently upgraded a clients theme to “redecorate,” the updated theme disabled the old One-click Child plugin completely. That’s not normal. We lost the customization CSS.
There’s more to this site conversion story:
Redecorating and speed improvements made no measurable improvements in web metrics. Those were measured by Google Analytics over a two-year period. That’s right. Bounce rate, dwell time, and traffic count didn’t change for the better. The site under test has 70 to 80 percent mobile traffic. These mobile-first changes made no difference over the short term or long term. The speed dropped below 2 seconds (goodness) from the old speed of 4 seconds (badness). The site was also enhanced to appeal to an all female audience. While this made the site feel much more credible, it didn’t change the traffic quality.
Why did speed and aesthetic changes fail to help this site? Simple. Content wasn’t valued enough by a larger audience. The pie didn’t get bigger or better. The content is written in a mix of formal and scientific writing. It’s boring medical jargon.
You can’t bore people and expect them to stay just because your site is fast and beautiful.
In the past, a child theme (plugin) was the best way to protect custom code during upgrades or updates. But that is no longer true.
Note: Using a child theme adds one request slowing down all pages.
Simple CSS plugin doesn’t add any requests. It’s a faster method.
Custom code is now placed in the Customizer. You’d do this in the “Additional CSS” section. But it is vulnerable, too. It’s better to use a plugin called “Simple CSS.” It appears right there in the Customizer ready to receive your custom code. And protects it during updates, upgrades, or theme changes.
We have an assortment of fast themes from the WordPress directory. We collect themes for evaluation. But those freebies don’t always meet the criteria of longevity and easy updates.
Using stock WordPress default themes for better speed.
Annual default themes are well designed. Annual generic default themes are usually conservative. No cutting-edge experimentation. Our recommendation is using a lightweight default theme. If your site has 70- to 80-percent mobile traffic, small screen users are a priority. Mobile-first ranking is number one for SEO. There are only six WordPress default themes working well as responsive designs and fast loading:
Often people think free themes are low quality. It’s quite the contrary. … Free WordPress themes are actually held to a higher quality standard. All themes in the official WordPress theme directory go through a strict theme review process.
There are some very talented folks in the theme review team who examine and test these themes before they are included into the directory. – source
Swapping generic themes is a torture test. It’s unrealistic to expect themes to be interchangeable. Swapping to another theme is often the same as “nuking” a site by upgrading. Then restoring (reinstalling) the original theme choice and assessing the damage. That demonstrates how resilient a potential theme upgrade is. It may be destructive testing.
You can create a staging area on your server using WP Staging plugin. But you can’t “push back” to the live area without buying the pro version. We do the tests using the free version. Then duplicate the changes on the live site after we verify in staging. It’s not that painful and safer than working live with unknowns.
How long are stock WordPress themes supported?
WordPress core supplies the last three-years default themes with the latest core version. That means they’re fresh. They’re updated for most-recent WordPress version compatibility. That means these are getting special support treatment and attention. Top of the pile.
Is their speed excellent?
Yes. We’ve had good results. Strip themes of Google Fonts, Emojis, other WordPress baggage, etc. Do those modifications with plugins not affected by updates.
Are they designed and supported by WordPress.org?
Yes. They’re prime. They’re portfolio pieces for the chosen theme designers. Based on history, we expect active theme support for at least 7 years. And there’s no marketing upselling to pro versions. They’re updated with each new core release.
What if your web developer croaks? Who’s their successor? Will they know what to do? Do you have to scrap everything?
If you fire your web developer today, you’re left with a difficult task of reverse engineering their work. You don’t deserve that vulnerability. It’s great job security for the developer – but not a good practice for you.
The last web contractor is always the fall guy. New developers will definitely change the theme for easiest possible upgrades – for them – not you, the site owner. They charge you to do that. They’d swear at us for our past choices – and say derogatory insults about our skills or mindset.
Most web designers and developers have an odd need to complicate sites. They add whizzy features like animation, sliders, parallax background images, etc. This makes it look like they worked harder. They love gadgetry. But these are the things that bloat a site.
Feature gaps are the things the theme of your dreams is missing.
The feature gap can be closed completely with the use of plugins. … The selection of the right plugins can make all the difference in turning a site that is almost what you want into one that meets or exceeds the needs of your business and your audience. – source
Free WordPress themes tend to be compatible with a lot more plugins than premium themes. This is because all the free WordPress themes in the official repository all have to meet certain standards to be approved. – source
Simplicity is our goal.
If we had a crystal ball, we’d predict theme difficulties or fragility. You can step into a trap. “Who is the guilty party?” That answer to future brokenness varies. Is it the developer’s judgment? The theme creator? The plugins selected? Or WordPress making big changes? WordPress is a dynamic landscape full of potential risks.
The goal is preserving the look and speed while hardening the site for future changes. Then you’re improving website “shelf-life.” Return on investment!
We document site specifications and upgrading procedure as a deliverable PDF. This is a style guide or brand manual. The goal is making it as easy as possible for another web technician to pick up the reigns and keep going. We charge $900 for style guides. That’s because once we hand it over, we’re out of the picture. No more income. Obsolete. The site owner can approach any developer and hand them our blueprint.
Speed websites are not canned or off-the-shelf. It requires a brain. You’re smart enough to figure it out. But how much of your life would you burn up? For us, we’re motivated and curious to learn new things. We enjoy theme experiments to increase our knowledge of future successful projects.
Theme changes have unknowable risks. All projects do. We fail many times in testing before there’s a success. You have to check the grief factor.
MIGRATION IRREGULARITIES We do important tweaking of site speed using the Plugin Logic plugin. It’s a secret weapon allowing us to activate or deactivate plugins using page or post URLs. The URLs are static instead of dynamic. Also known as Absolute versus Relative links. This means if the site moves to staging areas or is migrated, the URLs listed will not change to the new domain URLs. Plugin Logic settings are then pointing to the wrong addresses. Page and posts may appear broken because other plugin functions are not activated.
It makes WordPress look bad when a theme change causes big disruption.
You need a future-proof theme strategy. It needs to support long-term and be fast loading. Claims that a paid or free theme is fast loading doesn’t mean it’s true. We’ve done tests and written about this terrible marketing deception. It’s false advertising. Authors use exaggeration of better speed as a marketing differentiator.
We have investigated thousands of paid and free themes. They both share the same vulnerabilities. We wish paid (premium) themes would have better support and performance. Paid-themes are usually more complicated and have longer learning curves. There are no guarantees you’ll get what you hope or what’s in the demo. Theming companies and authors sellout or go bankrupt overnight. They are under no obligation to support a theme. They can stop at any time.
We don’t like this vaporous aspect of the WordPress world. It’s inherent in open-source volunteer communities. It plagues plugins, too. Even some of the biggest and most popular plugins can go sour.
Technical volatility is a motivating factor to not sell WordPress web services.
So what is the safest and fastest theme strategy? We examined a new website to see if we could speed it up. We asked, “What theme is this? We’ve never seen it before.” He replied, “It’s stock WordPress’ Twenty-sixteen – customized.”
We were stunned.
We never realized you could build on top of one of those generic themes and still have it look great. It requires stripping the theme and then building up features with well-selected plugins. He was using our “speed strategy.”
WordPress is proud of long-term support for their annual themes. It’s a matter of professionalism for them. That works to our advantage. WordPress is not going away. Neither are those annual, pre-packaged themes. From our tests, those featured-and-endorsed themes are fastest.
Don’t throw money at a premium theme. Give Twenty-thirteen through Twenty-nineteen themes preferential consideration as long-term theme solutions. We recommend that direction.
Any theme, paid or free, can be abandoned or even banned. That’s the risk of the WordPress world.
For example, WordPress suspended Zerif Lite, a theme with 300,000 installs for 5 months. Why? Because they didn’t keep widget content active after upgrades. That non-compliance cost that author (ThemeIsle) $35,000 per month in revenue. Today they only have 100,000 installs. Ouch. Big hit.
So we place our bet based upon reading the signs of theme credibility and longevity.
Accelerate theme uses a widgetized front page for homepage customization. It’s a common theme-developer workaround. In times past, its only flaw was when making upgrades. All widgets “retired” to the Inactive Widgets section. They then were installed again one-by-one (drag and drop) to the right locations. It was a confusing puzzle.
Accelerate theme is not the Lone Ranger. All themes dependent on “widgetized” front pages potentially have this same not-so-well-known bugaboo. It’s not advertised – that’s for certain. To fix the errant Zerif Lite theme, the author’s add a stopgap plugin to maintain widgetized page content.
PHP 7 is twice as fast as PHP 5.x and requires one fraction of the server memory.
Does activating PHP 7 from version 5.6 make a difference in speed test scores? There is no measurable load-time difference in milliseconds. None. Why no improvement?
Some claim to add PHP 7 instead of a version 5.6 speeds up page loads by 300 milliseconds on a cache-less site. We don’t see that improvement evaluating with normal online tests like Pingdom.com or WebPagetest.org.
On a well-optimized site, there is no speed change evident. But the same can be said about caching plugins, minification plugins, and CDNs. With proper origin optimization, there’s little benefit from inadequate speed fix-it attempts. Band-aids.
Our PagePipe blog is a well-optimized site. So why would we even want to risk a change? We write about plugin technology and speed. We stay current to “walk the talk.” We also have an insatiable curiosity. So we did it.
A 3,760-word article at WP Elevation is about the pain of producing websites. The article expresses everything we hate about website creation. The thought of building “explosive live hand grenades” stresses us. Just reading the article was stressful. Why?
Because it’s true. The nit-picky horrors described are exactly what occurs during web projects. Client or website owner expectations are high. Their technical knowledge is often low.
A new monster arose on the WordPress horizon.
The fragile nature of WordPress and PHP v7.x.
Why does adding PHP 7.x break your site? Our choice to transition our GoDaddy-host-server to PHP 7 rattled our nerves. And we’re initiated in this stuff. Our experience is a good example of what goes wrong. Upgrading PHP version 5.6 to version 7.x is a simple C-panel setting – but not without potential consequences.
PHP 7 released long ago on December, 3rd, 2015. GoDaddy didn’t add this server option for a year and a half. Why? Because they knew the changes might break hundreds of thousands of WordPress websites. They left it up to users to perform the update. And they delayed the service call costs for as long as possible. The GoDaddy default version was set to 5.4. Making users choose their poison was smart. Users then are responsible for breakage. Or dialing back the PHP version themselves – or tracking down fixes. GoDaddy is blameless – sort of.
Above: Pie chart – Percentages of WordPress sites using different PHP versions.
Risk breaking my site? Why even care about PHP version 7.x?
PHP is the code of WordPress, a server-side programming language. It first appeared in 1995. All themes and plugins use PHP, too. Upgrading your site to run on PHP 7 instead of PHP 5.6, you’ll improve the performance of WordPress core by 2x. That’s right. Twice as fast is the typical gain in core speed. We anxiously waited and watched for this no-extra-cost, speed opportunity. Free speed. Most vendors upgraded long ago. So we felt snubbed. But we didn’t change hosts. We like bragging about good speed achieved under the worst conditions!
So how faster does WordPress core load? We should see a 100- to 300-millisecond improvement. But we never see betterment in testing. Updating PHP is a theoretical improvement – not an actual improvement.
PHP running twice as fast doesn’t mean your website loads twice as fast. We’ve never seen significant, measurable differences switching back and forth between PHP 5.6 and PHP 7.3 on various hosts. For us, it’s a theoretical improvement in speed. We think it’s good – but if you don’t change versions – no problem for us. It might be a problem for you at a future date.
PHP 7.x isn’t going to break WordPress – it may cause some of your plugins to malfunction. Perhaps your theme. But the result is the same, your site appears broken. You can test all your plugins using a free plugin. Naturally! We tested with:
Display PHP Version plugin displays the current PHP version in the “At a Glance” admin dashboard widget. We like it.
So what broke after the change from 5.6 to 7.x?
Broken Link Checker – compatible – warning 1 – This plugin broke the site when viewed on an Apple iPad. Meaningless code was all over the screen. We disabled the plugin. This is plugin causing site drag anyway. Only use it during maintenance. Leave it disabled. On some host, they ban Broken Link Checker. Why? Because it overloads the server slowing down other sharing domains.
Simple Content Adder – We got a red flag for the file revisions.php. But we couldn’t find it breaking anything. We left it as-is.
SS Downloads – red flag – This favorite old plugin broke the site with PHP error screen. The plugin failed because it triggered a fatal error. The plugin is for email capture before PDF downloads. We had to dump the plugin. Presently, all our free downloads use MailChimp signup. We do product downloads with Easy Digital Downloads plugin on our store site.
Title Experiments Free – compatible but 7 warnings. We wrote plugin author, Jason Funk, and he updated the plugin to version 8.9 for PHP 7.1 compatibility. No more warnings. Thanks, Jason. [Jason later removed this plugin from the directory.] It caused global loading – site drag.
WordPress Popular Posts – compatible – 24 warnings. The plugin stopped gathering data for page visits. This is the primary reason we use this plugin. It’s very popular with 300,000+ active installs (v3.3.4) The new version 4.0.0 is now PHP 7.x compatible. It has slow Font Awesome onboard but it’s not enqueued. We’re thankful. We like the new GUI control panel for the plugin. The original plugin was a 125k zip file. The new one is 759k. Most extra weight is font overhead for the control panel. It doesn’t affect your site’s front end. This newest version is now available on the WordPress directory. We don’t use this plugin any more because of site drag.
How fast was PagePipe home page after the switch to PHP 7.x? 699 milliseconds unprimed cache and 440 primed. Superfast even on GoDaddy mechanical, shared server with no CDN.
So a quick comparison of primed cache: PHP 7.2: Pingdom NY PagePipe home page – primed cache: 559ms. PHP 5.6: Pingdom NY PagePipe home page – primed cache: 567ms.
8 milliseconds gain with TTFB fluctuations. Maybe? Insignificant gain on an optimized page. We’d be better off economizing in other areas.
Why no big gains? Because we have super optimized the homepage. There’s a point of diminishing returns. Only fat bloated slow websites benefit from the PHP version switch.
PHP gains are overrated and exaggerated. A bloated site gets the most improvement. So PHP is a cheap test of site bloat. Big speed improvements from PHP indicates a big potential from the investment in site origin optimization.
While we’re not selling jewels, the Shopify aesthetic at Awe grabbed our attention. The UX and speed leave a lot to be desired, but the feeling is good enough to make up for it.
Is it a cliché store design? Yeah. But that may be the expectation. If you get too creative (expressive noise), you won’t make sales (confusion of choices). People can’t find things when convention isn’t followed.
Four sample web pages from Awe ecommerce:
This site is a common Shopify example. They throw millions of bucks making their services fast loading. But silly designers still slow down Shopify pages with apathy. It could load in under 1 second. Instead, instead it’s 3 seconds. Is that horrible?
It depends upon the audience. If they are a mobile audience – YES. It’s a waste of bandwidth from being careless or apathetic about speed strategy. Speed is the first hurdle to a good user experience. UX is about hospitality.
UX observations of these Shopify pages:
Color is the biggest influence of quality and habitat cuing after speed. “Does it feel good to be here?” The large images set an elegant mood with female lifestyle inference. They use white space and solid colors. Elegant themes generally have higher usage of blacks and metallic colors. We could say tan is gold with no reflections and gray is silver with no glint – and that is believable. It has a blackish footer. We’ll judge this site theme as elegant. It looks like their specialty is gold and silver – so that works.
Sliders don’t work unless they tell a story. A single non-moving image is better. Is this a picture story? We can’t tell – but we guess not. Mere glamour shots. The slider was “present” and the designers couldn’t resist not filling it. We’ve written about this:
Some core web UX philosophy: Avoid sliders unless they tell a captioned 3- or 4-panel story. The imagery better be good. No image links. No rollover pause control. This flub disables the story – a negative feature.
Almost always, we’re anti-animation. No moving parts on a web page. Why? Have you ever tried to read a page with a fly walking around on it? Distraction.
That doesn’t mean we haven’t experimented with motion. But the attention trickery is usually a UX failure.
These pages have minimal animation – but no animation is good. It’s often a content-cramming ploy.
Our reference about building big hero images that are fast loading:
We’re not in love with double layer navigation bars. Bad UX. One or the other gets ignored. It’s a trend or a fad. Is it a solution? It shoves everything down the page a little further. Vertical waste is negative. We don’t like the dropdown menus that cover content. Is there no alternative?
Do we like the gimmicky photo image swaps on rollover? No. It’s hiding the content. It may be cleaner but it doesn’t promote product understanding. We’d rather have a magnifier show product detail.
The coolness feature of the day. Iron boat anchor and it’s live in a few minutes. We tested it. It says Jill will answer but Sharon did (Screener of Incoming Calls?) and she’s a real person. This costs money. She disappeared for 25 minutes when we asked her location and if she worked for Shopify. And she promised to “answer any & all questions.” Disappointment.
But usually chat is fake.
Then Real Jill came online and said she was the only one in the office on the holiday and was swamped. Now that’s the real story. We told her she deserved a raise.
No. You can easily overbuild slow sites there, too.
Our lame store-building experience with Shopify. Fast (under 2 seconds) but not good functionality.
Q&A with Kendra Heim, PagePipe Staff
1How hard was it to set up a new account? Super easy. Enter a valid email address, create a password, answer a couple questions about your business, and finally enter your personal information (physical address, name, phone number.)
2 Got any reference on how to set up an account?
3Tell us about the decisions you made when making the store. What was most difficult? The most difficult part was figuring out how to make the home page a viable storefront, or more accurately a viable catalog page.
4What are the limitations or design constraints with free Shopify? There are only 8 free themes to choose from. Each has its set parameters and variations. These include menu placement, logo placement, added functions etc. You need to do some looking to decide which fits your specific needs.
For our store, we chose Minimal with the vintage style. You can’t add much design to the page, only pre-designed elements, and this doesn’t work freely on every page. Some pages allow you to add elements, while some do not. This is good and bad. It keeps websites cleaner and better looking if those using the design tools aren’t familiar with design, but it also cripples your ability to design if you do understand how.
You can change colors. But you can’t add any color blocks or change the location of the text, or how the text alignment, etc. You can’t add a photo and change the size or move it around. The design is foolproof but kills a creative mind.
One problem was trying to make the homepage a catalog page. There were ways to add featured products and collections, however, clicking on these would then take you to a product page or another catalog page. It was adding steps. To get around this, we used a simple image and text element with a button and linked this button to an ordering page. While there is one extra page for ordering, it’s better than having multiple added pages.
Another reason this was necessary was to ensure we could use an app for quantity bundle discounts. There are some free apps. They allow you to apply a bundle to only one product before you need to purchase the app. We had 4 products. In order to get discounts on all 4 products, we created one product with 4 variations. This allowed us to use a free discount app. The app we chose was QD because the discount information wasn’t obnoxiously placed on the product page.
What do you hate most about Shopify? I can’t move anything or add what I want, ie. text and photos.
5If you had it to do over again, what would you do differently? Find a theme that matched as close to what I wanted first. I wasted time trying to get something to work that probably never would have.
6How much did Shopify tempt you upselling paid features? I didn’t feel tempted because I didn’t want to buy anything. But, they do cripple you so much that they make purchasing more apps and features appealing. It seems easier than figuring out a workaround.
7What was the hardest limitation to deal with? The set elements and trying to get their idea of a homepage to work with what we actually wanted as a homepage which was a catalog.
8What helpful tips would you give someone learning Shopify? Keep it simple.
9What were the original goals for the project? Did you achieve them? Create a storefront using Shopify. The home page needs to be a catalog page. There should be product pages for each product. It should integrate with a WordPress site. I think I did manage the original goals, but we also found out that more had to be done on WordPress than originally thought. Shopify would not let you have product information pages for what we wanted, so Shopify became strictly a store. The WordPress site will need to have the pages for product information and contact information while Shopify manages the store by using a catalog home page, an ordering page, and a cart. The ordering page is similar to the cart, but we can’t get around that.
Our contempt is about fake freebies. Shopify is a bait-and-switch worst-case and best-case an upselling ploy. They deliberately cripple the product so it’s practically useless. This is a common marketing strategy deplored by idealists. Unless it is the idealist’s business that profits from it.
NOTE: Why were we even tempted to examine Shopify when we write about WordPress speed? It’s a struggle to get 3- to 4-second page loads on shared hosting with WooCommerce. But when “free” is your criteria, Shopify is insufficient in that class.
Conclusion: We scrapped the idea of a hybrid site combining WordPress blog and a Shopify store. Too restrictive without shelling out money for extras.
Online speed test scores are especially useless. These include lame tests like Google PageSpeed Insights. Many test-result recommendations don’t make any measurable difference in speed. They’re not real-world practicalities. They sound rational and appeal to logic. But they’re a waste of time and energy. You’ll drive yourself mad attempting to achieve perfect test scores. Google doesn’t even use test scores in page ranking.
So if speed test scores don’t matter, what does?Forget scores. It’s load time in milliseconds and page weight. Cumulative file size (page weight) is expressed in kilobytes or megabytes. Load time and weight make the biggest difference, especially for mobile audiences. Speed is critical for a good user experience (UX). Readable, relevant content is most important for SEO. But impatient visitors won’t wait for slow pages. Then you waste your hard-earned relevant content and it’s never seen. Most sites are “not good enough” for mobile users.
Google gives you technical speed suggestions. These are published in their webmaster recommendations. Few make any difference in mobile speed. Why? Is it sadistic torment of website owners? We wonder. The most infamous waste is testing with Google PageSpeed Insights.
We still feel this version 5 is the most inferior speed test for site evaluation. We don’t recommend it. Google conveniently now excludes Google Analytics and Google Fonts from the scoring. What? You’ve got to be kidding! Those things slow sites down – and they’re sweeping that liability under the carpet? Absurdity. Biased results. They changed the scoring so everyone wins. It’s a fake guarantee. So what do we recommend? Pingdom.com or WebPagetest.org.
Here are some weird things Google often recommends *fixing*:
This imperious rule is a great waste of time. WordPress almost always breaks this rule. The return on investment in improved speed is impossible to measure. It’s so small and insignificant. And there is a high risk of breaking your site. If this *fix* tempts you, you’re a programmer and should leave PagePipe. You’ll hear disturbing things you don’t like here. Plugins don’t help much – like the oft-recommended Async JS plugin. Work on other things that are more important like image optimization or removing heavy assets like chat, sliders, header video, and third-party services.
2Minify and concatenate CSS, JS, and HTML files.
These improve your score – but rarely improve your speed. Scores are meaningless. It’s milliseconds of load time that count. Minification plugins frequently break sites. We don’t use a minification plugin on PagePipe. We’ve experimented with many. They make no difference in speed and only cause problems with Easy Digital Downloads plugin – among others. Does that mean we never use minification? No. If nothing breaks, fine.
Trimming HTTP requests cut down on the number of calls to your server. The big theoretical gain is from concatenation (combining or lumping) JS script, HTML and CSS files together. These can break your site – white screen of death. It’s not worth it merely for irrelevant score improvement.
3Reduce server response time (TTFB 200ms).
This goal is laughable.TTFB is time-to-first-byte. It’s a measurement of server response time in milliseconds. Even Google’s home page can’t produce 200-millisecond time-to-first-byte specifications. That is because of HTTPS / SSL certification which introduces an additional 400 to 500 milliseconds of handshaking delays on the server.
Try Google.com on the *peerless* PageSpeed Insights test. They can’t even pass their own tests on the simplest of pages. Enough said. Zero credibility.
How much load time does www.google.com take? Using WebPagetest.org online test (owned by Google), the result is 2.196 seconds. The full-load time can vary from 1.270 to 2.204 seconds depending on the test server location. That’s right: seconds. Not milliseconds.
When is optimizing images worth it? Well, we tested a page and Google PageSpeed Insights said we failed. Why? Because we could still optimize two images by 2 percent. What!? That’s ridiculous. And it didn’t detect resizing one image dimension to save extra weight. That was more noteworthy.
Is it worth optimizing to remove 2 percent? No. Why? Because images are loaded in parallel. Improving image weight doesn’t make as big of a difference in speed anymore. Browser image loading is faster now.
Should we still optimize images? Yes. It’s important. But we can’t only optimize images and say we’re done.
Imsanity is one of the best plugin solutions for automatic optimization.Why? Because it’s easy to set image size reduction to column-width size and also set the JPEG image quality to 70. WordPress default is 82 but only on cropped images – not on uploaded originals to the media library. Imsanity crops originals. Most other image optimizer plugins charge money once you pass a certain threshold number of images. So beware.
What’s better is plain-old, manual sizing and visual, save-for-web optimization in an image editor like Photoshop, GIMP, or pixlr.
Google.com scores a 90 mobile. And desktop: a 99. I guess now we know what the standard is for Google’s own self-created criteria. Google itself. Please. Mercy.
It’s not enough to just compress the images. Specify the dimensions of the image, or else the browser loads the entire image and then mathematically resizes it to the required dimensions. Stretching with browser math delays image rendering.
5Avoid landing page redirects.
Simply configure your WordPress site. Go to the “Settings > General” page. See if “WordPress Address (URL)” and “Site Address (URL)” options include “www” prefixes. If they do, remove them, save the settings and that redirect will be gone. Tip: Use the Redirection plugin when removing old posts to protect SEO.
6Leverage browser caching.
Browser caching is beneficial for return visitors. That’s about 20 percent of your traffic. On a well-optimized site, caching doesn’t help much. When an unprimed cache loads in 750 milliseconds, how much better do you need to get? Adding a caching plugin may get the speed down to 400 to 500 milliseconds best case. That 350 millisecond gain is good – but not if the caching plugin is complex or breaks your site with concatenation. We only recommend one caching plugin and it’s Cache Enabler. Make one simple setting: change the far-future cache expiry to 8760 hours (1 year). You’re done.
Now, will a caching plugin get your sluggard 20-second page load down to a wonderful 2 seconds? No way. You won’t even measure a noticeable difference.
Ridiculous. How do tests propose we do that on a WordPress site? This is sometimes referred to as above-the-fold content. Over-optimizing for speed is not worth the frustrations fixing the problems. Run a PageSpeed Insights test on Wikipedia or YouTube. Bet the scores are crummy, poor, or *needs work*. Suggestion: Ignore this silly test result.
Other odd recommendations that usually don’t help speed much include:
Content Delivery Networks – CDNs are servers located in various geographical locations. They are closer to the users’ location and can reach content to them faster than the original server. A well-optimized site doesn’t need CDN. It’s a band-aid. Cloudflare has a free plan that we recommend you avoid. That’s right. Don’t use it. It slows down or delays your pages.
Accelerated Mobile Pages – Ignore making your webpages Google AMP compliant. This wasteful gimmick was announced by Google in October 2015.
Database Optimization – A periodic check and spring cleaning of databases to keep them lean and easily searchable. Cleaner plugins remove duplicate data, unwanted post revisions and more. Do they help with speed? Not anything we’ve ever detected. But it sounds like a good idea. We do it regularly. But it’s a matter of being vigilant and sleeping well at night. We’ve never seen a speed improvement.
Removing query strings from static resources in CSS and JS files – Developers use “?” and “&” to bypass cached files before they are purged. However, URLs with “?” and “&” are not cached by some servers. You can use a plugin to remove them. This removal improves your score but not your speed. Remember, caching doesn’t help so much.
Combining Images Into One – CSS Sprites – CSS Image Sprites were born out of the need to reduce the number of HTTP requests made on a website. The typical use of image sprites is for icons. This is where you bunch multiple images together into one big file. This is not very helpful and frequently won’t work on mobile-size screens. We say, “Forget sprites.”
So if those test parameters and tricks only help with scores – but not speed differences. What does matter most for speed improvement?
1. Themes without bloat. Popular themes like Divi and The7 alone take seconds to load – without any content or plugins. A typical responsive, free, stripped, WordPress theme with no bells-and-whistles loads in 40 to 50 milliseconds. Simplify.
2. Good Hosting – And we don’t mean expensive. If time-to-first-byte is too long (over 1 second), the only choice is changing your hosting service provider. You can have fast TTFB on shared hosting. We get around 250 to 500 milliseconds on cheap GoDaddy hosting. Test at ByteCheck.com.
3. Lazy Loading – enable lazy loading for videos. WordPress core lazy loads still images as of August 2020.
4. Gzip – A compression technique reduces code files for faster transfer. It also saves mobile bandwidth.
5. Reducing Redirects – Get rid of as many redirects as possible. Redirects are good for SEO traffic. But you slow down the browser a little.
6. Disabling Trackbacks and Pingbacks – Trackbacks (manual) and pingbacks (automatic) appear in content moderation to let you know that someone else has put a link of your post on another blog or site. Most of these links are spam. It there’s too much of it, it can affect site speed. Disable them under Default Article Settings > Settings > Discussions. Or we can use a plugin that can deal with spam, like No Self-ping. Or use XML-RPC deactivation.
7. Disabling Hotlinking – Sometimes other *web criminals* use the content hosted on your site’s servers for their own websites. This is simply an extra load on your server. To stop others from using your server resources, the recommendation is changing your server .htaccess file code. A plugin will do the trick.
8. Identify plugins slowing down the website. How much will this help? Our experience is you may save up to a second. That’s great on a site that’s loading in 4 seconds. But if it’s a 20-second page, you’re still in trouble.
Remember, it’s not the number of plugins that slow down a site. It’s the quality. We have 43 plugins. Our homepage and most others load in under 1 second on shared hosting.
Use the performance report generated by the P3 plugin to remove or selectively disable the worst plugins dragging down site speed.
No WordPress theme has gotten more Elementor-related hype than Astra. Judging by the comments on Elementor groups, you’d think Astra is the leading theme. But when we’ve seen surveys counting actual users, Astra comes in below GeneratePress. Astra contains speed potential. But a novice overloads this theme like any other. It’s lightweight (less than 50K on the frontend). We tested that spec and it’s true. Astra has the potential for *unparalleled* speed as the authors claim.
The Hello theme is from the same team that created Elementor. You might think that means it would be best for working with Elementor, but you would be wrong. The Hello theme has been around for years. It didn’t get much traction until recently when they released the theme to the free WordPress repository. Lots of people jumped on it right away. But many of these people realized that Hello was not the answer for them. Why? Hello is so barren of features that it’s unusable for anyone who isn’t proficient in CSS. Another sticking point is the fact that Hello requires Elementor Pro ($49 annual rent). It won’t work with just the free version of Elementor.
Regardless of the theme, every website needs help to reach the most speed potential. Find out the basics of what you need to know: Start here.
Which is best?
Unless you’re good with CSS, Hello is a non-starter. Hello fails to take part in the race. The others are both great themes, are fast, and have good support from their developers. And they work great with the free and the Pro version of Elementor. They also have premium features available at reasonable prices.
But premium will double the theme page weight – and slow down your site. They forget to tell you that.
For example, Elementor boasts they replace 17 plugins with Elementor widgets. They claim this saves you money and improves speed. But the 17 plugins they claim to replace, we would never add these features to any website. They cause slowdowns or poor UX:
Maintenance Mode/Coming Soon
Email Marketing Services
Header & Footer
As a final insult, Elementor says 20 plugins is the largest number of plugins you should have on any website. That is hokum. We often have 50 to 70 plugins and our pages load in under 1 second on cheap hosting. The average number of plugins on WordPress sites is 25.
What about speed?
What will ruin the speed for your site is not the theme – unless it’s a slow one like Divi. It’s all the junk you add to it. It’s designer apathy that ruins speed.
Keeping themes fast requires site owner self-discipline.
We’re sharing speed numbers from a recent test. We built a fast Elementor page using our origin optimization recommendations. Then we built the same page without Elementor – and threw in an extra-large featured header image.
Automattic introduced the Twenty-seventeen default free theme way back in the fall of 2016. It loads in about 25 milliseconds. But only after stripping the theme and WordPress core of unnecessary features.
That’s right. It’s not even on the popular speed theme list above. Do we use this theme? Yes. Why? Because it won’t retire. It’ll be maintained for 10 years. That’s based on the historical performance of past default WordPress themes. Longevity. It won’t disappear soon. And it’s free. No pro or premium version to pay rent every year!
What do the pages look like for comparison?
And below without Elementor:
Do we remove or recommend removing Elementor? That may sound like a piffle. But when TTFB on a slow server (like BlueHost) is 1.7 seconds, you only have 300 milliseconds left to load everything. That’s achievable but it’s a speed miracle. Gaining extra overhead to pad your web performance budget is a relief.
Both test pages loaded in 1 second (+/- 30 ms) using WPT.org.
page weight 314k, 549ms, 18 requests without Elementor
page weight 715k, 1.57s, 19 requests with Elementor
1 second load time difference with this test! The two test sources reveal different results. Which do we trust in the end? Neither. We trust our browser timer most.
Timer results: 1.2 seconds with Elementor pagebuilder and 930 milliseconds without Elmentor. Both pages with unprimed cache. 270 millisecond is the difference on Firefox browser desktop.
1.71 seconds (with Elementor) vs 850 milliseconds (without Elementor) on Google Chrome browser. That’s 860 milliseconds faster. Go figure.
But no matter which speed test you use, it’s intuitive Elementor is slower. Elementor makes significantly more requests and adds more scripts and styles to pages than using a simple column grid plugin with a fast theme.
So does using Elementor slow down desktop pages significantly? No.
But the difference in page weight will slow down mobile. There is no noticeable difference in load time on desktop. It only matters to mobile users with limited remote bandwidth and/or metered data. How bad is that?
For some sites, that’s 70 to 80 percent of visitor traffic.
NOTE: The test pages are both optimized and enhanced for speed with the same performance plugins. Both are on the same site, host and theme.
While internet mobile speeds dropped by 7 seconds this last year. The average load time is still 15 seconds. Horrible mobile connection times.
This link below gives solid information about the state-of-the-moment for what is normal and bad mobile speed:
“Statistics” and bloviations from SEO gurus say more searches are from mobile devices. Some sources say the growth in mobile search makes desktop searches irrelevant.
Our skepticism increases.
You can bring in Google Analytics data from their online store. You can add it directly to your own Google Analytics account. It’s mostly to help you learn how to use Google Analytics when a website doesn’t have much going on.
Google’s online store is interesting. In the section showing what kind of devices people use on Google’s online store:
70 percent of visitors come from desktop
26 percent originate from mobile phones
And the rest (small percentage) were from a tablet.
How can *THAT* be? Has Google – and cellphone companies – been lying to us? Do SEO gurus not know what they’re talking about? Is all the conventional wisdom wrong? Is this an expected glitch from noon traffic?
Here’s a screenshot from the Google Analytics data for Google’s online store:
The title SEO “expert” is an immediate red flag of poor source credibility.
The trust level for purchase transactions on a desktop is still higher than mobile. In Africa or The Philippines, you have no choice. Mobile is your only option since you may not afford a table to but a computer on.
Most shopping happens from a company-owned job-related computer. That’s right. Goofing off. Especially during the holiday season. It’s estimated employees goof off 30 percent of the time. You can fake you’re working while shopping and listening to Pandora. Perfect. So is it real shopping? Uh. We don’t think so. Wasting time is not shopping with a purpose. It’s window-shopping from boredom.
OFF-TOPIC: Take for example the statistic that WordPress owns 1/3 of the Internet. Is that real? Or a myth they don’t squelch because it makes them look and feel important? No clue.
We don’t know what to trust in the world of metrics anymore. Google doesn’t disclose how it gets information. It’s secret. They don’t tell us if it’s skewed by something or biased. In a sinister world, it would be completely fabricated.
If their data is so perfect, why do they fix algorithms every month?
Google does more to slow down mobile traffic and page loads than any other culprit on the web. Yet they claim they’re advocates of speed. Smokescreen.
One question is who are the people in the sample? Are they global or domestic? If this is store data, what is the buying behavior of their particular sample? How stale is the sample – real-time? Are they last year’s average? Cyber Monday?
Anyway, a few simple questions would blow some holes in this information.
This much is real – Google spins data to their advantage. Vanity metrics.
Our experience working with clients around the world is “mobile is worrisome.” But not for everybody. For example, 85 percent of our readership is on desktops or laptops
The people who ask us to “fix” their websites have high mobile rates: 70 to 80 percent of mobile traffic. Where do they get that information? Google Analytics! Is Google Analytics a trustworthy source of metrics? Not for a minute. They have improved a few things to remove artificiality. For example, until recently they counted visits from bots as real visitors. That meant an easy error of 10 to 30 percent on some sites.
Last year, they did various adjustments to thwart people generating fake content. The people the hardest hit were selling snake oil. Those were the people approaching us for fixes. All brainwashed that SEO was the most important thing to mess with. They wanted speed for better SEO. Even though we told them that was a falsehood. They still wanted speed for SEO, just in case, we were wrong. Too weird.
We don’t have to brainwash people that speed is important. Google does that for us. Free sponsorship.
So we use Google Analytics to track post popularity. The reason is that it’s faster loading than a “Popular Posts” plugin. Those plugins track metrics inside WordPress using the MySQL database. The way we do it is faster anyway. Last year our traffic was up. Google did some adjustments and the number dropped by 20 percent. We also realized we had a slow tracking beacon on our homepage. That’s our showcase page. That’s the page most skeptics test for speed to see if we walk the talk. Google Analytics slows down pages. So we removed it from the homepage.
We turned off tracking for the homepage with selective activation. The numbers dropped another 10 percent. That was fine since that was before reported as 10 percent of our visitors. But we care less about how many people come to our website front door. We’re not interested in “funnels” and “paths.” We only want to track the popularity of articles. Not our site popularity.
Now with all this jockeying around, did our bundle sales drop? Maybe. They certainly didn’t increase. But we can’t attribute it specifically to lower traffic. It will always be unknown even if we get sucked into the SEO vortex – like many other web fools.
So many unknowns based on half-truths and misdirection.
What we’ve seen as far as web designers go, they still build for desktop first. Not mobile. Why? Have you ever tried to build a website using a phone? Crazy. But people do it. A phone is how many manage online stuff. Not the way we want to do it. We hate looking at tiny screens and trying to push things around with our fingers. We test with mobile devices. We don’t build on them.
Based on our site “data,” designers prefer desktops. And according to Google, 75 percent of our traffic is international. But do we build for mobile-first? Yes. To the extreme. Why? Because that is our positioning strategy. It’s what differentiates us from the herd of wannabee speed experts. That is a fragile position. But it gets us conversations with some interesting people and problems. We then become even more expert. We figure the 25 percent who look at our site on mobile are the same people viewing on desktop. They’re checking our credibility.
Anyway, what makes a difference? We don’t know. We may not even care. It has an opportunity-cost checking metrics every day. But content that solves problems (real or imagined) seems to work for us. Do we care what device they’re looking on? No. Because we build for mobile since that is the anxiety and pain for our readers. Real or imagined pain. We may be selling placebos. If it makes users feel better, that is OK with us.
So interesting side note: We’ve had 3 or 4 international bundle sales this year when downloads failed. Why? They were downloading a 3-megabyte bundle on an iPhone! The connection timed out. We ended up supplying those 4 bundles through WeTransfer.com. But that put the kibosh on our idea of self-hosting video from our media library. Can you imagine downloading on a remote phone something as big as a gigabyte? Crazy.
Google can’t measure user experience. Not really. Can they measure happy or sad or disgust? Nope. But they have a lot of people fooled into thinking they *can.* Even if they could see your face, they would still have to guess. Google can’t judge quality of content. They guess.
Google predicts 200 “precision” secondary signals. They make assumptions what those mean. This includes things like:
Do these signals show website goodness or quality?
Maybe they do.
There is some truth in what they’re doing, but not a lot. Even real-life people judging every website would still be inaccurate – and impossible. All the “signals” are prone to cheating and mismeasuring. It’s ridiculous.
People make a lot of assumptions about what matters and what doesn’t. For example, many think speed is critical.
Sorry. Relevant content is much bigger than speed. Anything is bigger than speed. Speed is a pimple on the Google algorithm. Less than a half-percent improvement. Big wow. It’s amazing how many people believe that speed is a huge ranking factor. It’s not, as we’ve said many times.
Doesn’t Google use PageSpeed Insights scores in search ranking? No. That test’s an ivory-tower exercise to drive site creators crazy. WordPress sites can’t pass that test. Not even Google’s homepage passes that test. Try it sometime.
Is speed anxiety serious for e-commerce? Yes. Big anxiety. Why? Google whipped them up in a frothing frenzy. Speed propaganda says fiddling makes a difference. Amazing.
We often open an article in a tab and then proceed to read something in another tab. That dormant open tab counts as dwell-time in Google’s book. Google (supposedly) counts that dwell-time for up to 30 minutes. Is that indicative someone is reading the content? We may never return to read the article – and close the tab.
This fake-out creating false data bothers mighty Google. It’s the motivation for more spying via Chrome browsers to gather data on user behaviors. Every time we use a Google product, we consent to them gathering data.
Google uses artificial intelligence called RankBrain to predict user intent. This means filtered information by an “unbiased” machine before presenting search result choices. We don’t filter anymore. Google does. Interesting. How does that taint the data? Isn’t data then analyzing itself? Too weird.
What’s to prevent Google giving preferential treatment to big advertising clients? They wouldn’t do that, would they? Well, yes. An example: When it came time to roll out the mobile-first page ranking algorithm, did they test on big accounts that brought them profits? No. They quietly excluded them. Instead, they experimented on millions of little sites as no-risk collateral damage. They weren’t about to test on any big fish. Profit first.
Hmm? So Google does game their own system.
Everyone says Facebook and Twitter over-control what people see on the Internet. But they never realize Google is the one who controls most:
What website pops up during a search.
What Google “thinks” we’re looking for.
Don’t believe everything you read or hear about the relationship of SEO, UX, and mobile speed. Especially if Google said it.
LiteSpeed Cache for WordPress is a site acceleration plugin. It features a server-level cache and a collection of optimization features.
LiteSpeed claims advantages over common Apache servers. Here’s why we’re doubtful.
1It is up to six times faster than Apache. And 12 times faster than NGINX.
Does this make a difference in page load time? Not really. Real culprits slowing down websites are:
2It is three times faster than Apache in SSL.
Would we argue that a stereo performing above the range of human hearing is better? Braggadocio.
Does that mean it’s faster than not using SSL? We doubt it. But we’re being snarky. It’s a metric. But is it more significant than loading bad-choice images? Such as large PNG photographs on your site? Of course, not. It’s insignificant by comparison.
3With its LiteMage cache, LiteSpeed makes Magento pages run up to 75 times faster.
Wow! That sounds really good. But, what does that mean for WordPress users? Uh? Nothing. Magento is an e-commerce platform built on open-source technology. It’s a content-management-system alternative for WordPress and WooCommerce. Big deal. It’s not even part of the WordPress world.
4It increases PHP performance by 50 percent.
Is PHP performance – or lack thereof – significant? Not from our tests. Immeasurable.
So let’s talk about LiteSpeed benchmarks? What are they saying:
We tested HTTP/2 implementations from LiteSpeed Web Server, Nginx, and Apache, to see how they would compare when loading WordPress. LiteSpeed beat Nginx by up to 12X, and blew Apache out of the water by a whopping 84X! The best available WordPress cache plugins were used for each server: LSCache for LiteSpeed, FastCGI Cache for Nginx, and W3 Total Cache for Apache. – REFERENCE
They later mention those specs are on HTTP/2 servers. And they weren’t using cheap shared hosting. They used Vultr Cloud VM. That service costs extra money based on usage. Vultr’s cheapest cloud hosting plan is $5 per month.
What are they not comparing? They’re not comparing milliseconds of load time for a website. They’re talking about server processing requests per second. Not the same thing at all.
On common affordable everyday HTTP 1.1 servers, the results aren’t as shiny:
LiteSpeed Web Server performs 5X faster than Nginx and 28X faster than Apache when loading WordPress.
These specifications mimic what developers claimed about PHP version 7 when compared to PHP version 5. WordPress runs on PHP server-side language. Did that version update shave seconds or milliseconds off of WordPress page load times? No. It only increased server processing speed. The benefit wasn’t even measurable with standard speed tests. Hand-waving specsmanship. Pure hype.
Are these processing metrics a valid indicator of potential website speed improvement? Not for a minute.
That’s right more plugins are better than less in the speed department. Only when using lightweight discreet plugins. One fat plugin like Yoast SEO – 150 to 250 milliseconds – negates all gains from discrete plugins.
Caching is a band-aid for inferior website design. Any page using origin-optimization strategy always loads faster. The complexity of many settings means the plugin isn’t dummy-proof.
Plug-and-play discrete plugins are more bulletproof. Caching often breaks site functions. As does minification (file concatenation). Because of fragility caused by complexity, we don’t always use a caching plugin. It’s herd mentality to always install a caching plugin. Especially popular ones like W3 Total Cache (1+ million active) and WP Super Cache (2+ million active). It’s better if you don’t use caching if you don’t get a significant gain. And we rarely do. Not with an optimized site built for speed.
2Using discrete plugins for individual features is faster. These plugins usually load in under 1 millisecond. LiteSpeed Cache adds 53 milliseconds of load time globally. That is on every page and post of your website. We call that site drag. We can install and activate over 53 discrete plugins in that same load time.
3The question-and-answer section of the LiteSpeed Cache website recommends third-party CDN workarounds. This is bad practice for speed. Again, it’s an after-the-fact repair band-aid. It doesn’t “fix” the sloppy site-origin problems. It masks the causes of real bloat. Apathy is the true cause of bloat.
4Using discrete plugins allows us to selectively activate or deactivate plugin functions on a page-by-page basis. This is always useful, especially on an e-commerce site. The store pages are dynamic. They often won’t work with caching plugins – or other plugins used for minification.
According to LiteSpeed developers, a single LiteSpeed server is capable of handling data equivalent to two Apache servers. We dislike LiteSpeed’s claim of “80 times faster.” Ridiculous vanity metric. Server execution time doesn’t translate into milliseconds of gain on a website. You can erroneously imply the same claim for PHP7 over PHP5. But change in server PHP versions doesn’t make a dent in poorly optimized images, ads, fonts, etc.
If LiteSpeed is free from your host, go for it. Use it. If you’re paying extra for it, think about what value you get and if it contributes to site profitability.
NOTE: Lightspeed was aggressively added to one of our sites by the hosts. That means they didn’t ask permission, they just added the Lightspeed plugin. That broke Easy Digital Downloads plugin checkout functions. Instant no sales. The solution? Because we knew they would just reinstall the plugin if we removed it, we used selective deactivation to turn off the Lightspeed Plugin on the checkout page. But that didn’t work. WE had to have the host deactivate LiteSpeed completely. There are workarounds with LiteSpeed for ecommerce.
LiteSpeed and eCommerce problems
Litespeed cache causes the WooCommerce shopping cart to be updated incorrectly. LiteSpeed requires a configuration tweak.
Is WooCommerce supported?
For some WooCommerce themes, the cart may not be updated correctly. LiteSpeed blog has a tutorial on how to detect this problem and fix it if necessary.
Do Not Cache URIs – used to exclude pages from cache. (List pages that have contact forms, logged-in pages, or any checkouts. Although WooCommerce checkout is already excluded by default.)
It’s reported: WooCommerce doesn’t want something cached (like, for instance, the Cart, My Account, and Checkout), LiteSpeed Cache will automatically respect that, and will not cache those pages. There’s no extra configuration required. REFERENCE: https://www.buyindemand.com/litespeed-cache-and-woocommerce/
But LiteSpeed doesn’t recognize Easy Digital Downloads plugin transaction pages.
You would not need AMP tags for a normal website. You can have a very fast website without using AMP separate markup. – REFERENCE
Google’s ambitious AMP project (October 2015) speeds up the entire mobile web. Right!? It’s a WordPress speed miracle – wait! No it’s not. It’s another Google sham. It promises better SEO results in exchange for people surrendering control and information. Google’s great boon for mobile publishing fizzled.
Three examples of Google AMP hijacking URLs and valuable screen space:
Google, to speed up AMP, stores publisher’s pages and serves from Google. When a reader clicks an AMP link, the address bar displays www.google.com instead of the publisher’s Web address. Oops? Where’d that site branding go?
Google says articles served from its Internet network is four times faster. That’s right. But developers disagree. Engineering or “designing in” origin optimization gives the same or even better results. That’s right. It appears Google has a hidden agenda not-so related to speed. That’s their mask.
What do these wonderful plugins claim?
AMP By Automattic Active installs: 500,000+ zip file size: 1.5M
The Big Promise: AMP makes your website faster for mobile visitors. SEO pros claim Google prioritizes AMPs in search results. Feb. 24, 2016: Google integrated AMP listings into mobile search results. Big deal.
Really? But when we search on the phrase:
“Why Google AMP sucks”
We’re inundated with blog posts by developers. They say unexpected derogatory words about how AMP doesn’t help. They say it’s all a ploy by Google. These aren’t crackpots chiming in here. There are tech heavyweights who think AMP is bad for the free web. Why?
How to Setup Google AMP on WordPress Site (Using AMP Plugin) When you read the above post the propaganda sounds wonderful. Google AMP is the magic silver bullet curing all mobile speed problems. And there are two WordPress plugins helpers to get you set up. One authored and endorsed by Automattic, the founders and owners of WordPress. What an amazing endorsement! (Question: What is Automattic doing in bed with Google?)
AMP hijacks the real URL of the page and provides no UI control to get back to the canonical URL. Consumers, publishers, sites and users are asking Google to give an opt out feature. AMP often loads broken links, takes longer to view, or may not format properly
When a user searches for an article on Google and clicks on an AMP link, it never leaves Google.com. Even if the AMP link is an external entity. Google rehosts pages from popular search destinations. But they truncate it and reformat it. It’s supposed to be helpful because it lets complex pages load faster, but it’s a pain because it frequently doesn’t format properly or they cut it off in an inconvenient place. You’ve got to waste time getting to the original page.
There’s no way to turn AMP off. If they’ve got an AMP mirror for a domain, they’ll give you those results instead of the actual page every time.
Smaller publishers have more to lose if they use AMP. For many, they think Google owns the small-guys story.
John Gruber at Daring Fireball says the following:
The lock-in aspect makes no sense to me. Why would I want to cede control over my pages to Google? AMP pages do load fast, but if publishers want their web pages to load fast, they can just engineer them to load fast. Best answers I got were that it wasn’t really strategic — publishers are going with AMP just because their SEO people are telling them to, because Google features AMP pages in search results. I suppose that is a strategy, but ceding control over your content to Google isn’t a good one in the long term.
Wikipedia negative quotes about AMP (there aren’t any positive ones):
Some tech media outlets, including The Register have criticized AMP:
Chris Coyier, cofounder of CodePen, writes that AMP would be better if it “was never used as search ranking factor (or anything that could be interpreted as such) by anybody.”
At AMP Conference, Gina Trapani described some aspects of AMP as being “scary to me.”
I’m on the record as being strongly opposed to AMP simply on the grounds of publication independence. I’d stand by that even if the implementation were great. But the implementation is not great — it’s terrible. Yes, AMP pages load fast, but you don’t need AMP for fast-loading web pages. […] It’s a deliberate effort by Google to break the open web.
Why is PagePipe against Google AMP?
Here’s a good offsite article about how Google AMP failed for Kinsta by actually impacting their SEO by a negative 59 percent drop in leads. Read it in a new tab.
Speed is a strategy not a band-aid. You build site and page quality in – not sprinkle it on afterwards. AMP serves Google’s purposes – not ordinary users. We’re surprised an AMP content blocker hasn’t shown up by now. Improving the web is a better answer than donating it to Google. AMP is for apathetic site creators. We can fix the problem without handing everything over to Google.
LONG but GOOD EXCERPT from: PixelEnvy Written by Nick Heer. “Given the assumption that any additional bandwidth offered to web developers will immediately be consumed, there seems to be just one possible solution, which is to reduce the amount of bytes that are transmitted. For some bizarre reason, this hasn’t happened on the main web, because it somehow makes more sense to create an exact copy of every page on their site that is expressly designed for speed. Welcome back, WAP — except, for some reason, this mobile-centric copy is entirely dependent on yet more bytes. This is the dumbfoundingly dumb premise of AMP.
That belies the reason AMP has taken off. It isn’t necessarily because AMP pages are better for users, though that’s absolutely a consideration, but because Google wants it to be popular. When you search Google for anything remotely related to current events, you’ll see only AMP pages in the news carousel that sits above typical search results. You’ll also see AMP links crowding the first results page, too. Google has openly admitted that they promote AMP pages in their results and that the carousel is restricted to only AMP links on their mobile results page. They insist that this is because AMP pages are faster and, therefore, better for users, but that’s not a complete explanation for three reasons: AMP pages aren’t inherently faster than non-AMP pages, high-performing non-AMP pages are not mixed with AMP versions, and Google has a conflict of interest in promoting the format.
It seems ridiculous to argue that AMP pages aren’t actually faster than their plain HTML counterparts because it’s so easy to see these pages are actually very fast. And there’s a good reason for that. It isn’t that there’s some sort of special sauce that is being done with the AMP format, or some brilliant piece of programmatic rearchitecting. No, it’s just because AMP restricts the kinds of elements that can be used on a page and severely limits the scripts that can be used. That means that webpages can’t be littered with arbitrary and numerous tracking and advertiser scripts, and that, of course, leads to a dramatically faster page. A series of experiments by Tim Kadlec showed the effect of these limitations:
AMP’s biggest advantage isn’t the library — you can beat that on your own. It isn’t the AMP cache — you can get many of those optimizations through a good build script, and all of them through a decent CDN provider. That’s not to say there aren’t some really smart things happening in the AMP JS library or the cache — there are. It’s just not what makes the biggest difference from a performance perspective.
AMP’s biggest advantage is the restrictions it draws on how much stuff you can cram into a single page.
AMP’s restrictions mean less stuff. It’s a concession publishers are willing to make in exchange for the enhanced distribution Google provides, but that they hesitate to make for their canonical versions.
So: if you have a reasonably fast host and don’t litter your page with scripts, you, too, can have AMP-like results without creating a copy of your site dependent on Google and their slow crawl to gain control over the infrastructure of the web. But you can’t get into Google’s special promoted slots for AMP websites for reasons that are almost certainly driven by self-interest.”
The AMP version of my website is loading very slow. I checked it with the Google Speed Insights and it says that the server response time is 3 – 4 seconds for the AMP Pages. For the Desktop version of the site, the server response time is 0.3 seconds, so it’s not a hosting problem.
People love to argue about which hosting is the best in the world. All hosts are pretty lame at some time. They’re for-profit companies run by fallible humans.
We apologize in advance for speaking ill of a host you may care a lot about. Can you tolerate trash-talk about your special-preferred hosting company? If not, you’re gonna be resentful. It’s a loyalty and pride issue for some team-type people. Hosting is a commodity product. We don’t get emotional about web hosts. But at the same time, we ‘hate” slow web pages. That sounds pretty emotional.
SLOW web hosting SUCKS
Let’s be upfront about our topic. We’re discussing common ordinary cheap shared commodity hosting. Not expensive specialty hosting like VPNs. And huge cloud-based services charging by the minute – or by byte volume.
A frequent question we’re asked is,
“What host do you recommend for speed?”
People are often surprised at our response, “We don’t recommend hosting.” Of course, you need a web host to build a website. But we don’t recommend hosts? Why? They’re cyclical from mediocre to bad to worse. That’s been the history.
If site owners asked, “What host should I avoid?”
Many diseased hosts — they’d sell their grandmother for $100.
It’s much easier to answer. We could give a long list. There are so many hosting problems people don’t know about. But there are easy ways to find out how good-is-good-enough.
A speed evaluation we do is measuring Time To First Byte (TTFB). There are three ways to determine TTFB. Two are using online tests and another is by uploading an HTML file into your media library.
We’ve talked on PagePipe about why TTFB is important. But we’ll review it here.
Consider TTFB the server overhead. It’s a delay in milliseconds. It’s how long it takes for the server to respond to a request for web assets. Then the browser can begin to construct your WordPress page in the device viewport. A good TTFB is 100 to 300 milliseconds. Ordinary is around 500 milliseconds, and poor is 750 milliseconds. Dismal is 1 second. And anything after 1.5 seconds is horrific.
How bad does TTFB delay get?
Worst case we’ve seen TTFB as long as minutes. This slowdown is often caused by plugins hammering on the server database. Or a plugin and theme conflict confusing the server. So it isn’t always the server’s fault. Plugins with heavy zip downloads are the most notorious for causing TTFB delays. They’re complicated plugins with lots of code – and lots of repeated calls to the server. That is bad for speed.
Hosts with impeccable service and wonderful uptime become bandits about their speed benefits. They brag about the quality of their magnificent servers. Boasting about superior SSD (solid-state drives) instead of mechanical spinner magnetic drives. As if these specifications matter for site owners. We’ve never seen the type of drive technology make any difference. Yes. SSD drives are faster. But in the end with all the other hosting variables, it isn’t significant. It’s not a measurable difference. The gain gets lost in the noise.
Who benefits most from SSD drives is the host company. SSD drives are quiet, smaller, cooler, and reduce energy consumption. They are good at reducing the server-facility overhead in floor-space and energy bills. But that has nothing to do with actual speed – or website load time. That is specsmanship hosts boast about. But the boasting never proves anything. Like bragging that your car can go zero to 60 in 3 seconds – but you use it for shopping at Walmart or driving in school zones. A waste of machine.
Because a host claims its servers are fast doesn’t mean you get fast. They exaggerate and omit details. They show perfect samples.
There are simple tests to find out if they’re deceptive. Here’s how:
Get the URL of the hosting company homepage. Plug that into ByteCheck.com. You will never get a faster TTFB on their servers than they present on their homepage. That is simple. Compare them. Take 6 consecutive tests in a row. Are you surprised to find your favorite host gets 200-millisecond TTFB — but only once per hour? The rest of the time it’s fluctuating around 1.7-seconds TTFB. That is terrible. What host would be that bad? Er. SiteGround. Yeah. They suck for speed and there are reasons. Popular hosts don’t guarantee good speed. Neither do expense hosts Like WP Engine, $100 per month: $1,150 per year. Ouch.
So what about places like BlueHost ($71.40 annual rent)? We wanted to prove a point once. Even on super-cheap hosting like BlueHost we could load an eCommerce store in under 2 seconds. What kind of TTFB did our store get on BlueHost? Typically there was a 1.7 second delay.
That meant we had to load our store pages in 300 milliseconds. We made it work but we had no room for error. Those tight tolerances made us uncomfortable. The load time would push out beyond 2 seconds to 2.2 to 2.5 seconds. We wanted to maintain our reputation as speed freaks. So we moved the store to Rochen later ($227.40 annual rent fee). We proved our point, but we didn’t want to stay there any longer than necessary.
So is BlueHost so bad? it all depends upon what your goals are. To run a simple blog, it’s fine. All hosts are bad and good at different times. What we like is hosts with rock-stable servers. We prefer a predictable 500-millisecond TTFB at GoDaddy. Rather than a fluctuating TTFB at SiteGround ($300 annual rent). When we work with speed clients, do we move them off SiteGround? Always. Do the clients whine about that? Yes. They do. They think SiteGround gives them good service. When we call their service desk, we get a correct answer half the time. The IRS 800 helpline gives wrong tax advice answers half the time, too. Is that good enough? Heck, no.
But what we despise most is hosting companies with speed claims that are flat out lies. They give no proof. They start throwing around technobabble jargon. It’s a smokescreen that they out spec the competition.
We do not appreciate these tactics. We aren’t ignorant.
When you find a good host, will they stay good forever? Not for long. For example, we used to love Pressidium managed WordPress hosting ($1,798.8 annual rent). But they had some policy changes and now TTFB speed – that used to be spectacular – isn’t so good. And they locked us away from writing code to the server htaccess file with plugins.
The htaccess file is important if you do selective activation of plugins. A cool speed trick. That ruins speed Karma for us. Most managed hosting tries to keep the client as far away from server access as possible. Why? They say “security issues.” We know the real reason. Services costs rise as they get more clientele and they can’t keep up with the ensuing chaos. So they start locking people out to reduce the service calls.
What about Kinsta ($2,400 annual rent)? Aren’t they good? Their mantra is “Built for Speed.” Yeah, except they give out bad advice for speed. They don’t speak of cheaper alternatives. They perpetuate myths to their advantage making their specsmanship appear dandy. Propaganda machines. We criticize their claims and advice more than we do their actual performance. We feel it’s deceptive and shallow.
We’ve written quite a bit about Kinsta policies and advice. So we won’t repeat the potential insinuations and insults here.
We want a host with integrity. It’s often fine until a hosting company changes ownership – or gets popular. That’s right. Popularity ruins hosts. For example, HostGator, we were checking why a New York seller of woman’s perfume had such a slow site. We used a tool called YouGetSignal. That online tool told us it could only show the first 1,000 domains. Our client was sharing the server with at least 1,000 more domains. That was the problem. The server was crammed to the gills with over 2,000 domains. What is normal? Well under 100 domains – and more like a dozen.
Mystified speed clients ask why their host server is so extraordinarily slow. We check with YouGetSignal, and bingo, there’s an adult porn site lurking unknown on their server. Kiss speed goodbye if that’s the case. You don’t need an advanced degree to use these online tools. It’s a simple copy-and-paste of your URL. The tool then identifies in RED the offending adult site. Sharing a server isn’t bad – if you have the right neighbors.
Generally, you get better speed by throwing money at the problem. Or renting more expensive servers. We see this on sites so corrupted we can’t update to a current version of WordPress. Instead of rebuilding the site and fixing it, the lazy owners buy more expensive hosting to solve it. The site is a hand-grenade with the pin pulled out. But they don’t want to fix it.
Where do site owners go to host a damaged site? Usually Kinsta. Got a lousy site. Go to Kinsta. It fixes speed with money. Shortsighted. That site is ticking like a time bomb.
30 seconds or even 1 minute load times, inevitably it’s a huge plugin fighting to get control of the server database. Usually a security plugin, a caching plugin, or metric-gathering plugin. These plugins usually weigh more than WordPress itself. But – hey – they are popular. Did anyone ever think to look at the size of the popular plugin? And wonder if that monstrosity is detrimental? Obviously not. It’s installed because everyone (The Herd) is installing it. It must be good. Right?
After pulling these fat plugins, the server miraculously calms down to a fast load time. What good is a lame plugin like that? Your bounce rate goes through the roof. These kinds of sites are the easiest tune-ups for us to fix and create the most dramatic results. We’re heroes. When you go from a 30-second load time to under 1 second, it impresses the client. They think you performed an exorcism when all you did was disconnect a power hog plugin.
WP ENGINE is not a preferred host for speed. Their server overhead (TTFB) is often about 800 milliseconds. A good TTFB (time to first byte) is 100 to 200 milliseconds. The average is 500 milliseconds. Poor is 750 milliseconds and bad is over 1 second. Hosts like SiteGround have erratic TTFB. It sometimes is 200 milliseconds – but most often is 1.7 seconds. So they claim the cherry-picked specification and ignore the real speed errors.
WP ENGINE’s homepage TTFB is 107 milliseconds tested with ByteCheck. Impressive. That would fool our usual rule of checking the host’s homepage as an indicator.
But 6 consecutive test of a client site on WP ENGINE shows:
WP Engine is using a sweet server for their homepage hosting – but not yours. You aren’t sharing your server with anyone. It should be fast.
Their speed secret: they are using Fastly CDN and Cloudflare CDNs on their page but not yours. Fastly services charge based on traffic and bandwidth usage. And Cloudflare is $200 per month estimated.
In other words, they throw money at their homepage speed but not yours. A deception.
WordPress.org Official Recommended Web Hosting GLOWING ENDORSEMENTS
There are hundreds of thousands of web hosts on the internet. WordPress recommends only three hosts: Bluehost, DreamHost, and SiteGround. Those companies “donate” a part of your hosting fee back to WordPress. That’s called an affiliate link anywhere else. Not a donation.
Listing is completely arbitrary but includes: “contributions” to WordPress.org – as their first criteria.
How much do those 3 companies pay WordPress?
They won’t tell you. But you can’t get listed unless you “contribute.” The WordPress endorsement is worth millions of dollars to these three companies. Uh? That’s more blackmail or hostage payments.
“The recommended webhosting page on WordPress.org is incredibly lucrative. Based on conversations I’ve had with employees of hosts listed, it can generate millions of dollars in revenue.”
Why are the top-recommended hosts in a Google search the worst hosts in the eyes of real users? Why can’t you trust WordPress hosting recommendations?
“Most hosting recommendations are unreliable for a simple reason: money. Like many other things, money corrupts hosting conversations. Recommending bad hosting leads to large amounts of money for the recommender. How does that happen? It’s the reality of the affiliate marketing model that dominates the hosting space.”
Here are some facts about WordPress.org recommended hosts:
We’ve mentioned BlueHost to prove we could make a store work even on poor-quality hosting. The TTFB for our site was 1.7 seconds. We loaded pages in under 300 milliseconds. Then things got worse and we moved.
Bluehost received a wooden spoon award for being the bottom feeder in a review of web hosts. Winning a wooden spoon doesn’t sound like a very awesome prize for a high-tech company.
BlueHost is owned by Endurance International Group who owns the 20 largest web hosts. They have venture-capital ownership in Automattic, the mothership of WordPress. EIG is owned by Clearlake Capital Group L.P. a diversified investment firm.
We’ve never used DreamHost ($203.40 annual rent for one domain). We never were tempted. But here’s the scoop on their homepage: TTFB: 195 milliseconds. Hey! That’s pretty good. Load time: 5.541 seconds. Hey! That’s pretty bad with such a good TTFB.
So why are they so slow? Maybe it’s because they’re sharing their server. That would be walking the talk. Nope? They share their server with no one. So why so slow?
Here’s why: remote third-party services located on distant servers.
Amazon CloudFront fonts.googleapis.com : Google widget.trustpilot.com : Amazon CloudFront www.google-analytics.com : Google connect.facebook.net : Facebook googleads.g.doubleclick.net : Google
HotJar, a user experience metric service, adds 500 milliseconds.
You get the idea. These guys need to do their speed homework.
We’ve already mentioned SiteGround as having fluctuating TTFB. Every client we work with moves if they are on this host. Do we make them move? No. They choose to after they see the speed reports. Do we tell them where to host? No. They choose. We don’t care where they go as long as they move away from this host. Then we can achieve our speed goals. The clients always say, “But everyone says they are fast.” Who is everyone?
“Everyone” is bloggers with affiliate links to SiteGround. Of course [forehead smack].
But … But … what about all the impartial reviews on internet blogs?
Check those referral links again in the “impartial” reviews. They’re affiliate links.
When an affiliate recommends a product to you and you buy it, the affiliate gets a payment. Someone paid for an opinion isn’t a great judge of truth.
$50 for 1 to 5 referrals per month, $75 for 6 to 10, $100 for 11 to 10, and $125 for 21+ sales. (This is the affiliate structure of SiteGround, one of our least favorite hosts.)
So, most WordPress hosting advice is dishonest. How do you find honest, real information?
You can’t. You have to test.
Do some of the simple tests we recommend in this ebook. Check their homepage and gimmicks by running a speed test on WebPageTest.org. Check their TTFB on ByteCheck.com. See who shares their server with them YouGetSignal.com. Look them up on Wikipedia and see who owns the host company.
Can we trust PagePipe?
We won’t answer that. Do testing for yourself. Do not trust any reviews or testimonials. Even ours.
Ask yourself, “How good is good enough?” Don’t waste money.
Here’s a full and clear disclaimer: There’s one affiliate link in our hosting report. It’s not on this page – or in our blog. We recommend a host and we collect an affiliate payment if you buy from them. Not if you click the link. Only if you decide to buy.
Who is that host?
You go there and buy and we get a kickback. The link is in our 1.5MB downloadable PDF. Do we deserve it? Or will you punish us for being blatant and honest? They don’t charge you. You pay the same price either way. It’s your way of saying “thank you” and giving us some applause. Do we need applause? Absolutely.
If you abhor GreenGeeks, feel free to write and tell us why. Email us.
After all these years of restraint, why endorse a hosting company now?
Ironically: The reason is integrity.
GreenGeeks doesn’t brag so much about speed. They could.
We moved from GoDaddy to them last year. We still use GoDaddy servers for testing and have an account with them. We moved because readers were asking if we’d like to have our SSL fixed. You know that little shield in the corner of your browser address field that promises you’re “secure.” They thought we missed that. They didn’t realize it was a deliberate choice for speed. SSL is used to slow sites by 500 milliseconds. But hosts have found ways to speed that extra burden up.
PagePipe blog now has a Time To First Byte of 167 milliseconds using ByteCheck test. GoDaddy was typically 500-milliseconds TTFB or worse. Our blog loaded in under two seconds. But GoDaddy charged $70 per year per domain for SSL/HTTPS certification. Other hosts as GreenGeeks offer that for free.
Our load time on PagePipe’s homepage is demonstrated in this test:
How much does it cost for shared hosting on GreekGeeks versus GoDaddy?
We chose the Pro plan at $4.95 per month for the first year and $15.95 per month after that. Their features include:
Unlimited Web Space Unmetered Data Transfer Free SSL Certificate Free Domain Name for 1st Year Free Nightly Backup Free CDN Unlimited E-mail Accounts WordPress Installer/Updates Unlimited Databases 2x Performance LSCache Included 300% Green Energy Match 30-Day Money-Back Guarantee
LSCache is LiteSpeed server caching.
LiteSpeed helps performance.We’ve written a snarky article about our experience here:
The downside is disabling your eCommerce WooCommerce or Easy Digital Downloads cart. There are special settings to disallow those pages. But we lost sales messing with the setting figuring things out. We got over it. Today, we are fans of LiteSpeed.
So how much is GoDaddy hosting?
For unlimited websites (Deluxe plan), it’s $4.99 per month the first year, and $8.99 after.
So after pricing settles in the second year, the annual difference is GoDaddy $107.88 US dollars. And GreenGeeks, $191.40.
But if we added the cheapest SSL fee +$69.99 per year, GoDaddy would be $177.87. And if we were registering a new domain that would add $17.99. Or $195.86.
The difference: $5.54 less for GreenGeeks. Does price tip the scale in GreenGeeks favor? Not really. It’s sixes. About the same cost.
So what does tip the scale to GreenGeeks favor:
GoDaddy doesn’t offer LiteSpeed server caching. That makes a big difference in performance for even our origin optimized websites. These speed sitesdon’t benefit from caching plugins because they’re so dang fast loading. LiteSpeed Web Server (LSWS) is a high-performance Apache drop-in replacement.GoDaddy uses Apache servers. Is this another useless non-reality-based engineering specmanship? No. LiteSpeed makes a real difference. A web page that loads in 2-seconds now load in subseconds with LiteSpeed.
Big deal who cares? Faster than fast? So what?
Mobile users care. Some site owners get 80-percent mobile traffic, it’s a big deal.
LiteSpeed incorporates selectable speed features we desire. Weadd free discreet plugins to strip WordPress non-features and make things faster. You don’t need those extra plugins. It’s unnecessary. The functions now reside on the LiteSpeed server. Much faster and efficient.
Adding LiteSpeed plugin is required. It causes 53 milliseconds of global site drag. But the loss is worth it – because it accelerates everything else.
But there’s another reason to use GreenGeeks. We admire their idealistic integrity. Even if it’s only a marketing differentiation ploy, we like it.
GREEN ENERGY HOSTING SERVICE
GreenGeeks started in 2008. They committed to being the most Eco-friendly green web hosting company in the World.
GreenGeeks is recognized by the United States Environmental Protection Agency since 2009 as a Green Power Partner.
GreenGeeks work with the Bonneville Environmental Foundation (BEF) in Portland, Oregon. “BEF” is a Green-e Partner.
GreenGeeks tell BEF how many servers, personnel, etc. they have. They calculate the yearly energy consumption and carbon footprint. BEF purchases 3 times what GreenGeeks consumes. They put that energy back into the grid.
They match the energy they consume as well as payback for 2 other companies their size. This is their commitment to the environment since the beginning.
GreenGeeks has an A+ rating with the Better Business Bureau.
They walk the talk.
We admire their idealistic values. We are idealists, too. We call it honest responsibility.
Endorsing GreenGeeks is mutually beneficial for our credibility.
People binge-read our entire blog. We are the “consumer report” of speed.
GreenGeeks is our only affiliate link. We don’t publishing that link here in our blog. It’s not on our website. You have to download our report, “SLOW web hosting SUCKS” to get it.
We’ve written about GreenGeeks competitors. But those bandits get zero links – only GreenGeeks get a link. It would be normal on an affiliate link farm to have links to even the hosting losers. Desperation to snag extra income. We’re not sending you there.
We move our clientele over to your GreenGeek servers as long as the TTFB stays fast. Our reputation as speed experts depends upon us recommending good solutions.
GreenGeeks is “the best shared-hosting deal.” GreenGeeks is our only hosting recommendation. We’re finished dating hosts. We want a long-term relationship with these guys. We’re engaged and hope to marry.
Domain names are purchased from a domain registrar, which is an accredited company that allows you to buy and register domain names.
Early in our online career, it just made sense to buy our domain name and hosting from the same company, so we did from Namecheap.
We found Namecheap to be an excellent domain registrar with reasonable prices. The customer service was 24/7 live chat.
They were nice and helpful.
As for the hosting, we were new. We had no understanding about or expectation of site speed or even what a web host’s ultimate responsibility is to a website owner.
But at the end of the first year, we decided to leave Namecheap web hosting in search of a “better” web host.
And do you want to know why?!
Well, we were running a WordPress website. And all of this website stuff was foreign to us and we needed a lot of hand holding — we mean support — on all fronts.
And at the time, Namecheap’s hosting was a new offer for them and their WordPress support was just less than stellar. So we hit a wall. We needed more and they could not deliver at the time.
So after the first year, we decided to keep them as our domain registrar and move on to a “better” host.
Lesson: Choosing a web host needs to be a thoughtful decision. Don’t be lazy and take the easiest path.
InMotion Hosting: Were They “Better?!”
After doing much research, we decided that the “better” host for us was: InMotion Hosting.
They “claimed” to be WordPress experts.
They had not only 24/7 live chat, but also 24/7 phone support.
For a still relatively new owner of a WordPress website, 24/7 phone support was sweet music to our ears.
And at first, all was well. We were happy. And boy did we use that 24/7 phone support.
But can you guess what happened?!
Well, it came to an abrupt end when John, a InMotion Hosting customer service rep, rudely informed us that helping us with our WordPress questions was, well, not their concern.
It was his insulting tone.
We felt stupid and we were upset enough to call InMotion Hosting back and ask to speak with a supervisor. To the supervisor’s credit, he apologized for John’s poor customer service. And then, said, “There are always going to be customer service reps that aren’t quality.”
He’s right, of course. We understand that, but it did not make us feel confident or “better.”
So before the year was even up, we were searching for a “better” web host again.
Lesson: Customer service can make or break a web host.
SiteGround: But Aren’t They the “Best?!”
So after we left InMotion Hosting, we landed at SiteGround.
They offered web hosting at a very sweet deal of just $70 a year for the first year — actually, you could lock in this sweet deal if you could afford to pay for three years up front. Truth is, most beginner website owners don’t have a clue where they will be in three years, so that’s an unrealistic commitment.
We did have the good sense to ask about what the pricing would be after the first year. And we were told, “Not to worry. There are always coupon discounts.”
Plus, they had 24/7 ticket support, 24/7 live chat, and 24/7 hour phone support. All this potential hand holding was a dream.
Plus, everyone – everyone – loves SiteGround.
Plus, everyone online shared how super fast they were. Everyone.
Honestly, they seemed like a web host made in heaven.
The first year ended and our web hosting was up for renewal. This is when we got the sticker shock.
Coupon discounts?! That was a total lie. (To be fair, nine out of 10 web hosts give you a sweet first year deal.)
Our annual budget for web hosting went from $70 to $240. That’s a lot.
But we were still clueless (and lazy) about how one chooses a good web host.
Because we had absolutely no idea how you even begin to choose a new host after you’ve been with the “best.”
So we stuck around for another year, paying their inflated web hosting prices of $59.85 every three months. That’s $19.95 a month for their Grow Big hosting plan.
This all came to a screeching halt, when they jacked up the price to 89.85 every three months. Now we are up to $29.95 a month. Interestingly enough, their website says $24.95/month for the Grow Big plan, so why did charge us $29.95?
But here’s what added insult to injury…
SiteGround’s proprietary Client Area and Site Tools are provided to new customers, while many of their old customers like us were still stuck using cPanel, the standard web based control panel many web hosting companies use.
We ask you this…
How fair is it to jack up the price from an already over inflated $19.95 per month to a ridiculous $29.95?!
And how fair is it that this price gouge happens to old, loyal customers, who SiteGround still haven’t migrated to their new Client Area and Site Tools a year after it was first rolled out?
Our wallets could no longer handle the steal and our self-respect made us get busy looking for a new web host again.
SiteGround can’t be the only “great” web host, right?!
There are two things that SiteGround is known for: Speed and Customer Service.
On the subject of speed, here is what SiteGround promises: “Our hosting platform is built on Google Cloud and uses its ultra-fast network and SSD persistent storage.”
Sounds nice, right?!
And along with caching implementation on their servers, they also provide a proprietary caching plugin called SG Optimizer.
But we were not convinced that SG Optimizer provided any meaningful speed boost.
Plus, like most caching plugins, it is not an easy plugin to configure. And it’s implementation on our website was never a smooth experience. So we deactivated and deleted it.
In fact, we go so far as to say that a lot of SiteGround’s server caching optimization tools were disconnected from each other — at least for the average website owner.
We would accidentally discover yet another server-side caching option that required manually flipping a switch even after talking extensively with a customer service rep about website optimization.
We think that if you are a web host that prides itself on being super fast, shouldn’t your server speed optimization tools be cohesive and intuitive for the average website owner?!
Fair question, don’t you think?
And on the subject of customer service, here’s what SiteGround has to say: “Our Customer Care team is among the highest-rated support squads online, fast, multi-skilled and helpful.”
Here’s what our experience has been…
Of all the web hosts we had tried to date, SiteGround was the most proficient at providing support on WordPress itself.
And, honestly, they were really nice and usually super helpful. They would actually offer to complete many site changes on our behalf.
But towards the end, the accuracy of the support was sometimes suspect and, at $29.95 a month, that seems unforgivable.
We think you deserve consistent, accurate information from web hosting customer service reps, especially at SiteGround’s price point. Don’t you?!
Lesson: Don’t believe the hype. Dig a little deeper and test for yourself.
Tip #3: You want a web host that delivers 99.9 to 100% uptime. Because a down website can reduce your bottom line.
Tip #4: You want a web host that allows you to edit or write to your .htaccess file, a critical configuration file. You just might need to do that someday.
Tip #5: You want a web host that provides 24/7 customer service. These days, with the right web host, ticket support and live chat are just as good as phone support.
Tip #6: Don’t automatically install caching plugins, even proprietary ones from your web host. As with all plugins, please do your homework. Test your site speed BEFORE and AFTER installing any plugin.
Tip #7: You want a web host with a stable Time to First Byte (TTFB) of less than 300ms. And if you can find a web host with a TTFB of 200ms or less, sweet.
TTFB is the time it takes a web host’s server to deliver the first byte of data for a requested page to a visitor’s browser. This is an important qualifier.
Knowing a web host’s TTFB is how you can verify if their speed claims are true. To check for TTFB, go to ByteCheck.
And run the web host’s homepage through ByteCheck six times. Then, you’ll know if you’ve got a winner or a dud.
There is likely no “perfect” host, but now you have information to help you choose web hosting more wisely.
It’s a miracle this theme loads in under 4 seconds (sometimes 3) with all of these obstacles stacked against.
The client wrote:
To make a long story short, the job was to improve the look of the website and improve the speed to 2 seconds or less. When he started the site was loading close to 3 seconds and as you can see it got worse to about 4 seconds. I believe it was $680.00 through a “developer” in India.
The theme “The7” cost $59 $29. The visual composer (drag-and-drop) is built-in for maximum customization. It’s “feature-rich” – meaning the theme authors from the Ukraine included the kitchen sink to appeal to everyone. Multipurpose usually means slow. The7 has been sold 219,000 times. It’s popular. Popular means “slow.” That is because people are attracted to themes that look pretty and have bells and whistles. But learning how to use The7 is probably as complex as a new computer operating system.
Million dollars worth of The7 theme have been sold to happy(?) customers. In the Ukraine, that’s a lot of money.
This is typical: 50 percent code weight and 50 percent image weight.
The goal is to balance aesthetics (branding) and speed (load time). Anytime, you increase one you decrease the other. Push and pull. More decoration slows down a site. Less speeds it up.
There are two types of aesthetic design: classic and expressive.
“The7” site leans toward expressive because of the colors, animation, and image usage. Image usage includes PNG transparency and large background images (layering effects). If too much expressive aesthetic is used, then the page gets visually noisy – and heavy. It distracts from the content (text). The goal is to get people to read – or click a response button.
Classic aesthetic is static, clean, and usually stark white. Sometimes referred to as minimalist. But it has it’s roots in Greek and Bauhaus design theories with white space usage, invisible grids, and golden-aspect ratios.
If a page is too classic, it gets boring and repetitive. If it is too expressive, it turns into distracting noise. A balance has to be found again for wisdom. How “good is good enough” is subjective and biased by opinion and perception.
It’s amazing The7 theme loads in 4 seconds with all the expressive design elements. 4 seconds is a typical load time for a WordPress theme that doesn’t have many images on the page — and no animation. But the Internet average load time is about 8 seconds. Which has been proven practically intolerable for users. The saving grace is the pages aren’t completely blank for 8 seconds. If it is blank, the site will most definitely be abandoned.
At this point, to get better speed, you’d have to throw money at the theme problem – or redesign. You’d have to sacrifice some expressive aesthetics – especially the animation. All for a few seconds of speed.
I do like the look of the site so my goal would be to improve the speed as you suggest. 4 seconds is just too long.
There are other things that can’t be changed: Fontawesome is included. It loads even if it isn’t used. SliderRevolution plugin is loaded on every page – even if there isn’t a slider present. Dashicons are loaded for every page. Various Google fonts are loaded. While all of these things can be removed with code modifications, they are part of the design and the site wouldn’t look the same – it may even break.
We’ve never had any success with W3 Total Cache plugin. So you aren’t the Lone Ranger. In general, if the site is already as optimized as it can get, caching just doesn’t make any difference. It doesn’t matter what caching plugin you use.
But recently Cloudflare and MaxCDN stopped working right so I disable both of them and they weren’t really working anyway.
Be sure to completely uninstall any plugins associated with those old CDNs.
Cloudflare CDN also failed in our speed testing. Same story as caching. Once a site is optimized, CDNs can’t help. Instead, they frequently slow down pages or cause “page not found” errors. It’s our opinion that CDN mainly helps with security and not speed. But if you have a grossly bloated website, CDN makes a difference. Or if you are selling to an international market (which you aren’t).
CDN and caching are band-aids for sloppy designers. Too lazy to optimize.
There isn’t anything we can add that would significantly improve the speed. We do believe W3 Total Cache is minifying your CSS and JS files. It’s also Gzip compressing all files. So you are getting some benefit from it. Those features could be added with other plugins – but there wouldn’t be a speed increase to remove the W3 Total Cache plugin.
Contact Form 7 is a heavy plugin. But changing it to something lighter won’t be significant. We’d leave it alone.
Conclusion: We suggest your website is good-enough to communicate for marketing purposes. Review the main goal of your site. Redesign should be postponed as long as possible.
Something we see bogging down homepages is the faddish inclusion of a huge Google Maps dynamic graphic. Often the trendy map isn’t needed on the homepage – or it isn’t needed anywhere! It’s gratuitous interactive bloat.
We understand needing a good map and directions. Especially if you’re a brick-and-mortar store – or have offices where you meet clientèle. Or you run a restaurant. Then people need to find you. We get it. So what can you do to keep your pages lightweight – and still have an interactive map?
First, let’s examine how heavy are Google Maps? They use an API (script) to call offsite web assets from Google’s servers. You can’t host these bits and pieces locally. That means the assets for maps aren’t cached. There are delays when servers talk to each other.
Often, Google Maps add at least 500k page weight. That’s our observation. But others have seen worse speed damage than this. More on this in a minute. Depending upon how you install Google Maps, the page weight loads on every page and post of your site. Even if only using a map on one solitary page. Global loading of a plugin or script is site drag. Most site owners don’t even know site drag is a potential liability.
To load the typical Google Map, it takes about 70 requests which is 2 megabytes extra page weight. Or up to 2-seconds load time. On slower connections and especially mobile ones it’s even more. – Offsite resource
Why does a map take so long and is so heavy? The majority of site owners use the easiest plugin installation – an iframe method. It gets the job done. It begins *building* the dynamic map from remote components during page load. It doesn’t take into account if the visitor is looking at the map – or interacting with it.
Instead of loading dynamic map data chunks, it’s better to load one single, static image. That’s about 50k file size. That takes a fraction of the time. The user clicks the static map. The interactive version then loads in a new browser window or on a new page. Simple offloading trick.
We prefer to open the actual Google hosted map page in a new window and not embedding the map into a local page. This completely offloads all heavy assets to Google’s servers and hosting.
How do you get a static map image? There are two simple ways we like. First go to Google Maps and use their tools to build the map the way you want and then do a screen capture.
Another way is to use an online free tool.
Go to Static Map Maker and use the form on the left to change the map. Leave the API field empty. We selected the retina option to produce a larger map. You can adjust the width and height to fit your site’s needs. Google imposes a maximum static map size of 640 x 640 pixels. But using the retina setting, you can get a 1280 pixel square PNG image. You’ll want the default roadmap type. Play with “address and zoom” to get the map you need. The preview on the right refreshes as you make changes.
This PagePipe speed tip is the fastest way to manage Google Maps for mobile websites.
Somebody changed something at Google. Now your map and driving directions are broken again.
We recommend future proofing you map page from Google changes.
Insert a screen-capture, static JPEG image and make it an image link to open a new tab with Google Maps. Keep maps heavy load off your site. Keep it on Google site instead.
There are two resulting benefits: your site is always fresh and current – and your site loads faster.
This is the best way to handle Google Maps especially for mobile devices. And it always works.
So deactivated your Google Map plugin. It’s no longer needed.
SiteGround isn’t always kind to their customers. We probably only get whiners coming to PagePipe searching for change. Speed anxiety is their motivation.
SiteGround’s home page says, “Latest speed technologies are our passion.” We also have a passion about speed. But we say, “You can get WordPress speed on ugly, cheap servers.” SiteGround thinks no one knows more about speed than they do. Experts? Really?
PagePipe home-page loads in 900 milliseconds cached in our browser (at this moment). 1-second – even with cache cleared. That doesn’t mean we walk with the speed gods. It means our load time is good right now. We catch it behaving “just fine” much more than we find it failing. If it’s good for 80 percent of the time. It’s “good enough.” How far does it drift, ±50 percent. Horrible, huh. We don’t have an expectation that GoDaddy delivers better than that. Nor are we paying for better speed.
But more often than not, GoDaddy delivers 200-millisecond TTFB or better. Go figure. At his moment, it’s 139 milliseconds. That’s strange – but what we usually get. The other day a test was the worst we’ve ever seen, 15-second TTFB. Why? We don’t know. But that’s really rare. But we caught it. Are we ashamed? Nope.
Do we recommend GoDaddy? Of course not. They’re cruddy. We’re proving a point about cheap speed results. No SSD drives. No hopped-up CDN. No server caching. Only vanilla, shared, magnetic hosting.
If site owners don’t care about speed and choose ignoring it deliberately, then no big deal. It’s a business decision and choice everyone gets to make.
Do 1-second speed reports matter for desktop? Not much. But for mobile, it’s significant. Those translate into less waiting. In fact, they theoretically load at user-expected desktop speeds. For site owners with 70-percent mobile traffic, it’s a godsend.
People’s expectations with SiteGround is 100-percent goodness. Why? Because SiteGround claims having the “latest speed technology.” But it’s just mumbo-jumbo, theoretical speed claims – not actual measurable milliseconds.
Many hosting customers don’t know about speed – or don’t care. In that case, fine. If they are happy, no problem. But to finger point and say, “It’s not our server speed problem. It’s Google or WordPress voodoo or you’re technically stupid.” That doesn’t sit well with us.
HostGator (claim: powerful hosting 2X Faster) and Bluehost (claim: 2-million websites worldwide). Both brag about their prowess. [note: 2x faster than what? a turtle?]
SiteGround is on our radar. No one’s ever written us about SiteGround wonderfulness. We have a self-proclaimed mission to counterattack speed incompetence, hypocrisy, and deception.
Had a wonderful experience with SiteGround? Congratulations. But have you watched your TTFB (server delay) bounce around over time – for top-tier GoGeek prices. Better check it out. READ MORE HERE
Speed trivia? Perhaps. Remember, our grand purpose. It’s saving the Internet from WordPress speed abuse – one website at a time. We help our little corner of the world.
Web work is disposable dust. In the future, new solutions will obsolete our speed expertise. Except humans will continue to abuse and overload websites. That won’t go away. Job security? Nah.
The best and fastest websites and hosts don’t exist yet.
“SiteGround was driving me crazy blaming WordPress plugins and my site’s coding for the problem of slow page loads. They made a simple php script to show that their servers were fast. It then got terrible scores on GTMetrix and Google PageSpeed Insight. It was a simple script. They tried to prove a point, but ended up disproving it. Their simple script loaded slow! Then they said it was a Google PageSpeed problem. I said the times were always inconsistent. They said that was Google’s problem. What?
I’ll be leaving SiteGround soon. Great customer service for slow servers isn’t worth it for me.
The tech rep who was dealing with my support ticket eventually started getting snarky. He repeatedly said, “… as I’ve said before…” and similar things without trying to understand what I was complaining about or without trying at all to offer or look for a solution.
Eventually, I said I will start looking for a new host, and he replied along the lines of, “Thank you for your time. Please contact us if you have any problems.” Hmm…
I was SOOOO happy to find your PagePipe article as it mirrored my experience and frustration. I really like your analytical way of thinking.
By the way, I pointed my servers to FastComet hosting and although the PageSpeed Insight scores are inconsistent, they are consistently better than they were at SiteGround. My TTFB went from an F to an A at webpagetest.org.
These are things Siteground said they couldn’t help – because it was the coding of my website causing the problem. More importantly, the GTMetrix and Pingdom times went down. Not by much, but as you know, every little bit is hard earned when you’re under 2 seconds.”
He’s the author of SiteGround’s free ebook. “SPEED MATTERS: 21 Expert Tips to an Ultra-Fast WordPress Site.” Hristo’s an expert on WordPress speed optimization. He has a video online from a 2016 WordCamp. But we have found a few ideas in his ebook that don’t measure up to our experience and testing. Naturally. But most of his speed suggestions are safe and sane.
SiteGround implies that somewhere there exists a mandated 1-second barrier. Is their hosting service the only method to break 1 second? Speed authorities think otherwise.
Why are they advocating an idealistic or sometimes impossible 1-second goal?
A well-optimized site and SiteGround’s servers on good days can achieve this. We’ve used SiteGround with clients. But it’s possible to do 1-second loads on cruddy hosting too, if you abide by certain principles.
Good mobile user experience needs the fastest page loads.
One second is instantaneous gratification for users – and has been for decades. But 2-second loads are a more realistic optimization and performance target. And those are desktop hardwired speeds.
What is realistic on wireless mobile? We suggest the performance budget is 3 seconds. That is also the user expectation – for today anyway.
From this web study, 4 seconds is average mobile speed.
You can waste a lot of resources attempting unreachable maximization (100 percent). Optimization, or 80 percent return, is more affordable and realistic. Avoid waste from gold-plating or over-engineering your website.
Note: PagePipe’s Home load time is under a second (most of the time) – sometimes 1.2 seconds. We use cheap GoDaddy “evil” because our goal is good speed results even under bad conditions. We practice what we preach. Just like Google. Ha!
1Identify and Prioritize Issues. SiteGround lists GT Metrix and Pingdom online tests for speed benchmarking. Easy. Knowing what the test means takes some educating and reading. Learning curve stuff. Our preferred test is WebPagetest.org that’s geared for professional optimizers.
SiteGround gives a good piece of advice about speed testing:
“Even though most of the benchmarking tools will give you a ‘grade,’ don’t go too far chasing it.”
And this supportive quote below is from WP Rockets FAQ page about ratings:
“Performance ratings are mainly indicators of good practice. Ratings tools check that the optimizations have been made. These ratings do not indicate, however, the actual speed of a site, they are only indicators. Good ratings do not guarantee a fast site and vice versa. The actual page load time is the most important metric to look at.”
We’d add to that our puny opinion: “WordPress – by it’s very nature – cannot pass many speed tests. At least, not without major expenditure and effort. In the end, those improvements will not necessarily make the site faster. And speed (human perceived load time) is the only thing that counts – not scores.”
2 Reduce the number of Posts Shown on the Index Page. This is only a problem when the blog-listing page is the Home page. Then it may show too many featured images – if they are used. Installing a lazy-load plugin fixes this. We recommend Rocket Lazy Load plugin.
SiteGround recommends aninfinite scrolling plugin or changing WordPress for “show at most” values. Those are good ideas. But, infinite scrolling can activate jQuery and add page weight, too. So test before and after installing any infinite-scroll plugin.
SiteGround also recommends pagination of long pages into sections using the <!–nextpage–> tag. An easier method of insertion is using Page-Break plugin. It adds a control panel button.
3They recommend getting rid of sliders and using just one image. We agree. We’ve written about slider extravagance before:
SiteGround recommends two other slider plugins we’ve tested before. We weren’t that impressed with the load times. Our extreme speed philosophy is no sliders on the Home page. Period!
4 SiteGround recommends using appropriate image sizes. Again, we agree. They don’t have a plugin solution recommendation but our safeguard utility is using Imsanity plugin. We also use this plugin to solve the next problem they talk about:
SiteGround recommends using EWWW image Optimizer plugin. We don’t recommend it because it can cost money. We’re pro-free stuff. And there are plenty of other free alternatives and strategies. EWWW isn’t the best image optimizer as supposed and reported by many. It’s just one of a multitude of options.
Nothing will ever beat just optimizing images by-hand. Use an image processing program (like GIMP or Photoshop). Do that before uploading images to the WordPress media library.
Our most unpopular but best speed recommendation: Don’t use images whenever possible. That’s right. None. Just design with text and unicode symbols.
If you must use large images, don’t make them JPEG photos. Use PNG illustrations instead with limited color palettes. This produces the smallest, fastest file sizes and reinforces your branding.
6Reduce the usage of external fonts. We agree with this suggestion but we go even further. There are various plugins that can “exorcise” this font fluff. Our recommendation is Disable Emojis and Remove Google Fonts or Disable Google Fonts plugins. We’ve used all these plugins.
Or choose a theme that doesn’t use any webfonts – just websafe ones. Yes! Those themes do exist. We’ve written about those, too;
To get rid of Font Awesome, you’ll need to use Asset Queue Manager plugin. This plugin can “break” your site with some themes so proceed with caution. But we love this plugin and use it a lot to strip down bloated themes.
7Manage the volume of comments on your site. This isn’t the first we’ve read about comments slowing down sites. But we haven’t seen any real data to prove it (yet). We know a few reasons why comments in theory cause slowness from database issues. SiteGround makes two plugin recommendations for comment management. But we’re more hardcore about achieving goals and streamlining sites. We say, “Get rid of comments completely.” Read about our radical ideas on comment management:
8Enable Gzip compression for your pages. SiteGround recommends editing your .htaccess file but don’t say how to do that. So SiteGround must not have Gzip enabled like on some hosts. This .htaccess file edit is not an easy thing for newbies. We think it’s a pain. It’s easier to change the .htaccess file on your server with Far Future Expiry Header plugin. Read more about Gzip:
9Enable caching. OK. Unlike the rest of the world, we don’t think caching helps much on a well-optimized website. We just never see speed benefits. SiteGround makes two plugin suggestions. WP Rocket, a paid plugin, and WP Super Cache, a freebie. We’ve experimented with both of them and like we said: “We aren’t sure they help much.” Yeah. They’re band-aids for sloppy-built websites. But don’t help quality sites.
THEME AND PLUGIN OPTIMIZATION
The recommendations in this section are kind of silly or else just common sense. Someone must have been trying to fluff up the report with filler. SiteGround then goes on to tell us many things we’ve already covered:
10 Select a reputable theme from a solid provider. We only use free WordPress themes from their repository. That is our choice and self-imposed limitation. It speeds up our decision making and creative process. We don’t have time for shopping. We don’t spend money on complicated “premium” themes. Not us.
11 Avoid bloated themes. Their explanation of what is a bloated theme is good. Avoid sliders.
12 Always use a child theme when creating your website. This is common sense for safety sake. It prevents future updates from overwriting your customization. But what does a child theme have to do with speed? Nothing. Child themes load another CSS file. So this recommendation doesn’t make sense.
13 Optimize for mobile devices.Really? You need to tell people this? And they recommend the plugin WP Touch (non-native mobile conversion). This isn’t good. Even when they afterwards say, “Having a native mobile version is always preferable.” Even that isn’t a good idea. What’s preferable is using a responsive WordPress theme. A mobile version is a second version of your website that sniffs to detect a small screen device and then redirects to the mobile version. Not the same as responsive which just serves one site – no duplication.
14 When using icons, use an icon font. We despise icon fonts. They are heavy and slow-loading – the bane of speed. We disable icons whenever possible.
15 Don’t overlap functionality with plugins. A good suggestion but isn’t this just common sense again. Don’t duplicate stuff. Simplify.
16 Always keep your plugins up-to-date. This is just good housekeeping. Staying updated and current helps speed? We haven’t seen any evidence yet. SiteGround claims it’ll give your site a huge performance boost. Serious? Got some experiential proof of that? Or is it just theory? Or exaggeration?
17 Cleanup your plugin options from your database. We do this as best practice. Again, it’s seems like just common-sense good housekeeping to us. We’ve never seen any speed boost yet from cleanup – even with big fat databases. There are several plugins to do this. We don’t make a recommendation. We’ve tried them all and it gives a nice feeling that you checked the databases and verified. But no speed improvements whatsoever measurable.
SERVER & HOSTING OPTIMIZATIONS
Now we’re into the promotional selling materials of the brochure. In other words, SiteGround specific features they hope to motivate us to buy.
18 Take advantage of server level caching.
This is pure specsmanship and boasting about SiteGround’s capabilities.
19 Use a CDN. We’re not sold on CDN’s. They don’t help much for a well-optimized site. Just like caching. If you need a CDN, it is indicative you didn’t build your site as well as you thought. Another speed band-aid. The author said in his video to test CDNs since they can slow down your site. Good advice. That’s our experience with free CloudFlare. Slowed by server 500 and 501 errors. Longer TTFB (time to first byte). Badness.
20 Use SSL and utilize HTTP/2. This is a nice way to upsell and make money for SiteGround. These aren’t necessary for speed. SiteGround was kind enough to mention a free alternative. Secure third-party transactions like with PayPal means this stuff isn’t needed.
So what’s missing?
We think social links are a big culprit for site drag. That isn’t mentioned anywhere in the Ebook. Most site owners don’t realize social media gives little benefit. But they feel like an outcast if they don’t have it. Yes. There’s stigma to conform to the herd. Don’t give into this peer pressure. Social buttons and likes slow down pages.
Remember what your mother taught you about peer pressure and popularity. They can be bad news.
If you have to use social media, use static image buttons or CSS buttons links instead. The fastest loading social-sharing button is none. Do value analysis. What kind of return are you getting on your social media? Is it a time waste generating that social content?
Hristo advises on another blog against going overboard with Social Media widgets or plugins. He just forgot to include it here. Social widgets and plugins ping their respective servers, delaying page loads. Particularly, Hristo says not to use IFRAME. He recommends using one plugin that covers all the social networks. (Facebook, Google+, LinkedIn, Twitter, Pinterest etc). Don’t use a separate plugin for each network. Again, we think social media is as useful as a cast-iron paddle in a chicken-wire canoe.
We disagree with this ebook, web hosting isn’t key to great performance.
We’ve seen sites on great hosts (including SiteGround) with lousy speed tests. It’s more essential to use speed strategy for balancing aesthetics and speed.
They said nothing about failing some basic Pingdom tests and the simple plugin solutions:
Speed Booster Pack 82k Speed Booster Pack allows you to improve your page loading speed. You’ll get a higher score on testing services. (GTmetrix, Google PageSpeed, YSlow, Pingdom, or Webpagetest). We’ve tested this plugin. We found it’s minification features succeed where other minification plugins caused conflicts or white screens. It’s not for everyone but worth noting here.
The SPEED MATTERS ebook is better than most with good speed suggestions. But plainly selling services. At least it doesn’t perpetuate the myth that speed improves search engine optimization (SEO). That’s always a disturbing lie told by many site optimizers.
Ideas we agree on:
We agree performance test “scores” are meaningless ratings. Only speed measurements in milliseconds count. Or even better user perception of fast speed.
We agree sliders suck. But we say get rid of them on homepages.
We agree on image optimization. We even define how good is good enough in a downloadable PDF.
We agree on Gzip. We tell how to activate Gzip using a plugin without editing the .htaccess file by-hand.
Social links are baggage. Even though the author left that off. We know he agrees from other blog posts he’s written.
Ideas we disagree on:
We don’t think a good host is the key to site performance. We think speed strategy is the answer.
They recommend a 1-second performance budget. We recommend 2-second loads for desktop and 3 seconds for mobile as best practice. Even though PagePipe is a 1-second site! Other experts agree. Even Google says 2 seconds is good enough.
They recommend EWWW image optimizer plugin. We say that’s poo. We recommend free Imsanity plugin instead. It’s configurable with a maximum width, height and quality settings. And authored by the same guy.
We recommend using no images – or substituting PNG illustrations for heavier JPEG photographs.
They recommend using Google Fonts sparingly. We say get rid of them completely by substituting websafe fonts for speed. We also say eliminate Font Awesome icon font when possible – and always get rid of emojis. We recommend various plugins to do these eliminations. These are drastic but necessary measures.
They say manage (reduce) comments on your site. We say get rid of them completely with a plugin.
They say use a caching plugin and CDN. We say these are unnecessary band-aids if you build a quality speed site using strategy.
We think recommendations 10 through 17 are common-sense housekeeping or plain silly.
We think items 19 and 20 are less-credible selling promotions.
So half the report – the first 9 items – are worth reading. We think they aren’t aggressive enough to achieve their one-second page goal.
The LTI SEO plugin is the lightest SEO (search engine optimization) plugin built. It’s installed and still works on various client sites. Why? It’s old but not broken.
Clients insist on SEO gyrations. It’s a placebo effect known as “SEO tweaking.” A futile and silly habit for neurotic, obsessive-compulsive dullards. It’s an addictive tranquilizer for the vigilant nervous. Then they sleep at night thinking all will be well in Googleland.
SEO tweaking is self-defeating behavior. It’s insanity. Go ahead! Kowtow to what Google publishes as metric truth. See if it makes any measurable difference. It won’t.
Hundreds of third-world promises: “Get easy number one Google page ranking.” These promises flood my email inbox. Isn’t source credibility a clue of the valueless remedy called SEO?
LTI SEO plugin is not installed on PagePipe. Why? Because relevant desired content is the only thing to quickly alter SEO. The slow way is focusing on speed or rewriting post titles. Yes. We do those things. Those alter user experience. UX has future value. But no guarantees.
Relevant content is the hope. Not SEO plugins! Not gaming the system. There’s no system to game. No game to systematize. Vaporous provable claims or fanboy testimonials are weak. But no measurable proof, research, or data of real ranking improvement.
Millions are wasting and squandering resources on SEO.
We don’t defend any lightweight SEO plugin. Nor defend any SEO plugin – period. They’re an utter waste of human resources. Can you feel our oozing contempt for those selling SEO? And especially anyone claiming speed makes an immediate difference in SEO. Absurd.
LTI is the only SEO plugin installed at customer’s dictatorial insistence. Never volunteer installing SEO baggage. You could substitute 14 other site features with discrete plugins consuming comparable speed.
The “very popular All-in-One-SEO Pack plugin”?Tortuous slow-loading nonsense. We’ve tested it’s speed – 174 milliseconds added on every page and post of your site. But speed isn’t the point. It’s a waste of content writing time!
Just did a quick test. HotJar adds 500 milliseconds to your page load time – globally. THAT move blows 25 percent of the entire performance budget. You want to reconsider activating that plugin and API?
HotJar – an all-in-one analytics and feedback platform – provides heatmaps, visitor recordings, conversion funnels, form analytics, and more.
We find Hotjar often on commercial websites. Especially websites with speed problems.
What Alfredo Gutierrez of FunctionLabs says about the benefit of Hotjar:
Why user recording?
Hotjar allows us to literally see what people see, rather than guessing at what happens between pages.
Google Analytics shows the number of people who clickthrough or purchase, but Hotjar shows us friction.
It shows where someone scrolls to and between, what they see, and what happens when they do something.
This saved us numerous times. Here are two I recall:
– Android users would click on a button, which led them to bounce instead of go through to the next page. We didn’t test thoroughly enough, and turns out the email code would kick them out of the browser app entirely, and there was a bug where the back button wouldn’t work. iPhones would just pop up their Mail app, allow for the email to be sent, and immediately send the user back to the browser. We were about to stop Android traffic altogether because it didn’t seem like it was converting, but it was actually just an Android bug. So, we excluded Android from this particular funnel split test, and it converted again.
– We tested a new long-form sales letter. GA / Mixpanel showed low conversions on two of four variations. Policing that data with FullStory showed us that although the conversions were lower on the two, the users were much more engaged with certain page sections. We could’ve nixed the two variations and committed to a different hypothesis, but instead we took the section that saw a lot of interaction, and mixed it into the original variation. That resulted in a 27% bump in conversion.
Why does a small site / business need so much data recording?
My philosophy on this is that the smaller budget you got, the more data you need.
With our previous big business, we had so many media buys and so many transactions coming through, that we could ignore issues and still make a profit.
The smaller businesses have a much slimmer margin of error.
They’re also gun shy, and more demanding of information, even though traffic is about 3x more expensive now than it was 5 years ago. They want information after $1k adspend. They want to know which variation wins after 40 conversions.
They want to squeeze ‘insight’ from insignificant information.
So, user recording gives us a *direction* to go down, if we want to test something. Instead of a hunch, we can test something backed by something.
And when we do have a lot of information, it’s also a great policer of our assumptions.
Thanks, Alf, for sharing your knowledge about Hotjar benefits.
What others have to say about Hotjar and speed:
The one script Hotjar has you add to your site adds a tremendous 479k to the size of the fully loaded website. To put that into perspective, my site was only 1mb before adding the code. This means that hotjar is almost the same size as HALF OF MY WEBSITE!
Despite the asynchronous loading, you can feel that the website is sluggish when Hotjar is enabled.
So you have to ask yourself: “Do I value the data hotjar is providing over a faster website for my users?”
My website typically loads between 1.5 – 2.5 seconds without Hotjar, and up to 4 seconds with Hotjar – Andrew Curtin
“HotJar significantly reduced the loading speed of our website. This became a serious problem and we eventually had to remove it all together.” Patrick Eng, Marketing Technologist
“Snippet. Sounds small. Lightweight. Not a big deal, right? Well, wrong. One script alone adds just a tiny bit of extra to your load time, but scripts can really take a toll once they are combined.”
“Unfortunately neither you nor WP Rocket (nor any other caching plugin) can control the speed or performance of resources which are located on external servers (like Google, Hotjar or Facebook servers).” –Alice Orru
You insist you must have HotJar information? Fine. But after testing for 3 months, deactivate it. Test again in another six months if you must. Don’t just leave it sitting burning up speed. Give your site users a break.
Normally, we write about releases of default themes – and we torture test it for speed. Not so in the 2018 year. Why? Gutenberg delayed Twenty-eighteen theme completion schedule. The powers that be said, “Spring 2018” for introduction. That never happened.
Twenty-nineteen theme‘s official release was October 16, 2018. There is no Twenty-eighteen theme. A hole in the dynastic chain. So someone pounce on that theme name! Great opportunity to confuse the world. And perhaps make money from it.
Twenty Nineteen is part of WordPress core version 5.0 and up.
Disable Gutenberg Completely disables the Gutenberg block editor and enables the classic WordPress post editor (TinyMCE aka WYSIWYG) for lighter coding and simplicity.
Twenty-nineteen theme is Gutenberg-specific. We recommend two alternatives for speed:
LOOKING FOR A NEW THEME REPLACING DEFAULT 2018? READ THESE:
You’ll weep when you read Google’s 200 Ranking Factors: The Complete List (2018). Because it’s so sad? No. Because it’s so overwhelming. It an encyclopedic explanation of Search Engine Optimization (SEO). The article makes SEO sound so complex and mysterious – and confusing. It implies little nitpick details make a big difference. It’s anxiety producing.
But still, we recommend reading the whole thing anyway. Some people may then try gaming all silly 200 SEO factors. Don’t go there!
Will these “tricks” help more than writing good content?
SEO fiddling is a waste of time.
Be calm. Good page ranking is within your reach if you:
Write about topics people want to read.
Write content in an interesting way that keeps visitors reading more.
You make text readable. What’s readable? Readability is the appearance or perception text may be easy to consume. That mean placing subhead and captions for skim readers. Use short words, short sentences, and short paragraphs. Then they will spend more time after skimming content cues and clues.
The article doesn’t reveal the hierarchy of ranking factors. What matters most is the summary list at the very end. Many will never make it there. It requires a lot of boring scrolling to arrive at the real pithy basics.
The author presents a shortlist at the end. These are the real fundamentals of what counts. They’re the most important Google ranking factors (or signals) according to the SEO article. But they aren’t explained in plain English. So we’ll attempt translating some more.
Here’s his list with our commentary:
Referring domains This is other websites linking to yours. It’s them choosing to advertise your site’s valuable content for free. Again, relevant content is good writing about interesting things. So get rid of your dud articles and uninteresting posts. Don’t make site-noise diluting “user attention.” That’s simple positioning strategy 101. Referring domains is the biggest influence on SEO. If you game inbound links with a link farm or purchased backlinks – there’s bad news. When Google gets wise to your ploy, they’ll punish you. Even blacklist your site. That sharp retaliation indicates the significance of this “ranking factor.” No mercy.
Organic click-through-rate Organic means Google non-paid listing. CTR is the percentage of *impressions* resulting in a “listing click” for a website. What’s an impression? That’s the number of times your listing (page title) gets viewed on the search engine page. You can view a page of 10 listings. If your your page title is chosen – bingo – that’s a “click.” If you own most of the 10 first-page listings, that’s called page dominance. When a searching reader suspects finding relevant content on your site, that’s information scent. What affects visitor suspicion or cues most? 1) The page title. 2) The “snippet” constructed by Google RankBrain, and 3) your publication date (freshness) if indicated. Publication dates are changed in WordPress for freshening up evergreen content. The snippet refers to a description extracted from page content.
Domain authority The only thing controllable here is the longevity of your domain name. That’s right – the date when you registered your name. You can buy an old domain name that’s in use and re-purpose it. Gaming the system. But then it’s back to writing good relevant content as the main influence of authority. Serve up user-valued information.
Mobile usability Mobile-first ranking is only two things: responsive screens and fast speed. And avoiding certain stupid web practices anyway. Like interstitial ads. Google AMP and Mobile Applications aren’t mentioned as good tricks. Praise the Lord!
Dwell time This is also called engagement. It’s time spent reading or consuming your wonderful page content. What helps with engagement? Good writing and interesting images. And suggesting relevant articles to keep people on your site reading more once there.
Total number of backlinks You can’t game or cheat backlinks without penalty. See the first item “Domain authority.”
Content quality Isn’t this about writing quality? Learned skills. Writing stuff people want to read.
On-page SEO A page title is a solid suggestion. This is an interesting and attention-getting headline. But it also needs to contain your keywords (positioning statement). Example: Yoast SEO plugin affects mobile WordPress speed. Then change it into a question: How does Yoast SEO affect mobile WordPress speed? or 10 ways Yoast SEO ruins mobile WordPress speed. Use good headline writing styles developed during the direct-mail years of graphic design.
ESOTERIC ON-PAGE-SEO DETAILS FOR THE TERMINALLY BORED
Outbound links are a relevancy signal. PagePipe uses outbound links (resources) for credibility enhancement. Readers appreciate offsite links. How do we know? Feedback! They tell us in emails. And they see it as courageous. Because we might be sending them away from our site for good. Risk taking or confidence our content is good enough. But most often, they return to our tab.
Internal links are good (of course you reference your other written material – duh. Common sense).
Speed affects repeat visits – is that a surprise?
Use synonyms for keywords – another “shocking” suggestion.
Use ALT and title tags with keywords on image file names.
Longer content ranks higher. Increase average dwell time by writing long, engaging content that keeps people reading. If you love your site topic or focus this shouldn’t be a burden. If you don’t have a fascination about your chosen field, you’d better quit now.
Isn’t “on-page SEO” obvious best practices and common sense for writing?
Here’s the bottom line:
WRITE CONTENT PEOPLE WANT TO READ
That includes readability – not mentioned anywhere in the article or list. Make words look fun, easy, or interesting to read is a goal affecting SEO. Or at least, get out of the way of reading the words like fast speed or responsive sites remove barriers.
On websites, transparent features mean being invisible or undetectable. Speed is transparent when it’s fast. No one notices a fast page. But everyone hates a slow one. The best speed is instant page changes when clicked. Good speed is a transparent feature differentiating a site from competitors.
It’s our opinion, social sharing doesn’t affect page rank directly. But in general, it takes traffic away from your site – and when there the seduced visitor never returns. Social rarely brings quality visitors. Social 1) slows down your pages, 2) causes link clutter, 3) and takes people away. Is that helpful?
We ask clients if they quantify how much profit is because of social media. They can never answer that question. Why? Because it’s immeasurable. They’re following the web herd. All paid themes come with social links built-in. So that feature must be good. Right? Themes sometimes have heavy sliders, too. Uh. Not good. Theme authors include every feature trying to please everyone.
In another article, the author recommends using “2018” “best” “guide” and “review” in titles. Make the title an “H1” tag. WordPress already does this.
Your SEO mission: learn writing skills.
Content IS the user experience.
Please remember relevant content is number one for SEO. Speed affects User Experience (UX). Good UX then influences metrics like dwell time, bounce rate, and click through. Google interprets those as user intent.
User intent is a major factor in search engine optimization and conversion optimization.
Speed affects page ranking less than 1 percent. But everyone hates a slow page. That’s not being hospitable or polite.
Speed is about kindness!
Common-sense tip number one: Good Titles.
Writing good titles for your WordPress posts should be obvious. But we’re always stunned at how many sites don’t use this simple tactic to improve Search Engine Optimization (SEO) and click through.
Page title is important. People choose to click your listing on the Search Engine Results Page (SERP) instead of nine other competitive page titles. A title arouses human curiosity. If it doesn’t, it’s a loser. The WordPress permalink is written for machines (search). It doesn’t need to match. Title is an important controllable indicator of relevant content in the search listings. This affects findability, too.
People read your article based on it’s title.
A good title reads like a headline. A plugin we used for a year is a good teacher for writing better titles. Title Experiments is a free plugin available in the WordPress plugin repository. The plugin allows you to test multiple title variations for any post or page.
Title Experiments relies on the old classic WordPress editor. It won’t be updated to support Gutenberg block editor added in WP 5.x. This is the author’s excuse to ditch the paid plugin. It’s plain he’s disappointed by the lack of plugin income. No enthusiasm to go on.
Our workaround is simple. Use the Classic Editor Plugin with the Classic Editor Addon. So even if your core version is 5.0+ and your running PHP 7.x, things still work.
Title Experiments is a helpful plugin. We learned a lot about what titles work and what doesn’t for user engagement. But the heavy plugin was a top contributors to site drag. So we removed it after our education on writing better headlines (page titles).
Title Experiments relies heavily on the old editor of WordPress and will not be updated to support Gutenberg (WPv5.0+).
This is an author’s excuse to ditch the plugin.
Every year we review the last 4 months of traffic and see what is performing and trending. We’ve found our worst performing posts always have a lame headline (title). Renaming the post is the best thing to try first. We also dump dead posts or consolidate posts. This has proven effective for three years now.
For example, a mere label such as Ferritin and Hypothyroidism could be rewritten for human interest.
“What are your optimal ferritin levels if you have hypothyroidism?”
That makes people curious and they click. Questions are always good. And including the word “you” is beneficial. Answer the readers question, “What’s in it for me?”
Purging your site is wise and focuses your content. That’s good positioning strategy. It affects perception of your site credibility.
How to find out what you should be writing about?
Intuition is needed for what future content to add. Not just metric history evaluation. The best article to write probably isn’t even on your radar yet. Our best post ideas come from reader’s emails who have questions. When we’re done writing long answers, we convert the email into a post – or add to an existing post.
Analyzing your inquiries isn’t something Google Analytics can do. Except for one helpful thing:
If you go to Google Analytics > Behavior > Site content > View full report (down in the right hand corner), you’re shown the top 10 of xxx pages. In our case 378 pages, we then change the “show rows” to 400. You then can see all posts and pages by popularity. You will see some entries with the following format:
This line above originated from our WordPress search box. A human couldn’t find something they needed on our site. Important info. They wanted to know more about Beaver Builder pagebuilder plugin.
We don’t have a Beaver Builder article. Do we need one? Maybe.
Going to the top of the GA page, there is an Export function on the right. We download the entire set for whatever period we choose and import that into a spreadsheet.
Then we categorize and sort the “searches.” The results reveal what people were looking for. We then test by doing a Google Search on the terms with the name “PagePipe.” That reveals what kind of placement the search phrase gets in the rankings.
This influences what we write about based on reader’s questions we’re not answering. So far this is helpful. How else can you learn what you don’t know?
From our recent analysis, we generated the following preliminary titles for future posts:
Why don’t we write about good hosts? Why only the bad ones?
How does cookie consent compliance affect speed?
Measuring HTTPS/SSL drag with ByteCheck
Why we don’t review paid themes
Why we don’t recommend CDN
How to use Cache Enabler plugin for speed.
Is Imsanity plugin good for speed?
How to use Autoptimize plugin for speed.
Magnetic versus SSD hosting for speed
What is site-origin optimization?
Speeding up Astra theme
Speeding up WooCommerce sites
Why use twenty-seventeen theme instead of twenty-nineteen?
One of the biggest decisions is deciding what to not sell, who not to sell to, and what not to promise. Telling people who you’re not is just as important as telling them who you are.
You must have something of value to sell. This is called the value offer for other people. This is about some pain or anxiety visitors are trying to resolve. This is their motivation.
Creative positioning strategy is a short cut to buyer’s motivation.
An offer includes terms, warranty, delivery, price, incentives and more.
People won’t be instantly convinced you’re a credible source. Web visitors are suspicious of every website.
In Google-speak, motivation is “intent” or “relevance.” Search engines attempt to rank the content of your site for intent, relevance or credibility.
Relevance is often determined by how many people searching for a key phrase go to your site (click-thru).
So. Why aren’t you getting traffic? First, look at your first-page Google listing competitors. They’re big companies with tons of articles, authority and credibility. You have to outperform them.
Does your site appear on the first 20 pages of Google for your preferred search term. Or after ten pages do you still have nothing? Focus on what matters most to people (relevant topic).
If 33 percent of your blog traffic is funneling through an unrelated topic page, these people have no intention of buying services or products. We recommend monetizing the page with a relevant paid download – or spin it off as a separate website – or else get rid of it.
Why get rid of a page creating 10,000 visitors a month? Because they aren’t qualified leads. It’s causing noise or dilution. Content pollution.
What’s a qualified lead?
A qualified lead (visitor) has money (budget), authority to buy, is ready to buy (timing), has the problem you can solve (need).
If a site bounce rate is 80 percent, those are people who don’t care and leave immediately. While we’ve seen worse, that isn’t goodness. Only 20 percent of people visiting stay and read something. The most they’ll read is a partial article. 80 percent leave instantly.
Maintaining an email list may not improve business profits. Selling an Amazon book may not increase business profits. They do increase “credibility” but they don’t increase profits.
Credibility consists of three components: trustworthiness, expertise, and enthusiasm. Credibility influences people or persuades them you can deliver what you promise. Remember they’re suspicious of all websites, not just yours.
Content is the user experience. What helps convince visitors you’re credible is how much content you’ve written related to solving their problem. That’s an “authority” goal of 50 articles. But if the articles don’t help them and are just fluffy, they won’t be convinced. “Fluffy” is referred to as thin content by Google. When you give away valuable information, visitors (and Google) are more prone to trust you more. The trust less if you charge for every little thing.
We dump 30 percent of PagePipe’s technical content each year. Typically the 30 to 40 least popular articles. Why? We revisit our core pages, homepage, etc. and improve how well they target our specific audience. We don’t want to dilute our best content with thin content. That makes it hard for people to decide what to read.
Our goal is selling:
Our break even is incredibly low. We don’t use any paid plugins or themes and we host on a cheap, shared server ($70 to $95 per year) with no paid services like CDN. We don’t advertise or monetize.
If you have no enthusiasm for your site topic, people will know. They will be unconvinced you’ll improve their life.
Will you produce valuable content for your audience? That’s more important than speed.
If I use ByteCheck to run a test against PagePipe’s homepage, it returns the result of ~663 milliseconds TTFB.
If using PageSpeed Insights which seems to be the go-to tool, it returns ~ 740 milliseconds.
Is a TTFB difference of 77-milliseconds high?
That 77 millisecond deviation between different measurement methods is common. Time to First Byte is server overhead. It’s affected by many things that slow down the server. Some servers fluctuate wildly: SiteGround hosting in particular. Some oddity is because of security and link checker plugins (server resource intensive).
All speed tests give different results. If you’re getting an offset of only 161ms, on two different tests, that’s amazing.
What we trust most is our desktop browser timer addon. And our eyes.
We don’t use a real stop watch. We use a browser load timer addon or extension.
Google doesn’t use PageSpeed Insights – or it’s criteria for page ranking. It’s not connected to their algorithm.
It’s a separate idealistic and ivory-tower test created by egghead scientists. It’s an inexplicably complex toy puzzle to solve.
A good score or bad score on PageSpeed Insights doesn’t affect the ranking in any way.
We’ve seen pages that load in 12 seconds get all “greens” and a high score. It is impossible to have WordPress pass the test with a 100. Yet, 1/3 of the internet is built with WordPress. What kind of test is that? A biased one.
The limit of human tolerance for machine delay is 10 seconds. A page that slow can still pass their test as “green.” Not always. But it deceived with trickery.
This silly speed test causes website-owner anxiety.
The only valid measurement is a browser timer (stopwatch). It’s only accurate for its geographic location. All speed tests are flawed. There are too many physical variables. Most use simulation. They don’t give absolute results but are instrumental in measuring incremental improvements. Relative measurements.
You cannot improve if you can’t measure. Cruddy machine measurements are better than no measurement.
Many things cause TTFB delays. Most variables are beyond the control of the site owner. Except one. Move your site to a different host and wait for their great TTFB to deteriorate. Will throwing money at the server problem make you more profit? Doubtful. The ROI is poor.
What affects page ranking is the user experience.
Speed is a primary blocker of a good user experience. Number 2 is aesthetics. User experience is how people *feel* when they use your site. Google can’t do the primary measurement of that *feeling*. But they can measure secondary things with machines like:
These are indicators of inferred quality content that satisfies the user’s need. But if users don’t wait because of long delays to see the content, what good is the content if never viewed?
The bounce rate goes up. Credibility drops as an authority site.
Speed affects SEO indirectly over time. Not directly. Not instantly.
So, “If tools such as Google PageSpeed Insights return a metric of 95+ consistently, is this a leading indicator of a site that is performant?”
Performant means functioning well or as expected. This test is not an indicator of “performant.”
It doesn’t even mean “good enough.”
As mentioned, we’ve seen horrible slow sites fake out this test. We never use this test to check page speed. Nor do any other professionals we know. Paid optimizer services gladly guarantee a passing Google PageSpeed score (an easy out). But not actual load time in milliseconds (real hard work). They know how to game the test. So do we.
But we’re not playing that Google game.
You are experiencing the frustration of the PageSpeed Insight test. It’s a gas-lighting ploy. Google makes you question your sanity.
I understand that Time to First Byte is a constraint or limitation that the server is experiencing or exhibiting.
I am assuming TTFB is a key indicator in improving page speed.
If a site is highly optimized and the server is the only remaining factor, I presume that this impacts page ranking. In reading through your articles though, it seems to indicate otherwise.
If tools such as Google Page Insights return a metric of 95+ consistently, is this a leading indicator of a site that is performant?
Should I avoid tools such as Pingdom, ByteCheck, GT Metrix, Google Pagespeed Insights if they’re not an accurate measure of speed?
We want to locate and correct speed issues. The waterfall many of these tests produce is most useful. You can examine what is being loaded. We can then do value analysis on the components.
Borrowed Industrial Concept A manufacturing method is a value analysis to streamline processes and components. It includes:
We use these methods to optimize websites.
“Value engineering began at General Electric Co. during World War II. Because of the war, there were shortages of skilled labour, raw materials, and component parts. Lawrence Miles, Jerry Leftow, and Harry Erlicher at G.E. looked for acceptable substitutes. They noticed that these substitutions often reduced costs, improved the product, or both. What started out as an accident of necessity was turned into a systematic process. They called their technique ‘value analysis’.”
Online speed tests are relative – not absolute. Using the same test, we watch (not measure) improvements from our site changes. Reducing requests is a false goal. It doesn’t always improve speed (load time) in milliseconds. In fact, often concatenation (minification plugins) break your site (unstylized HTML).
Relative changes on small numbers often look big.
Relative changes on big numbers often look small.
Absolute changes on small numbers often look small.
Absolute changes on big numbers often look big.
Explore both types of changes when looking at data.
A browser timer or stopwatch is a valid tool to measure speed for its location geographically. How does one measure for sites that have customers globally?
Good question. Again, you can only use approximation testing techniques. WebPagetest.org allows for selecting many different locations and browsers.
Pingdom.com also has a selection of geographic regions.
That is enough. It’s not accurate but results are repeatable. That means you can detect improvements from changes.
Assuming I’m on the right track, if speed is a primary blocker for a good user experience, is a stopwatch the only method of measuring it?
We rely on Pingdom most and WebPagetest.org second. Only because Pingdom gives faster results. Pingdom is a best-case scenario and WebPagetest.org is a worst-case scenario. Pingdom results are always faster. WebPagetest.org is always slower. Different methods of timing.
The browser timer is our acid test of the test. We trust it more.
You can learn about the differences between the two test above on our Site Tuning services page:
If yes, are tools such as the ones found in browsers that offer a measure of network speed also inaccurate and ineffective?
We have no evidence. Repeatable results are most important. If there is a conflict of results, I always trust the timer solution most.
If actual load time is the key determinant, what factors play a role?
Page weight is the total sum of all files creating a browser page. This is usually expressed in “k” (kilobytes). Page weight is an indicator of optimization and speed.
Page weight is key for mobile user experience. Mobile speed is 2X to 3X desktop speed. Much slower. The image shows the time tolerances of impatient visitors for browser response time.
If speed indirectly affects SEO over time, are the only factors content and design of a site (aesthetics)?
Borrowed Science Concept Hurdle technology ensures poisons in food products are eliminated or controlled. Stuff like mold, bacteria, and fungi. For food products, hurdles are pH and water activity measurements. These approaches are hurdles the pathogen has to overcome for safe food production.
In the end, “how fast is fast enough” is human perception of waiting or impatience.
We evaluated 35 theme candidates. All are available as free downloads from the WordPress theme repository. They came to our attention via email newsletters. Some are new and others are older but still popular. The alphabetical list is at the bottom of this page.
Here we report some theme trends that affect website speed and how to evaluate what is important.
Remember, our goal is a two-second, home-page load time using WordPress. The lighter the theme overhead the more headroom we have in our performance budget for adding things like images, forms, and other features.
There are certain cues and clues that help us evaluate a theme for speed potential (fast page load) – without installing it first. The download package size is number one on our list. Any theme download that is under 1M (zipped) usually is a sure bet for speed. But larger than that compressed file size doesn’t always mean a speed failure. You have to dig deeper by downloading the package and examining the contents. The 1M package size is a quick-and-dirty selection method.
A compressed theme package may contain many resources – or sometimes non-features. When we say non-features, we refer to fluffy bloat intended to deceptively entice website owners. Theme authors include these things to make the theme appear “feature rich.” But most features have a speed price.
Modern WordPress Speed Traps
When evaluating themes for speed, you need to consider the following:
Google Fonts (or worse the heavier Typekit) will slow down a website. It’s now rare to see a theme using fast-loading system fonts. Die-hards argue that most Google Fonts are now present in all browser cache. There is no way to prove this with present online testing tools. We estimate a 100 millisecond slow down uing Google Fonts instead of websafe fonts.
Fontawesome icon font – or an equivalent such as genericons or glyphicons. Icons are a popular inclusion. The downside is they load universally on all WordPress pages and posts whether the icons are used or not. How much this slows down a site depends on variables. We feel icon fonts are justified if you leverage them in your page design. But usually, it’s better to dequeue the font in the WordPress functions.php file and simply use icon PNG image files. If you remove the font files from your server using FTP or C-panel, you will end up slowing down the site even worse with 404 errors. Fontawesome can slow down a site by 500 milliseconds or more – 1 second delay isn’t uncommon. But how bad it really is depends. More typically, you’ll see delays of 100 to 200 milliseconds. There may be some improvement by using Better Font Awesome plugin and loading from their external CDN. We’ve played with this in tests and it has some potential for time-saving parallel loading.
Sliders are common additions now. When they are built into a theme, they may universally slow down all pages – even when they are only used on the Home page. How bad this delay is depends upon the slider. But for example, Nivoslider slows down all site pages by minimum of 200 milliseconds – even when no slider is present. The most popular in the 35-theme evaluation was Nivoslider (5 instances). Other sliders included: bxslider, OwlCarousel, Flexslider, Sldr, lightslider, and few custom ones. 19 of the 35 themes used sliders – that’s 54 percent. Over half. Sometimes, a better strategy is adding a standalone slider plugin and then selectively activating the plugin only on pages where it’s needed. The best plugin to do this is Plugin Logic.
Every theme includes a screenshot that serves as an icon in the WordPress theme control panel. These are rarely optimized images. In this evaluation, they vary from 100k to 2.5M Jpeg file sizes. The typical size is about 600k. This extra theme package weight doesn’t slow down the theme. But it’s an indicator of the theme author’s attention to speed details. A bloated screenshsot is simply benign, but wasteful, overkill.
It’s common for many themes to include sample or demo images. Like screenshots, these are rarely optimized images. 16 of our themes included header or slider images for easy setup. But they are extra baggage in the package. Worst case we had one theme with over 5M of Jpegs. But more typically, they were around 200k to 500k. These images are not intended to be used. They will be slow loading.
Theme packages may include language translations. These can add from 100k to 600k. But only 6 of our samples had these files. They don’t slow down a site. They just increase the package size.
The final proof is installing the themes and doing benchmarks. But, as noted above, there are things that can streamline our selection process before installation.
Fatness and popularity: Site owners prefer bloat.
The pressure to inflate theme package sizes is almost irresistible for theme authors. This is evidenced by the popularity of two examples: Zerif-lite (2.7M zip) 100,000 installs and Hueman (2.1M zip) 80,000 installs. And the incredible, 7M-download Enigma theme with 30,000 installs. Most themes are lucky if they get 10,000 installs. The fatter the more popular.
So if a theme is popular with the most installs, you can bet it’s a slow loading theme.
The biggest reason to ignore speed scores is they don’t improve SEO or ranking. That’s governed by readable and desirable content. Are we preaching to the choir?
After relevant and quality content, what improves long-term SEO is user experience (UX). That is how a person feels when they use your website. Speed is a component in UX. So is readable content. Speed’s a hurdle or barrier. Some will never see your beautiful, clean site aesthetics – or read your riveting content. That’s a worst-case scenario. Visitors sense danger in their cynical and suspicious mind. Your slow page is a subliminal indicator of poor quality and apathy.
They assume your site is about to rob them – or steal their identity. Credibility is what overcomes those fears. Credibility is trustworthiness, expertise, and enthusiasm. But they won’t sit waiting. They need to feel it in an instant.
Speed then becomes a critical marketing differentiation. It’s how you feel when you walk into the lobby of a motel. Were you greeted at the entrance? Or did you wait while ringing a chime at the motel desk?
That’s *hospitality* and Hospitality Management is a career discipline. Speed is hospitality. You care enough to be present immediately. Instant attention.
So how bad is your speed? Is it above average? The internet norm is an 8-second page load. Ridiculous. What’s the longest people wait before 100-percent bailout? 10 seconds.
People are impatient. They won’t wait. Their expectation is 2 seconds on desktop and mobile. They begin bailing out at that point and all disappear by 10 seconds. That’s for the initial page they land on. After that, they’ll tolerate 3 seconds if they think they’ll find what they seek. That’s information scent. Speed is a strong precursory cue delivering *the right scent* — like with a motorized fan. No waiting for the yummy scent to waft through the internet.
The next erroneous assumption is “requests” are the best indicator of performance improvement. That metric isn’t valid either. A site with a ton of requests can load fast.
All that matters is milliseconds of load time first. And second, for mobile-users: page weight.
Do tests claim you have vital speed problems for mobile? Are you certain they are real problems that make a difference in profitability? Remember, scores are meaningless to us. Milliseconds load time is the metric. But even more than that is profit.
So let’s answer those speed questions:
1You can’t use CDN. That’s great. Not using CDN is a blessing in disguise.
2You want to use webP format? Why? WebP format only improves image page weights by 10 percent. Because of parallel image loading in browsers, this gain is insignificant. It’s not worth it. It doesn’t even translate into a 1 percent gain in speed (milliseconds).
3Not changing your theme or removing your pagebuilder is whimsical. It’s like saying, “Don’t do serious speed tuning – instead give me a speed miracle.” But you’re right, we charge for miracles. $500.
4What should you do instead? For better speed, get rid of the following popular heavy plugins or find substitutions:
“I’m using GeneratePress (or Astra or whatever) and I’ve heard that the Hello theme is a lot faster, do you think I should switch to the Hello theme?”
If you have to ask that question, you’ve given yourself away as someone who should definitely not try using the Hello theme. – Christian Nelson, friend of PagePipe
Hello Elementor? Don’t use the theme. Seriously. Elementor pagebuilder works with any theme just dandy. It’s not your theme slowing things down.
But 60,000 other people are using it! And it’s free. That means it’s good. Right? Sorry. Popularity is overrated. Your mother told you that. And 60,000 active installations is achieved by many themes that SUCK.
EXCEPTION:Don’t use Divi theme. Please. Ever.
Newsflash: It’s the huge graphics and popups and things like chatbox sliding around the page slowing down your site. Trendy cool features are non-features for speed. They are bricks, doorstops, and paperweights. If you can’t divorce yourself from these boat anchors, you’ll never be as fast as Google yearns. But when is their opinion on speed and SEO ever plain truth. Weasel words.
We *wish* browsers cached special web fonts. This would make fonts transparent and immeasurable in tests. One hopes jQuery is already in the cache, too. There are plugin workarounds for this. But they still aren’t as fast as no special font usage.
But the better conservative assumption is these assets aren’t ready. Webfonts never enhance performance. At best, they’re benign.
Why aren’t Google Fonts there in the cache? They are. But Google forces a 24-hour cache refresh of fonts – and reloads. Why? Duh? To gather data on usage. Is that spying? Maybe. The font files don’t change every 24 hours. Ridiculous.
We strip fonts unless a site is achieving under target speed goals. Then they’re not deleterious. But on principle, we disable them. Site users are oblivious to the font differences. Nondiscrimination.
A telling recent trend is new themes vying for mobile speed. Themes like GeneratePress, Tiny Hestia, and the Twenty-nineteen default. Out of the box, they load a mobile font stack defined in the style.css file. They don’t use Google Fonts. Fonts have a stigma for mobile. They consume unnecessary bandwidth and data limits.
Does it make prettier sites? No. Pretty on mobile is a moot point. Single-column, 350-pixel wide images or resized images are what’s served up. All that matters is legibility and readability. Color choices create mood.
How much “mood or emotion” can you muster in a small 5.5-inch screen space? Be honest. Small blocks of color whizzing by with the flick of a thumb. That’s a mobile user experience? More like roulette.
Onscreen fonts once again are back to the stone ages. Even though there are a few more mobile system fonts choices to add.
Stripped-down “speed themes” often don’t enable jQuery by default either. If you resist the temptation to install a slider, top-of-page button, or some other jQuery intensive plugin, you avoid the extra load. Themes removed jQuery before but were seen as esoteric or fanatic – and not well received. Now it’s a mainstream trend. The Herd wants faster font loads. They think speed might save their site. Speed never compensates for poor content.
Discarding font baggage and jQuery are nuances for speed. They make for fast themes. But you can destroy that in an instant with a few popular plugins. Then the gain is minuscule by comparison. Dwarfed by thoughtless plugin abuse.
Websites are pretty generic on desktop and mobile. They’re mere containers serving content on little screens. And that is what matters most: relevant content. Fussing about choosing Google Fonts is wasteful.
Today’s cutting-edge developers desire mobile speed more. Theme authors retrofitted changes in the “font” loading. The demand was high enough and offered the chance to sweeten the deal.
Google Font usage is now seen as unfriendly to speed and UX.
You don’t have to install a plugin to remove Google Fonts on these modern speed themes. The theme loads are transparent: meaning less than 50-millisecond load times. But if you buy the premium version, they’re no longer the fastest kid on the block. They don’t tell you that fact. They then are more average than special.
If you search the phrase “Essential WordPress Plugins,” you’ll get about 7.4 million results. They all tend to regurgitate suggestions for the same old plugins. Copycat content. No wonder the identical plugins keep getting more installs. Even when better alternatives exist.
Includes important tips for mobile speed without coding.
Sorting and testing all the new plugins is too much work. So people don’t test. They assume. The assumption is “popularity” is good. For plugins, that is usually decided by looking at the number of active installs. Active installs is not a sign of quality or performance. It’s a standard of herd mentality.
Herd mentality, or mob mentality, describes how people are influenced by their peers to adopt certain behaviors. Examples of the herd mentality include nationalism, stock market trends, superstition, and home décor. —Wikipedia
To Engineer is Human by Henry Petroskiis a book about engineering failures – mainly of buildings and bridge structures and airplanes during the 1980s and before. The main takeaway from the book is still applicable – and maybe even more so today: When technology or ideas are changing rapidly, there is never the opportunity to build a history or library of experience. This increases errors. Experience is what prevents accidents and disasters.
New upgraded versions of WordPress come out multiple times each year. And new plugins are being introduced at a breakneck pace. In 2013, 15,000+ plugins were in the WordPress plugin repository. In 2014, there were 29,000+ plugins in the repository. By 2015, the number was 35,000+. By Sep. 2016, over 46,000+ free plugins in the repository. And today, over 55,000 plugins. It’s difficult to stay on top of that rapid rate of change. It’s staggering.
To make a fast decision, it’s plainly easier to select from the most popular plugins – and consider that good enough. There are over 55,000 free plugins in the repository. And this doesn’t count any of the plugins available on GitHub where authors refused to go through the WordPress red-tape of acceptance.
You can choose any plugin from the TOP100 and from our experience it will be the slowest and most bloated plugin in its class. For example: #1 Akismet: 52M installs, #2 Contact Form 7: 42M installs, #3 Yoast SEO: 33M installs, #5 Jetpack: 28M installs. These are all heavy plugins and either directly or indirectly affect load time. We see these plugins installed on most slow sites.
Plugin popularity is rarely an indicator of good value. People assume they must be good. At one time, they were either the-only-game-in-town or repaired or compensated for WordPress deficiencies that later became solved with new WordPress versions. So even though the need for “repair” was gone or obsolete, the herd kept installing out of habit and myth. It became de-facto standard best practice.
Many recommended “essential” plugins have negative speed repercussions.
Our rule of thumb is: the more popular a plugin is (active installs), the higher the probability it’s a slow loading plugin. Why? We don’t know exactly why this correlates. But it holds up in our speed testing.
It’s the quality –not quantity– of plugins that slows down a site. Speed testing free plugins and themes is our specialty. Millions of herd-mentality WordPress plugins slow down the Internet, waste web resources, – and use up your precious time.
PagePipe.com (our blog) has 53 active plugins. It loads in under a half second in the USA and about 1.2 seconds for Europe (Pingdom.com). It can vary. That is using the cheapest, shared, old-magnetic GoDaddy hosting located in Arizona. No CDN. It will go even faster when GoDaddy updates to PHP 7.1 – but they’re running on outdated version 5.4. We share our server with 24 other domains. Why? We want to prove a point: You can use “speed strategy” rather than throwing money at load-time problems.
Our Mantra is avoid popular plugins. High number active installs means they’re the slowest.
We don’t know why “popular = bloated.” We speculate the plugin authors are content and apathetic to speed and quality. Popular plugins existed first and use old unoptimized coding techniques (obsolescence). They tend to get heavier with revisions instead of lighter (kludges).
The authors of old plugins don’t have competitive motivation to be lean for speed. This isn’t true for newer, less-installed, lighter plugins. Speed (load time) is now a desired feature we’re seeing more because of mobile devices. But fresh, fast plugins are not easy to find. There are 55,000+ plugins in the free directory. Wow! An ocean.
What is more characteristic of “goodness” is retention rate. That’s calculated by taking the active installs and dividing by the number of downloads for all time. A plugin with a retention of 20 percent is pretty good. If it’s 5 percent or less, it’s a danger sign. They were tried – and dumped.
Slow plugin’s download file size is a clue. Bigger files load slower. There are some exceptions – but they are few.
In our new Toxic WordPress, we present typical time-wasting herd plugins suggested on thousands of WordPress blogs. And we give you speed alternatives.
… having read your book and browsed your site I had installed pretty much every plugin that you warn against using! I’ve spent I don’t know how much money buying plugins … I’ve reassessed the plugin functionality I actually need and struck a line through most that I had installed; the rest I think I can replicate with the lighter versions you’ve educated me about. … I am hugely grateful for the help and advice on your site and in your book: it’s great to know that good things are possible with WordPress on shared hosting!
There’s no WordPress dipstick to check SEO levels.
If you’ve seen a reduction in page rank, the estimated time for Google purge is 6 months after correction. Yoast SEO has a plugin-addition fix (extra plugin) to accelerate correction with Google. We doubt anyone needs it. But who know what the truth is since there is no “future history” to analyze?
Can you switch to another plugin without disrupting SEO? No one can promise SEO performance. And if they do, they’re selling a wish. But switching SEO plugins is apparently less of a risk than past Yoast weirdness. So migration is possible to another SEO plugin.
Some ideas from those affected by the Yoast 2018 bug is Yoast enhanced page-rank was artificial fragility at best. It set users up for disappointment. Removing the Yoast “benefits” reduced the site ranking to it’s real position. There is no proof. But it’s like saying, “You can cheat but eventually all tricks will be ineffective or eliminated.” There is a plugin to help migrate the Yoast SEO database when changing to select alternative SEO plugins. The next most popular plugin after Yoast is All in One SEO Pack.
We’ve documenting why not to use Yoast SEO plugin. No SEO plugin provides provable benefit. Have you ever seen benchmarks ranking improvement? We haven’t. How do you measure SEO reduction caused by competition or change in market needs? Impossible. It’s about consumer behavior.
Ensure your content is based on intent and what users want to find when they search. That’s the bottom line.
SEO is now controlled by a machine learning algorithm called RankBrain. RankBrain is smarter at identifying patterns and penalizes unscrupulous actors attempting to game Google.
RankBrain is an algorithm learning artificial intelligence system (AI), used by Google since October 2015. It helps Google process search results and provide more relevant search results for users. RankBrain is the third most important factor in the ranking algorithm after links and content. RankBrain interprets the relationships between words. I suspect it has more importance to Google than we know.
RankBrain allowed Google to speed up the algorithmic testing for keyword categories. They now choose the best content for any particular keyword search. This means old methods of gaming rankings with false signals are less and less effective. The highest quality content from a human perspective is being ranked higher by Google.
Content frequency, recency and relevance were previously rewarded with good ranking. This isn’t the case anymore. Search engine users and search engines are now trying to find amazing content with great value.
Gaming with plugins has negative, detrimental value for future SEO.
Technobabble sounds like sophisticated language. But it’s incomprehensible techno-jargon. It conveys a false impression of meaningful scientific content. It’s deceptive, disingenuous, unfair, or nonsense. It’s a method of misleading with pure presumptuous rubbish. Meaningless technical language overwhelms and confuses the audience, masking the presenter’s dishonesty. It’s an indicator of propaganda.
Kinsta is a managed WordPress hosting provider founded in 2013. They say they’re obsessed with maximum speed performance. Hey! We’re obsessed with speed performance, too. We should be happy cousins. But we’re not. Why?
Because Kinsta expels technobabble. Spewing.
Technobabble bores us to distraction, frustration, and irritation. It smells funny.
Technobabble gives an impression the speaker knows things the audience doesn’t. Try decoding jargon. It’s then obvious it’s unclear, pretentious, and unacceptable. Even novice listeners detect careless technobabble as a sign of dishonesty and insincerity.
Technobabble doesn’t describe reality. The presenter picks things out or makes them up, to suit his purpose. Deceived by web myth, an innocent person repeats misinformation without intent to deceive others. A false manipulation or misrepresentation is perpetuated. Speed myth erupts from the mouths of supposed experts. Often with hidden agendas of secret or ulterior motives.
Ivory-tower pseudosciences within an industry convey confusing, misleading, or nonsensical ideas using technobabble. Multi-syllabic scientific jargon gives false impressions. It implies bold laboratory research and hard facts. Technobabble takes a simple concept and describes it in an overworked scientific manner. This masks its inherent simplicity.
Intentional technobabble convinces audiences the science explained is true. Even though it may not be. Serious people will accept a meaningless idea wrapped in enough impenetrable language.
So we sat through a 1-hour boring seminar on speed by Kinsta representative Brian Li. He spread many common speed untruths. We share the speed errors below. Keep reading:
Here’s what Brian had to say about speed and Kinsta benefits in the free hour-long online seminar.
PHP version benefits are puny.
First Brian happily told us, “PHP 7 speeds up your site 3x.”
This is wrong. It speeds up the PHP code transfers. Your website also has a ton of other web assets that completely overwhelm this meager gain. Things like heavy images, offsite scripts, advertising, email automation, chat boxes, sliders, animation, videos, etc. These are where the speed problems lie. Not in speeding up PHP. That’s insignificant.
Brian told us “MariaDB is a faster alternative to MySQL database used by WordPress.” But didn’t tell us how much faster or how to get it. Only that if you buy Kinsta Hosting, you get it free. Well, it’s included in the price anyway. We suspect it’s not much help.
MariaDB shows an improved speed when compared to MySQL. MySQL exhibits a slower speed when compared to MariaDB. With the Memory storage engine of MariaDB, an INSERT statement can be completed 24% faster than in the standard MySQL. The memory storage engine of MySQL is slower compared to that MariaDB. – SOURCE
Speeding up the database by 25 percent is fine and dandy. But MariaDB doesn’t speed up your website by 25 percent. It’s the same gotcha as PHP gains. It doesn’t speed up the worst heavy assets like ads, third-party scripts, or image loading. Bragging about this is specsmanship.
Disk Drive Type and RAM are important? HDD vs SSD Brag
Bryan said, “SSD like Kinsta uses are better than magnetic drives.” Why? “Because read/write times are better and fetching is faster,” he says, “A fast server is better.”
This is true. But does it make a difference in real load time? Nope.
Specsmanship is the inappropriate use of specifications or measurement results to establish presumed superiority over competitors. Especially when no such superiority exists. We also call this vanity metrics.
Sorry. We’ve never seen server options make any difference in actual load time. Mechanical spinners give the same poor quality Time To First Byte as Solid State Drives. That’s right. In real-world comparisons, the gain is unnoticeable. Vaporous specsmanship.
Brian Li then talks about Kinsta’s Fast RAM and caching benefits. And mentions that virtual machines are better. At Kinsta, you get this trendy stuff even on the starter plan. Wow! PHP 7, Maria DB, SSD, RAM, Virtuals. Sounds great! Doesn’t it?
Brian says, “These alone give solid performance.” You mean to tell us a garbage site with heavy theme and plugins will be fast with these toys. Sorry. Paying attention to origin optimization trumps this technobabble stuff always.
An eCommerce store on a cheap shared host with 1.7-second TTFB – can still, load in a 2-second performance budget. That’s right 300 milliseconds is the tiny headroom remaining to build the page. How? Don’t load fat popular security plugins like iThemes Security or WordFence. Or plugins like Yoast SEO or Google AMP.
Brian’s claim: “With Page Caching your site can handle 10X more traffic. But it breaks eCommerce, forums, and any interactive site.” That’s not very reassuring. And it doesn’t help non-cacheable cache. Let see? That would be third-party scripts like Google Analytics, Google Fonts, and Google Captcha. And many more.
Bryan is convinced NGINX and FastCGI on Kinsta do the world’s best caching. He claims we’ll see TTFB improve from 230 milliseconds to 138 milliseconds. Sorry guys. But that is only 92 milliseconds. Thanks, we’d gladly take those savings. But we can save 300 milliseconds by dumping Google Fonts with a plugin for Pete’s sake. Brian then tells us, “Fast TTFB is important for page ranking.” Yeah. We sort of agree. But it only makes less than a 1-percent difference in SEO.
Bryan gets excited about a future *someday* when HTTP/2 is standard fare for hosts. But for today, it’s sort of geeky and beyond the reach of common site owners. Yeah. Sure. You can pay extra and have the future today. But is it essential to get speed? We don’t think so. He talks about HTTP/2 and HTTP/3 performance boosts. He recommends migration away from hosts that don’t provide it. Of course, his recommendation is his employer Kinsta. No bias here. Both require SSL certification. That forces you to be *secure.* He thinks that’s great for the web. We aren’t so impressed. You can learn more here:
Then Bryan dives into concatenation and minification as if they are necessary. But he warns: “You might break your site.” Surprise! He then recommends customizing to rid conflicts. But sadly, he neglects telling us how that’s done. He notes that minification plugins don’t help HTTP/2 because of multiplexing. So why is he talking it about it? He just recommended HTTP/2 on Kinsta. We suppose he’s showing alternatives to Kinsta. He then recommends using Autoptimize or WP-Rocket plugins as helpers.
Here’s our free article on minification and concatenation:
Now the hour-long seminar takes a turn into saving load time reducing the page weight of heavy images. He says, “Using right format. JPEG not PNG for photos.” This is a basic truth and big error for novice website builders. But he doesn’t tell how to retrofit a media library polluted with fat PNG photographs.
He recommends using ShortPixel or Imagify plugins. Those aren’t our preferences but they work. He says to serve webP-formatted images to supported browsers. We rarely see much benefit from this Google-endorsed trick. It will reduce image sizes by 10 percent. But that isn’t significant because images load in parallel. Cloudflare Pro converts to webP. WebP is not supported by Apple. A downside is webP uses more disk space with duplicates. We never use webP format.
If you’ve botched your media library uploading huge PNG photographs instead of JPEGs, it can make a big difference when it’s fixed. Otherwise, image optimization doesn’t give as big of a speed boost as it used to. Why? Browsers are smarter about handling images fast. A lazy load plugin may solve many problems for you instead.
UPDATE: WordPress now incorporates lazy load in core.
This is a geeky thing to do. It was faddish for awhile after WordPress added the Heartbeat API. Have we ever done it? Yes. Some hosts can’t handle the extra server load. Bryan recommends pinging wp-cron.php with better frequency. High traffic sites are most affected. And of course, the reason he mentions this puny feature is it’s a default on Kinsta.
Why do professed speed experts recommend rewriting code to eliminate the page blockage rendered by scripts? Bryan recommends it. This is so esoteric, time-consuming, and costly. Few plugins help. Always it involves custom work. But the payback is so small compared to getting rid of popular multi-function plugins like Yoast SEO or AMP or iTheme Security. We’ve written about Async or Defer flag for JS loading. But we’ve found this more often than not breaks your site. Bryan also mentions inlining critical “above the fold” CSS styles. These are costly make-work projects invented by programmers and coders. Don’t go there! Have we ever done it? Yes. But we were building experimental pages loading in under 300 milliseconds. That is unnecessary overkill.
CDN the wonder band-aid – promising after-speed-damage repair.
Nothing makes us more rabid than speed gurus recommending CDN as a solution. This a weak excuse for building a cruddy bloated website. Bryan recommends CDN of course. But at least he mentions CDN may not help. CDN can slow down page load time. Bryan warns: “Don’t just slap CDN on by default.” We agree. If you have a database problem, CDN won’t help. We steer clear of CDN with origin optimization. That means not being sloppy.
Now Bryan goes out on a speculative limb. He recommends Query Monitor plugin to identify slow components. This is a complicated plugin for professionals. He then tells us if we don’t know what to do “hire a developer” for solving problems. Absurd focusing on the wrong target. These need coding solutions. Ineffectiveness.
Is Kinsta expensive hosting?
You decide. Their pricing per month is here. Have we ever used them for client speed repairs? Yes. The sites were so broken we couldn’t migrate or backup. They contained obsolete WordPress versions and stale plugins. Absolute nightmares. Updates weren’t possible without breaking the site. Kinsta moved those sites. They were then faster loading – but still, time-bombs waiting to explode. The site didn’t *get pretty* just being moved to a different expensive host like Kinsta. Postponing the inevitable implosion.
You can get under 2-second load times on shared hosting. Pocket the money you’d spend elsewhere.
Kinsta is not the only speed solution. Build a better site from the ground up. Don’t add unnecessary junk.
I got 100/100 in GTMetrix. Is my site now good enough?
Scores represent principles that make little speed difference for most user experience.
load time in milliseconds
page weight in kilobytes
the number of requests
Scores aren’t on the list of importance for evaluation. Even the number of requests isn’t super important because of browsers loading assets in parallel.
Speed scores are artificial or superficial criteria based on the original concepts of the Yahoo Performance team. In particular, Steve Souders the mastermind behind Yslow Score. Google pirated Sounder away to their team where he invented PageSpeed Score. It wasn’t innovative but a clone of what he’d done before. Now more extreme. (Impossible?) He coined the title performance engineer and page speed. At Google, he established the same obsessive-compulsive criteria that evolved into PageSpeed Insights test. He also influenced other ideas like introducing speed into the Google ranking algorithm to persuade compliance.
Google hand-waving about speed is self-serving. It’s about them making advertising money. Making the web a better place is a noble cause to get site owner buy in. Ads slow down websites the most and are unmanageable. So Google has us focus on other distracting trivialities.
Google never let speed influence page rankings more than 1 percent. In fact, it’s closer to 0.5 percent. But no one knows for sure because Google doesn’t publish this proprietary secret information. The idea that a “standard” exists influences myths regarding web speed. And pushes site owners to waste time attempting to achieve a Utopian ideal. It’s inefficient.
All online speed tests (Pingdom, GTMetrix, WebPagetest.org and many others) are interpreting Google criteria for goodness. It’s extreme or invalid to the point even Google can’t pass their own insane tests. This ivory-tower speed snobbery is out of control. In other words, it’s excessive and impractical.
The good news: scores are NOT used in the Google ranking algorithm. Only one speed parameter counts and that is Time to First Byte. TTFB is host-server dependent and beyond the control of website owners. They can only cherry-pick a better host with lower server overhead and change hosts. A good test for TTFB is ByteCheck.com.
Our best-practice suggestion is to take 6 consecutive TTFB readings and average the results. Assume your worst TTFB in milliseconds is more common than the average reflects. How does that affect your 2-second performance budget? Take 2000 milliseconds and subtract the worst-case TTFB. That’s how much time is left to load your best pages. Hurts doesn’t it.
Speed doesn’t have the same valuation on all pages. You don’t have to be under 2 seconds all the time. What? Not all pages have equal importance to visitors? There is the primacy effect or halo effects influencing how a site is perceived. First impression counts. After the first page experience, people are more forgiving if speed was initially exceptional. Your 10 most popular posts or pages need to be the fastest possible. Less trafficked pages can be a little slower. If you’re using WooCommerce cut yourself some slack. Relax the performance goal to 3 seconds.
Speed does NOT affect ranking immediately. It affects SEO over time. Speed improves the user experience (UX). That’s measured by user intent. That’s derived from metrics like dwell time, bounce rate, and return visitors. Google search then knows visitors are finding what they search for on your site.
A few articles help explain Google’s speed oddity:
Even Google.com scored average with this Think-With-Google test. Only 1 percent of websites have speeds of less than 1 second. Focus on website content instead. If you’ve made it under 1 second: Time to stop. Remember less than 1 percent of the Internet achieves this perfection.
Try not to use CDN band-aids for speed. If you are using free Cloudflare, it’s not good. It’s especially a waste of money if you’re paying. You can achieve the same results with free origin optimization. Avoid annual or monthly paid edge optimization.
On WebPagetest.org, my site’s TTFB scores a big red F. Does TTFB really matter?
Yes in a big way. You can’t get under 2 seconds load time if your TTFB is 1.5 to 2 seconds. TTFB is server speed overhead.
Mobile-first ranking is a different and undefined criterion. Its influence on your ranking is still a mysterious and long-term strategic plan by Google. Punitive shaming affects those not complying. Google experiments on the little guys first. They can’t afford to upset their biggest advertising client accounts with inferior rank changes. Google isn’t stupid – even when it appears they are.
“If you do decide to go for cheap WordPress hosting, you should expect your site to go down from time to time (since at $10 per month, you’re most likely sharing a server with hundreds of other users). Also, expect that most issues won’t be resolved all that quickly. It’s just how the numbers work out.” – Kinsta
This quote is from an article written by Tom Zsomborgi. Tom is the Chief Financial Officer at Kinsta, a WordPress hosting platform.
Update: Our store is now hosted on Rochen. $4.95 per month. We aren’t an affiliate. Why did we leave BlueHost? For a better repeatable TTFB of 600 milliseconds.
We agree with Kinsta mostly. Their article about cheap hosting exaggerates some things. Like on BlueHost supposedly PagePipe’s store is the only domain on the server. They didn’t promise us that. Does that give us phenomenal speed – no. The TTFB is 1.7 seconds sometimes. That means PagePipe loads pages in under 300 milliseconds. And those are Easy Digital Downloads store pages. Every page reloads with an Ajax request! Boo.
And cheap hosts go down rarely. What do you expect for so little money? Most brag up times of 99 percent. And it’s often true. Some of our GoDaddy issues are resolved amazingly fast. And we mean hard technical problems. But we have a low expectation. We also think GoDaddy is a smuck for charging for SSL and privacy. Robbers. But we don’t buy that stuff from them.
Update: GoDaddy now offers free privacy. It formerly was $12 per year per domain.
We tested a client’s site on BlueHost with the same conditions as our store. He gets 500-millisecond load times – with a TTFB of 100 milliseconds. His SSL loads in 100 milliseconds. How? He has no clue. And neither do we. An Act of God. He pays the same as us. Go figure.
We’re not an advocate of cheap hosting. Whenever we can, we get clients on the most expensive host they can afford. It makes our life so much easier for obtaining speed. But not everyone can do expensive. And high price doesn’t translate into good always. We help resourceful site owners, too.
We walk the talk to show others how cheap-by-choice works. Severe self-imposed limitations.
It’s weird how good hosts and bad hosts are all bad and good at some time or another for speed. There is little consistency (repeatability) in performance between domains on the same hosting – pricey or cheap.
Now we have to consider this “cheap-hosting” article was published by Kinsta. Oh, they’re a host! They sell competing services. Would there be any bias against cheap hosts (since they’re not cheap)?
Have we ever put clients on Kinsta to solve speed problems? Yes. Do they always solve speed problems? Nope. Sometimes they make them worse. Go figure. We’ve found this to be true for every host. Variances. Fluctuations. Unpredictability. Voodoo. But we’re not paying for repeatability. We’re not relying on them to do a good job. We expect them to do a lousy job. We build accordingly with origin optimization. No tight tolerances.
The article’s testimonial is by Joe Hanley. He now hosts his site (https://www.audiblegenius.com/) on Acquia Hosting – not Kinsta. The Acquia homepage opens in 8.83 seconds in our timed browser. Lots of clutter on screen. Wow! Can’t wait to sign up with them. Built for Drupal? What?!
We bet Joe moves his site a few more times. Four times just isn’t enough.
Doesn’t Kinsta specialize in WordPress – not Drupal? Odd testimonial choice. Oh, Joe’s a software developer. No wonder.
So yes, if you’re a novice web virgin, you’ll lose a lot of money on cheap hosting trying to make things better. All a waste.
Conclusion: All hosting sucks – at sometime, in some way.
DISCLAIMER: We have no vendetta, grudge, or beef with WP Buffs. They seem like nice people.
But, WP Buffs baited us with email offers to sign up for two juicy PDF downloads. Curiosity got the best of us. Does our advice – about free plugins and themes – address the exact same problems?
Both WP-Buffs ebooks are 7-pages, color, and A4 size. Nice graphics. They tell you the same stuff we do. But they left something out: the good information solving the problems listed. Oh, we see now. We buy their stuff and it’s supposed to solve all our problems. How much? Read on:
The biggest difference is the price of their recommended solutions. In PagePipe articles – we focus on free ways to get speed and security. Let’s examine the two WP-Buffs colorful downloads below:
“The 21-Step Checklist to Ensure a 99.9% Secure WordPress Website”
The promise: WP Buffs implies reading these PDFs makes your site 100% secure. They don’t mention it costs you $1,553 US dollars.
This is the checklist below. You add these features purchasing the three recommended paid plugins or paid services. WP Buffs then receives a sales commission or finders fee.
Daily Cloud Backups
Force Secure Passwords
Daily Malware Scan
Brute Force Protection
Install a Firewall
Custom Login URL
Daily Link Scan
Blocking Fake Google Crawlers
Comment Spam Filtering
Daily Plugin + Theme Scan
Authentication Keys + Salts
Daily Database Optimization
DNS Change Alerts
Verify Trusted Sources
Manage Inactive Plugins
Note: These security features sound wonderful and amazing. They’re overkill. These practices are notorious for server overages and resource consumption. They slow down both TTFB (time to first byte) and page loads. Please, don’t do it! Save your hard-earned bucks.
WP Buffs publishes 3 affiliate links for you to click. WP Buffs earns a commission if people end up buying the advertised service or product. This sales tactic is also called revenue sharing. It’s a way to make a profit.
Their affiliate link solutions for security:
iThemes Security Pro plugin. $588 per year annual rent.
Malcare malware detection plugin $96 per year annual rent.
WP Security Audit Log service $89 per year annual rent.
TOTAL recurring annual fees: $773 US
WP Buffs makes an offer to secure your personal website or a client site as an extra paid service. They ask you to set up an appointment call. The goal is to sell you a maintenance and service plan for $780-per-year annual rent.
WP Buffs has affiliate links for managed WordPress hosting services on its website. We do not recommend any of the 6 they mention for speed. They’re expensive and you can get equal speeds for less elsewhere.
“The 12-Step Checklist to Achieve Loading Times Under 1 Second”
The promise: WP Buffs implies reading their PDFs makes your site load in under 1 second. They don’t mention it costs you $1,345 US dollars.
This is the checklist below. Add all these features by purchasing the three recommended paid plugins or paid services. WP Buffs then receives a sales commission or finders fee.
We wanted to know how they achieve these cool things. But alas, they don’t tell you how – only what you should buy. To get the things done, you must hire them – or click on an affiliate link in the PDF.
Our delight fizzled. This is mere bait. No answers at all. Not knowledge – a mere sales pitch.
Their gold-standard metrics: PageSpeed scores. An absurdity we despise. It’s load time in milliseconds that count for us.
And some stuff they suggest is contradictory. Many ideas presented won’t make your site more secure or faster. At least, not without a fat price tag attached to it. Actionable recommendations are affiliate links to premium paid plugins or services. As speed engineers and web developers, we don’t appreciate this sales tactic.
It’s a fluffy self-promotion piece. Marketing hype. We wanted technical solutions. Dang, it! We’re disappointed. Same old hype from other blogs. We wanted to learn something new or clever. Sadness engulfs our speed souls.
So if you want to find out how to achieve these marvels for free start by reading PagePipe. Visitors often binge read our entire site content. Are you obsessed with speed? Join us.
CONCLUSION: Keep your money in your pocket. Learn how to use duplicate plugins that achieve the same WP Buffs goals for free.
“If it ain’t broke, don’t fix it.” We couldn’t agree more. But we know, the WordPress world breaks on a recurring basis. That’s the price of an open-source community. Things become obsolete or incompatible.
Looking into our crystal ball we predict some near term WordPress trends and how they affect sites. Presently, here’s our status:
A common blog has 170 posts and 21 pages. And 32 active plugins (the WordPress average is 25).
There was technical bumpiness (total panic!?) during past core updates. Many major themes formerly used homepage widgets for customization. Upon WordPress new customizer addition, widgets moved without prediction. These errant themes became non-complaint and had to change their code to match new WordPress standards.
The new standards make things more bulletproof. But actually nuked everyone using old standards forcing emergency compliance. We recovered from that heart attack. WordPress makes changes now by adding a Customizer for CSS code normally stored in a child theme. Child themes are considered obsolete – but there is no documentation saying this is so.
We recommend using Simple CSS plugin to safely add customized code to your CSS. It won’t be overwritten by theme or core updates.
Presently, custom CSS is stored in the Theme Customizer > Additional CSS. This is fragile if the theme is changed. The way around this is to transfer the code to a nice plugin called “Simple CSS.” It acts in the same way, but is protected from theme changes. It also doesn’t cause an extra call like a child theme plugin. It’s faster loading. But CSS code isn’t always compatible in every way when themes are changed.
We don’t use child themes on new sites any more.
On the horizon looms the Gutenberg WordPress changes. We’ll get nuked again. But there are workarounds as stopgaps to get us into safe territory and through the learning curves. Gutenberg has already proved and predicted not being compliant with over 8,000 plugins in the WordPress plugin directory. Is there a published list? No.
Are paid plugins better for Gutenberg compatibility? No. 15 percent of those are predicted by Gutenberg developers to fail.
In the past, WordPress maintained backwards compatibility with legacy themes and plugins. This will not be the case with version 5.0 forward.
Gutenberg introduces unknown and unforeseen problems. They claim presently they will force the interface over the top of the traditional editor. It’s a god-like power play.
Don’t fear. Plugins are already created to defeat Gutenberg when automatic update occurs. Postponing Gutenberg hassles is a strategy improving the return on investment for existing sites. We’re talking increasing shelf-life or longevity. We can make changes when we’re ready instead of having them shoved down our throats.
The life span of a typical website is 3 years.
One can’t know for certain but some of your favorite plugins may fail. Authors may chose to repair them or we may have to find substitutes. There are 55,000 plugins in the repository so we’re not worried about fixing it. We’re only worried about “breakage events.” We assume it won’t be plugins with active ongoing version updates. But we don’t know. No one knows. Just because a plugin is stale or abandoned by it’s author doesn’t indicate potential failure.
“Gutenberg has three planned stages. The first, aimed for inclusion in WordPress 5.0, focuses on the post editing experience and the implementation of blocks. This initial phase focuses on a content-first approach. …
These foundational elements will pave the way for stages two and three, planned for the next year, to go beyond the post into page templates and ultimately, full site customization.” – Source
The goal of Gutenberg is to become a site builder (full site customization) and replace all page builder plugins. Fortunately, we didn’t go down the path of page-builder-plugin temptation. But millions of sites will ultimately be affected. Why is WordPress doing this? Do they want to destroy page builder plugins? No. They want to destroy WIX, Weebly, SquareSpace, and any other CMS competitor. Matt Mullenweg said authors and users of page builder plugins are collateral damage.
So here are our strategic recommendations:
1. Install a preemptive-strike plugin against Gutenberg. This is a precautionary safety measure and stopgap. It will add a years time to adapt your site.
There are presently 3 to 4 plugins to do this, but the one we recommend is:
2. Install the Simple CSS plugin and transfer the code in Additional CSS to it.
3. Twenty-nineteen theme deserves investigation for potential as a theme replacement for long term reliability and longevity. It is built for the Gutenberg editor. This theme is fast loading because the authors stripped it of features and external font requests.
The goal is future-proofing your site and improving the return on investment.
Now the bad news, we recommend a complete rebuild of your existing site. It can be planned for. We expect planning a switch in 2020 would be prudent. It’s gonna cost (again). Start budgeting now. Your site will then have a 3 to 6 year life (ROI) from that version. One never know exactly since the WordPress market is dynamic (chaotic).
“CloudFlare has released a new privacy-focused DNS service that runs on IP 22.214.171.124. They supposedly rotate logs every 24 hours and don’t store anything long-term. Seems cool, but I wish it did security filtering as well.” Link
This 126.96.36.199 Link told us who the real benefactors are:188.8.131.52 is a partnership between Cloudflare and APNIC. The speed project is at a research phase. Just like Google AMP has been for years. Both are ideas or concepts under test. 184.108.40.206 DNS services launched April 1, 2018.
APNIC is the Regional Internet Registry administering IP addresses for the Asia Pacific.
Many think PagePipe is technocratic. That’s anyone who thinks technology will save the future world. That’s a bad assumption. Ironically, we’re safety-seeking, risk-adverse late adopters. Or even laggards – when it comes to technology changes. That’s because “new” often means “buggy” – or worse a ticking time bomb. But our attention is now on this method.
What’s the downside of Cloudflare’s claims of reducing site speed by 54 milliseconds? No one knows yet. We’re optimistic skeptics. But we usually wait and see how things work for Guinea-pig innovators. We find it odd that Cloudflare brags about saving 54 milliseconds while at the same time boasting about how they converted the biggest chunk of early adopters to SSL certification. Using SSL/HTTPS slows down every site by 400 to 500 milliseconds. Speed hypocrisy!
Some international service providers are blocking 220.127.116.11. Why? That’s yet to be revealed. 18.104.22.168 doesn’t work in many countries because it’s blocked. What? Why would it be deliberately blocked? And in some cases, it’s not blocked but slowed down. Again, why? Odd mysteries to solve. 22.214.171.124 is plainly not a panacea … yet. Or maybe ever.
Cloudflare CDN publishes deceptive time-to-first-byte (TTFB) speed specifications. Because Cloudflare uses marketing weasel words, our level of trust is low.
Cloudflare is getting free PR and press. Proof of concept really lies in user testing. That’s where they’re at today. Testing on users. We wouldn’t adopt this technology. Our intuition says there will be revealed a hidden downside – or that this service makes little to no difference.
Cloudflare has low source credibility. They promote something for nothing often with a speed-gotcha embedded. From our experience and our clients, using Cloudlflare services increases site fragility.
Other elements have greater impact than Cloudflare 126.96.36.199 speed claims – like TTFB, SSL, heavy plugins, page builders, webfonts, email APIs, video, etc.
The 188.8.131.52 gain is the equivalent of disabling a related-posts widget plugin. Maybe.
So we’re watching and waiting.Benchmarking Cloudflare 184.108.40.206 services against Google’s 220.127.116.11 – and others like Quad9 and OpenDNS is the norm. Who is using those services? Geeks? Multi-billion dollar corporations? Certainly not non-programmers using shared hosting. None that we’ve ever seen anyway. We’re talking a difference of a millisecond per parallel-loaded request. Is this significant? Probably not.
18.104.22.168 is a distraction from speed fundamentals making a real performance difference.
If 22.214.171.124 makes a difference, it indicates the web page under test had too many calls (requests) in the first place. A bloated page always benefits most when optimized. How fast would the page be if built properly? Where do these DNS calls show up in speed testing? They don’t. They are smothered in the TTFB.
They aren’t giving us real benefits yet in language understandable to normal website owners. They’re using GeekSpeak.
For example, if a site has 24 calls. How much difference does using a special DNS make in real-world speed results? 24 milliseconds? We doubt it.
Why not eliminate 200 milliseconds by getting rid of a plugin like Social Warfare and stop linking to Facebook?
CONCLUSION This 126.96.36.199 DNS trick is misguiding site owners from true solutions: discipline, Pareto-based measurement, and value analysis of website components.
“I was thinking of using a theme like GeneratePress – or Astra. They seem to be the fastest but, now I’m unsure.”
Free Elementor is 61 milliseconds and adding paid Elementor Pro delays another 71 milliseconds. That slows down every page by 132 milliseconds.
GeneratePress (22 milliseconds) and Astra (36 milliseconds) are plenty fast. But only when using the free versions. If you buy the GP Premium plugin you add an additional 70 milliseconds. Buying Astra Pro adds 50 milliseconds to the free version. The theme authors don’t tell you those speed details. But that’s still minor compared to the real problem.
Should I use GeneratePress or Astra theme with Elementor for mobile speed?
For comparison, WordPress core loads in 215 milliseconds. Five times slower than your speed theme.
A typical free, discrete, single-function, no-setting plugin loads in less than 0.5 to 2 millisecond.
25 percent of plugin speed overhead is often consumed by one plugin.
80 percent of total plugin load time is burned by your 5 heaviest plugins.
The average WordPress website has 25 plugins.
Of PagePipe’s 70 total plugins, 12 load in under 3 milliseconds each. And 29 of the 70 load in under 1 millisecond each.
There’s usually one big-fat plugin killing speed – like WooCommerce 262 milliseconds – or more. Or perhaps Yoast SEO Premium plugin loading in 240 milliseconds.
We repeat. These plugin speed problems are minor compared to the real problem:
Speed killers: Undisciplined, novice site owners.
They’re the real problem. And some professional developers are apathetic about quality, too. You can do anything and everything dreamed of with a pagebuilder. That should be good. Right? But start adding incremental features and soon the site is overweight and slow. It’s not one slow thing – it’s everything. Egomania. Then site owners ask us, “Can you fix this Elementor speed problem?”
Sorry. It’s not Elementor’s fault. They didn’t make you put in all that web junk. They only tempted you.
WP Rocket is a $49 speed plugin with annual renewal (rent overhead). Most don’t realize WP Rocket caching plugin adds 109.1 milliseconds to global page loads. The irony.
Can you fix speed on a bloated Elementor site? The answer is probably “no.” Pagebuilder plugins can’t be selectively deactivated to reduce site heaviness. Not without white screening an entire site anyway. Discrete plugins can be selectively activated or deactivated on pages and posts. This is important for speed. It’s not the evil your pagebuilder does, it’s the goodness that got left out. It’s a sin of omission.
In fairness, Elementor doesn’t activate their widgets on pages where you don’t use it. But that’s not the same as selective activation. If you must use Elementor, don’t use it on your most-popular landing pages. Use good judgement for speed.
Desperate site owners throw money at speed problems. Usually by adding CDN, caching, or more expensive hosting. Things get costly. The speed investment is worse instead of better. Discarding whimsical features prevents speed waste.
How do I build a fast website from the start, without using a full-time developer?
What is the right decision about builder plugins?
Our speed advice is design without a pagebuilder whenever possible. Pagebuilder’s are slower, add more requests, and have a big learning-time commitment. But if you have no idea what you’re doing and are new to the game, go ahead – be a make-believe designer with a pagebuilder. It’s OK.
Pagebuilders are not the speed panacea you seek.
Will your pagebuilder site be slow?Most likely. The odds are high it’ll be slower than you ever dreamed. Why?
The answer: Because you own a rifle doesn’t make you a hunter. Just because you own a car doesn’t make you a racing champion. Owning a pagebuilder doesn’t make you a skilled web designer.
Using a pagebuilder doesn’t guarantee design quality. No surprise. There’s a pagebuilder learning curve. You still need to learn good universal design principles – aka best practices. It’s disappointing when your site is of low quality. You need a speed strategy before you start. Say, “No!” to dull, faddish fluff that doesn’t matter and adds no real value.
How do you design a website to be fast from the start? Building for speed is called “origin optimization.” It happens even before the project begins. It’s not an emergency, after-the-fact, speed repair. It’s strategic.
Here’s what to do for WordPress origin optimization:
1Get the best shared hosting you can afford. What’s best? Find a normal host allowing writing to the .htaccess file on your server. Special hosting – like WP Engine – won’t allow this. That ability is important for plugin speed tricks. Don’t choose SiteGround. Their wild TTFB fluctuates and is erratic. Their servers are worse than mediocre for TTFB.
2Choose a host with stable TTFB (time to first byte) on your server. Excellent is 200 to 300 milliseconds. Ordinary is 500 milliseconds and poor is around 1 second or longer. One way to find out is by testing the hosting company’s home page TTFB using ByteCheck.com. That’s the best possible it will ever be. Do at least 6 tests. PagePipe.com gets about 500 millisecond TTFB on GoDaddy (blog) and 1.7 millisecond TTFB on BlueHost (store). We host at these services to prove our point. You can get good-enough speed on cheap shared hosting.
3Do NOT install SSL certification – unless you’re doing ecommerce. You don’t need SSL for simple email signups. SSL handshaking slows down your site globally by 500 milliseconds average. SSL does NOT improve your SEO. There’s no proof. But you can measure the heavy toll on speed.
4Don’t put an email signup API (like MailChimp services) on every page. Have a single page with signup and use image or text links to that page. Use selective activation and only turn on your email plugin for that one signup page. Isolate the site drag.
5Use Twenty-seventeen default theme (48.4 milliseconds – after stripping Google Fonts). Live within its limitations. There are tons of articles online about how to customize Twenty-seventeen default theme. Why use it? Longevity. It’ll have an 8-year shelf-life. Don’t use Divi theme. It has a 1-second load time. Yes. That’s only the theme: Half your performance budget gone! Any theme is faster than Divi. Rather consider longevity a high value. Astra and GeneratePress are cool and fast. But they don’t have the potential longevity and risk-reduction of Automattic authorship.
NOTE: We’ve tested Twenty-nineteen theme for speed. While it is not as versatile as Twenty-seventeen, it loads in only 15 milliseconds. Dang fast.
6 Do NOT use free Cloudflare. It slows down your site with delays and 501 errors.
“Thanks for your time and feedback. You are definitely right about speed inconsistencies with free Cloudflare!”silvercoast-apartments.com, Portugal
8 When in doubt about some feature or frill, leave it out. What makes for a good website is content, not fancy things that move or animate. Like sliders, rotators, accordions, dropdowns, etc.
9Optimize your images with free “Imsanity” plugin. Other optimizer-plugin promises are seductive – and cost money. Don’t use free Smush plugin. Don’t use PNG format for photos. Use JPEG images and compress quality using your judgemental eye – and not a robot machine for a brain.
Can you survive without social media links? Do you have to have comments? Are you using Avatars (Gravatars)? These extras slow down your site and add little value.
Don’t use heavy, globally-loading plugins like Contact Form 7 plugin. Don’t use Yoast SEO plugin especially the paid version (or any SEO plugin). Don’t use a multi-function security plugin or any multi-function plugin. Stick with discrete plugins. Use more plugins – not less. Doubling the number of “good” discrete plugins will halve your load time. That’s right. 50 plugins are better than 25 if you choose the right plugins. Our ebooks are about this stuff. And the PagePipe blog is full of free plugin information. Buy both the “plugin alternative” bundle – and Toxic WordPress.
Paid Yoast SEO plugin is the speed equal to 250 discrete plugins. Bad site drag.
Accept that your learning journey requires frustration and failures. Nothing worth doing is easy. There’s a price. Pay your dues by investing in your brain power. “You” have future value.
BONUS TIP – If you use WooCommerce plugin, the best speed you’ll often achieve is 3- to 4-second load times. Reduce your expectations.
A conversation with Arisara Thitimoon about speed, Elementor, and Astra theme:
Arisara: Just purchased the bundle, I can’t wait to get into this!
Steve: Thanks for sharing your speed learning and web adventures. Also, thank you for caring about speed. It’s all about being kind to users and showing them you want to provide a good experience. They sense that when the site loads.
Arisara: First of all thank you so much. After days of researching why my site was !@#$%^*, I finally come across your article that helped me understand that a website built with bloated code is unfixable and is better off being built in a clean manner.
Steve: I wish more readers would realize this. I usually get requests to fix the impossible. Web miracles!
Arisara: What I do: I work on small projects $1000 to $2000 and so workflow speed plays a significant role in how much my hourly rate is effected. This is why I opt to use a pagebuilder (Elementor Pro). I use this in conjunction with Astra Pro and the combinations is damn near perfect for my type of projects.
Steve: This is a good combination. We’ve used it before and we’ll use it again on some client sites. But not all sites need these methods. Some can be simpler.
Arisara: But I’m not satisfied with performance … I’m learning that a pagebuilder needs to be used with respect and my goal is to master every little detail that allows to build a clean website that loads fast.
Steve: Value analysis is something we borrow from industrial manufacturing. It’s a discipline to improve profitability and efficiency. It consists of 5 things: combination, simplification, elimination, standardization, and substitution. It’s attention to details and creativity combined. Those use two different sides of the brain.
“Respect” is the keyword. If you use strategy, doing value analysis on features and functions, you’ll build a fast site. Assuming your host doesn’t have a TTFB of 2 seconds or more. But many shared hosts have TTFBs below 1 second. 200 to 500 milliseconds is ideal. You can find out by using ByteCheck.com and type in the URL of the host’s homepage. Their speed will never be better than that. And of course, avoiding SSL if you can.
Arisara: My speed goals is 2 – 3 seconds per page.
Steve: Good goals. This is the performance budget. Everything you add nibbles into that budget.
Arisara: My speed question: What is your advice to an Elementor user for building fast sites and avoiding bloated code and crazy amounts of HTTP requests?
Steve: I steer clear of pagebuilders when I build solo. But my colleague, Matt Stern, uses this combination often. Elementor makes my speed job more challenging. If it’s a WooCommerce site then we’re fortunate to get under 3-second load times with Elementor.
Here’s a free plugin to help you create or restore the look of your multi-column homepage without Elementor:
It is authored by Tom Usborne, the developer of GeneratePress. While it is possible to handcode this “CSS3 grid,” the plugin simplifies everything. No pagebuilder needed or extra paid addons. Much faster loading solution.
Keep reading below for more ideas.
Arisara: So far in my research, I’ve learned to use Astra features as a first priority (so create headers and footers in Astra, set up typography, etc)
Steve: Yes. You are on target. Here’s a quote from Matt Stern, my collaborator:
Here’s what I like about Elementor and Astra, having used them on different projects.
• Allows you to customize the entire theme (not just specific pages), this means you can use a drag-and-drop interface to customize your header, footer, blog layouts and more.
• Allows you near complete control of woo-commerce pages. See custom woo product page here: https://naturespiritherbs.com/product/kelp-fronds/. Again, you may not need this feature yet, but it will be a game changer if you decide to update your product pages.
• Leaves behind “clean code,” which can be reused if necessary. Divi leaves behind shortcodes which only apply to Divi and can’t be used elsewhere.
• Designed to be both customizable and fast. The fast part is an important consideration that many themes (even popular ones) seem to forget. One way Astra does this is by making some of its features modular. For example, if you don’t need to customize a particular aspect of a site, say the typography or the blog layout, you can turn “off” the customization for that specific part of your site, keeping your site leaner. See more here: https://wpastra.com/features/#performance
• Good Woo-Commerce customization “out of the box.” While Elementor allows for more in depth customization that take longer to implement, Astra has a nice set of options that are quick and easy to customize. https://thetoolmerchants.com/ was customized primarily with Astra and only used Elementor for certain pages.”
Steve: It takes good planning and cautious testing to not enqueue jQuery with a feature or plugin. Astra says their theme loads in under 500 milliseconds. But our real-world research shows it loads in less than 50 milliseconds! It’s WordPress core that adds the extra load time – about 300 milliseconds.
Arisara: Use custom CSS in my customizer to set default styling in elementor (btns, section padding, etc) so I don’t have to restyle every single element
Steve: Good job.
Arisara: Ditch the gimmicky features Elementor offers.
Steve: Amen. Try to avoid installing the Pro version if you can. Not only does it increase annual overhead $$$, but it doubles the load time of the free version.
Arisara: Reduce unneeded functionality like carousels, etc.
Yep. Sliders suck for UX and everything else that moves. Use static images instead.
Arisara: Optimize all images. I do this using Shortpixels site (everything on lossy).
Steve: ShortPixel plugin is good. But it isn’t always the best alternative. It adds 30 milliseconds of site drag to every page – if you keep it active. Resizing and compressing in a standalone image processor offline is always best (Gimp or Photoshop, etc.). Machines don’t make good decisions. One size doesn’t “fit all” in the image compression department. It requires a human eye.
Steve: Absolutely. Meaning avoid “popular” plugins.
Arisara: Don’t put complicated things in footers.
Steve: Good. Like email signups (MailChimp, etc). Use a central signup page and link to it instead.
Arisara: Disable any theme options / plugin / widgets that are not active.
Arisara: Reduce animations.
Steve: Mentioned that. No moving elements.
Arisara: Avoid all performance plugins (leave them till last).
Steve: Yes. And some speed plugins help scores – not milliseconds. Like minification, caching, etc.
Arisara: Don’t use a CDN (at least not right away).
Steve: CDNs are indicators the site wasn’t built for origin optimization.
Arisara: I’m probably barely scratching the surface. But if you have any tips, I would HIGHLY appreciate your expert opinion!
I’m really getting passionate about creating high quality designs and I’m prepared to invest the necessary time it takes to learn how to do it. So excited to read your ebooks.
Steve: Be patient. There ‘s a learning curve and new vocabulary. But I feel confident you’re a talented person and will be building “good stuff” soon.
Is there a reason we don’t like Astra for longevity?
Astra is a fast theme with 1 millon+ active installations. That seems remarkable. And it is.
We know the theme (without WP core included) loads in under 50 milliseconds. Much faster than advertised. Bless them!
If you buy the pro version, it doubles the load time. If you buy their pagebuilder (templates?), things get slower and slower.
How many people support this theme? 45 according to their about page. And they’ve been in business for many years. That all seems pretty good. But is it better than WordPress theme authorship for long-term updates? No one matches that. Especially for default themes. Astra is NOT bad – but they don’t have the strong future of WordPress / Automattic. WordPress has a market valuation of over 1 billion dollars.
We use Astra on client speed sites. But not on our own sites, we try and stick with customized default themes. Not everyone codes. We can write CSS and figure things out. So it’s OK for us – but not everyone.
We include Astra on our *list of goodness*. Along with GeneratePress – who is only one guy and an assistant. But Tom Usbourne only supports one theme and that’s it. We also recommend Tiny Hestia and Twenty-seventeen default theme with Google Fonts stripped. They’re all about the same speed as Astra.
If Tom gets run over by a bus, GeneratePress is dead, too.
If all of San Francisco, USA (Automattic’s headquarters) fell into the ocean, miraculously the WordPress torch (core) will be picked up by others around the globe and carried onward. That’s a harsh incident we hope never happens. But it demonstrates how bulletproof WordPress is today and tomorrow.
We don’t hesitate to use Astra on a client site. But rarely our own. Hmm? What does that say about out opinion?
A site is in serious trouble. Seventeen of 33 plugins have package sizes above 100k compressed (gzip). The site’s pages slowly load in 29 to 47 seconds. All pages are dragging beyond belief.
Time To First Byte (TTFB) is a measurement of server delay. We call it server speed overhead. We subtract it from our performance budget. The budget is 2 seconds. Sadly, the TTFB for this site varies from 2 to 6 seconds. Do we have to change hosts?
Is this site bad?
Rotten! Horrific! One of the worst we’ve repaired. The client is on GoDaddy hosting. We use GoDaddy hosting. It’s never that bad. We ask GoDaddy to check the server. They say it’s fine and dandy. There are no other domains on the server. They claim junk code in the server htaccess file is slowing everything down. We check using:
What we should see 10 lines of code in the HTaccess file contents. That’s the standard WordPress code (below).
Instead, we see 23,376 extra lines of code.
Where is this garbage coming from?
We clean out HTaccess. Then iThemes Security plugin puts all that junk right back instantly.
iThemes Security plugin is writing to the HTaccess file as if it were a wastebasket. Incredible confusion. Is it the plugins fault? Unlikely. It’s a slow and fat plugin but it isn’t that ugly. We guess an “operator error.”
Someone made settings for this plugin on overload. Why? Paranoia from being hacked and fear of getting hacked again with malware. A knee jerk overreaction into the red zone. Amazing security! We’ve written about this nasty speed-hog plugin before:
Many popular plugins often do calculations and file data on the server. These delays don’t show up in a speed test waterfall. Find out how fat plugins affect your site by checking server resource usage in server Cpanel. But one sure indicator is time to first byte gets unacceptable. A good TTFB is around 200 to 300 milliseconds. Average TTFB is 500 milliseconds. And bad news is anything over a second. This site is 6 seconds!
So what’s the alternative?
Use fast-loading discrete plugins instead. These are free single-purpose plugins that usually have no settings. They load in around 1 millisecond and the zip folder size is definitely under 100k – more like 10k. You only install the features you need. One plugin per feature. You don’t throw in the kitchen sink with over-engineering and gold plating.
But isn’t using many plugins worse than using one plugin? No. You could load 250 one-millisecond discrete plugins instead of Yoast SEO plugin. Think about it. You don’t need but a few lightweight plugins to duplicate the fat plugins features.
For example, the AMP plugin mentioned above doesn’t help mobile speed. Instead it slows down the page by 400 milliseconds because it’s so complicated. We suspect it also loads down the server. We don’t need to replace or substitute this plugin, we need to remove it. That’s right. Get rid of that Google-Dog plugin. It’s not helping anything.
We get busy removing the bad plugins and replacing them with discrete plugins. We trash the following plugins:
Lazy Load by WP Rocket
What happens to speed?
Global page loads drop from 38.7 seconds average to 5.7 seconds.
We next pull the iThemes Security plugin and replace it with 4 discrete plugins.
Speed gets even better:
3.5 seconds average. 10 times faster page loads.
Can we make the pages load in under 2 seconds? Yes. But we need to remove the SEO plugin – and replace the theme and pagebuilder. That costs too much time, energy, and money for this client. So we postpone these recommendations until a future rebuild.
We now understand how popular plugins slowdown TTFB (server overhead). Hurrah! You won’t need to move from your hosting provider after all.
Some managed hosts blacklist plugins to prevent the installation of vulnerable and disruptive plugins. Disruptive means server resource hogs. These include WordPress Popular Posts, Broken Link Checker, and Google Sitemap Generator.
You may ask, “How do we reduce server activity during backup when the media library is super big?”
Good question because that can slow down TTFB, also.
We use three free plugins to reduce a server’s burden during backups:
1UpDraftPlus We set site automatic backups to happen every week. We know if we lost one weeks worth of activity it wouldn’t damage us too much. Sad? Yes. But not ruined. That one-week interval reduces the amount of time the plugin is hitting on the server. (6.6M zip folder download).
UpDraftPlus takes into account server throttling and potential shutdowns. It doesn’t cause overruns of server resources. It sends packages in segments and waits for the server to recognize it’s not under attack or overload. Then it sends the next backup package (zip folders).
Set the weekly start time to a day when you know you have the least site traffic. On smaller sites, with few changes, we update monthly. Do you really need daily updates?
We retain two backup copies on remote cloud services (free Dropbox). But quarterly, we download a backup to our computer desktop for archiving. Don’t save backups to the same server you host on except for disposable copies. Be safe.
2Exclude Image Thumbnails From UpdraftPlus Backups This small 1.6k plugin excludes WordPress generated image thumbnails from Updraft backups, saving space. The original, full-sized image is included in backups. If a restoration from backup is needed, a thumbnails plugin is used to regenerate thumbnails using the original, full-size images. The plugin works for all image formats. It includes both native and custom image sizes added by themes and plugins.
There are 4 WordPress default image sizes normally created by core: original, large, medium, and thumbnail. Every time an image is uploaded these are placed on the server. Server space is not an issue. Resource consumption during backup and restore is a potential problem. We’ve seen themes (like Enfold) automatically create up to 18 different thumbnail sizes whether they’re used on pages or not. This bloats the media library backup.
3Regenerate Thumbnails Regenerate the thumbnails for your image uploads. Useful when changing their sizes or your theme. Regenerate all thumbnail sizes for one or more images uploaded to your Media Library. We keep it disabled and use it only when needed. (79.2k zip download).
We’re testing the free Ajax Load More plugin. We’re not happy with the styling yet. But we’ll get that figured out.
We originally wanted an infinite scroll on PagePipe. But we couldn’t find a trustworthy and fast plugin then. We tested every plugin we could download from the WordPress plugin directory. And most plugins acted flaky, were slow, or didn’t work at all out-of-the-box. Ajax Load More plugin can be loaded on only the blog posts which is great. But we want to make sure it doesn’t do global loading (site drag) on all pages.
So now to get geeky and show you 2 desktop speed tests concerning this plugin:
1This test is a PagePipe blog post and has an infinite scroll feature on it. Notice item 9 in the waterfall. That is Ajax adding 754 milliseconds to the post’s load time. This is typical for Ajax. Note that TTFB is 1 second. Ajax calls affect server activity thus increasing TTFB, too. Fortunately, the rest of the page assets are ultralight. So the load time is 2.192 seconds. Not under our 2-second performance budget but close. Fortunately, much of Ajax appears to be lazy-loaded. But on mobile??
Notice TTFB — same server — 664 milliseconds. Load time: 1.585 seconds.
Ajax for the plugin is NOT being loaded. This is a good plugin. It’s been built properly. It uses selective activation. Ajax is always bad for speed (WooCommerce uses it heavily globally). But if you going to use Ajax, only turn it on where it is needed. Don’t leave the lights on in empty rooms.
Someone who loves infinite scroll, Jon Dykstra:
One of my biggest wins this year was increasing time-on-site across my growing portfolio of niche sites. Gains: 58% to 330% increases. How? It’s so ridiculously simple. I added infinite scroll to individual posts on my sites. With infinite scroll, as visitors scroll to the end of an article, the next one (in the same category) comes into view. Because my content is so awesome they can’t resist and keep reading. The result: visitors stay and enjoy my sites much much longer. The benefit: I can’t say revenue increased but as far as I’m concerned, time-on-site is an important metric for any site. Any time I can increase it, I do. These gains are astronomical. – FatStacksBlog.com
And to be fair, someone who hates infinite scroll, Fatih Kadir Akin:
Footer is a very basic unit of web-page anatomy, just like a header. Sites keep some detailed information and links in the footer such as phone numbers, addresses, and help and support links. If users are searching for this detailed information, they mostly scroll down to find the footer. With infinite scrolls, users can have a hard time trying to find the footer. Infinite scroll makes finding the end of the page impossible. Not being able to reach the bottom of a website can make the user stressed (which is not great). – LogRocket.com
“I’m trying to make my website as fast as possible. I want to learn the best method and technical know how. I already watched WPfaster.org video on udemy. But they use W3 Total Cache plugin. So many technical settings – and difficult.” – Adzalan Yanggang
Surprise! We’ve watched their video, too. We don’t agree with their shady speed philosophy. It cost too much. READ WPFaster review
They recommend W3 Total Cache plugin. It’s not a good choice. Complicated.
We recommend Cache Enabler plugin (20k download file size). And three simple checkbox settings.
What about Swift Performance Plugin? It has so many cool features. It’s a multi-function speed plugin. It’s compressed download file weighs a massive 2.8M zipped – and 7.4M decompressed.
Very heavy plugins usually consume database and RAM resources on the server host. With these specs, we’re not interested in Swift Performance plugin. We prefer using single-function discrete plugins weighing about 4k and loading in under 1 millisecond. With discrete plugins, we do selective plugin activation on a page and post basis. This form of conditional logic significantly improves fine tuning a site.
The main valued functions of Swift Performance plugin:
Gzip Compression < This is activated by default on host servers.
Remove Query Strings
Async Execute Combined JS
Proxy 3rd Party JS
Inline Small Images
Google Analytics Bypass
You can add these features with single-purpose plugins with zero settings (no Wizard needed). Some are only used during maintenance and could be deactivated. But many don’t make a difference in speed at all. Just scores.
The Swift Performance plugin backend has animated advertising! Ugh!
Swift Performance Lite adds 131 milliseconds of site drag to every page and post of a website. Equally, we can install 131 discrete plugin instead. That’s the equivalent of adding the database-intensive Yoast SEO plugin (free version). Paid Yoast is even worse – 240 milliseconds. For the 20 features listed above, it should be a mere 20 milliseconds maximum: 15 percent of Swift Performance sitedrag.
Swift Performance Lite adds 6.2 milliseconds per feature whether it’s used or not.
And last – but not least – Swift Performance Lite plugin nuked the front end of the test site. All we had were gibberish characters. Our guess is this damage was caused by either concatenation in the minification process – or some caching weirdness – or a plugin conflict. Anyway. Not a fun plugin to deal with. We uninstalled it.
So how about using the monthly paid Pegasaas.com speed service? They charge $9 per month.
You can do all this for free – with plugins.
PagePipe’s homepage normal load time is 1.8 seconds according to the Pegasaas test. With their service tweaks, it’s 1.6 seconds. There’s easily that much drift for shared-server TTFB (time to first byte). The Pegasaas service essentially makes no difference – or a theoretical 200-millisecond potential improvement. You can get that gain by simply disabling Google fonts with a setting-less discrete plugin.
On a well-optimized page, minification rarely improves speed – only scores change. And test score are meaningless. Caching and minification are speed band-aids compared to website origin optimization.
‘I got a 100% score on Pingdom, GTmetrix and Google PageSpeed.”
Scores don’t alter SEO page rank or indicate good speed. Concentrate on milliseconds of load time as a better benchmark. Test scores are esoteric tweaks that make no significant speed difference.
1Use a theme that’s fast loading. And we don’t mean “mediocre” fast. We mean faster than greased lightening. Built and tested for speed. Don’t believe theme author’s speed bluster. Test it – or don’t buy it. If you must use a page builder, we recommend Elementor (with caution) – or wait (forever?) and see what Gutenberg offers.
We prefer a free theme because they’re not loaded with features. Paid themes are usually gold-plated and over-engineered with non-features. Free speed theme recommendations include: Basic theme, Tiny Hestia, Astra, GeneratePress, and Twenty-seventeen default theme. Many don’t activate code baggage like jQuery or Font Awesome. You can strip them of anything lacking substance.
WARNING: The pro (premium paid) versions of the above speed themes double theme page weight. This is not super significant. But we find it annoying. They brag about the free version’s speed then don’t publish the additional drag added by the premium version. That’s an advertising sin of omission. So if you’re really into extreme speed (1-second or less load time on a shared host), use the free theme without the premium extras. That takes creativity.
Creativity is the inverse of dollars. C=1/$
Do you think these insignificant improvements? Think again. Speed theme authors are deliberate in removing non-features for mobile speed benefits. It’s unconventional and bold. If pages weigh 5 megabytes to 3 megabytes – or even 2 megabytes, they’re doomed to fail for mobile user experience. The goal is superb quality pages weighing 100k to 500k.
2 Add features using discrete plugins. Not multipurpose plugins like Jetpack or Yoast SEO. This means also living within the theme limitations. Keep It Simple, Stupid. The KISS principle.
3Install proven fast-loading plugins. Avoid popular plugins like Yoast SEO and Contact Form 7 and many others. This includes WP Rocket, which functions great, but adds drag. Yep. 32-milliseconds of site drag to every page. Yes – believe it – a caching plugin slowing things down while speeding things up – oddity. We build WP Rocket’s features with free discrete plugins. It takes at least 4 plugin – but adds only 4 milliseconds to load time instead.
Discrete plugins allow activating features where most needed – instead of globally.
4 Find ways to either not use Facebook or using it in a limited way. Reduce the load as much as possible. Is Facebook making you money? Be honest. Or do you dream it might?
5Abandon the grandeur of Google fonts. If they’re in the theme you choose, disable them with a plugin. Even though they only add 100 to 300 milliseconds. On a fast mobile site that’s a 30-percent loss. Google Fonts are stinky bad for mobile.
There are more non-surgical extras for speed we place into the fine-tuning, tweaking basket.
Do you suppose examining your site that images are the biggest problem? They’re not. It’s rarely the case any more. The two biggest factors are usually theme-related and Facebook. Even worse than much-hated, third-party ads – but barely. PS- We optimize your image library for speed.
We don’t hate Facebook because we’re demure introverts and antisocial. We despise Facebook for what they did to speed innocents. Dirtbag destroyers of velocity. Apathetic.
Thanks for trusting us to help improve your site’s mobile health.