Updated: November 2019
There are no affiliate links on PagePipe.
You’ve done everything to make your site fast. You don’t want to start over and ruin your investment. You could buy expensive CDN services like https://www.akamai.com/us/en/cdn/ . But there’s no guarantee of success. You’ve maxed out and from here forwards there isn’t a good return on investment in tweaking speed.
At this point, the only reason to start over is if slow speeds affect your quantifiable sales. In other words, justify repairs or new construction for profits. Don’t tweak speed for mere speed’s sake. Do it only to increased profits.
The bad news: nothing will create further speed improvement without spending money.
WordPress is the CMS to work with. Not because *it’s best* but because there are so many solutions and people who can help. You can build a fast site with WordPress. But not usually straight out-of-the-box when loaded with images and widgets. It takes care and attention to details.
The speed problem usually isn’t WordPress. It’s the whims of the site owners. Or worse – committees!
We often recommend modifying WordPress default themes. They have longevity. They will have updates and support for years. Themes vendors can’t say that. There’s a shakeout in process among theme authors. That means mergers and acquisitions. It’s a volatile place. Businesses vanish overnight. WordPress won’t. Not soon anyway.
Simple WordPress-authored themes usually load in about 50 milliseconds. A twentieth of the time of premium themes. Even the ones who claim speed is their main feature don’t match that time. Premium themes authors can’t restrain themselves. They throw in the kitchen sink with gold-plated features.
We’ve seen a trend for theme authors to now offer two versions. One is the desktop, nice-looking version. The other is a stripped-down mobile version (usually removing Font Awesome and Google Fonts). The stripped down Premium version still isn’t as fast as free default themes for a foundation.
Even with a bare-bones theme, it still requires more removal (with plugins) to make it fast. These include removing icons and fonts and emojis. Baggage.
The more popular a plugin is – or if it’s a premium plugin – the slower it will load. It’s true in 99.9 percent of the cases. One popular plugin usually weighs more than the theme itself. Sometimes more than the theme and WordPress core combined – like Yoast SEO plugin.
You can duplicate the multi-features of paid plugins with free or open-source ones. Sometimes, you have to load several plugins to create all the needed features. But it’s still faster. A typical single-purpose (we call it *discreet*) plugin loads in under 1 millisecond. A discreet plugin usually has few or no settings – it’s plug and play. We can load 80 discreet plugins if we don’t use popular Yoast SEO free plugin. The load time would be the same or less. PagePipe has 70 active plugins with 6 on “standby” for maintenance. The homepage loads in under 1 second on GoDaddy hosting. We don’t use any SEO plugin.
But Yoast SEO has millions of active installs! That’s right. Poor fools.
Developers and site owners don’t understand plugins are often global loading. We call this “site drag.” Let’s say you install Contact Form 7 plugin’s shortcode on your contact page. You’d think, “Well, it only slows down that page.” That, unfortunately, isn’t true. Contact Form 7 slows every single post and page of your site by 42.8 milliseconds. That’s site drag.
Contact Form 7 has millions of active installs! Yes, popular. Let’s all follow The Herd?
Site drag is never revealed anywhere in plugin specs or read.me files. You have to test. Contact Form 7 is the same as adding Google Fonts or a large header image to every page. But the plugin author doesn’t tell you that bad news. There are ways to selectively deactivate plugins on URLs. This is a good trick to learn. It requires using more free plugins. But they don’t add site drag.
We test the site drag of plugins and report it. We always focus on the 20 percent of plugins that cause 80 percent of the load. This is Pareto’s law or the 80/20 rule. The worst, of course, are pagebuilders. Not all pagebuilders are bad. Elementor can be fast. The problem with pagebuilders isn’t the plugin itself. It’s the fact they encourage bloat. It’s too easy to add more features without design pain. The site then has organic growth instead of strategic growth. The pain comes after the fact when speed is finally measured.
Now everything we’ve told you for speed benefit can be undone by adding one third-party ad. Or a real-time social media “likes” counter. Ads and real-time counters are death to speed. You have no control over their connect time or weird delays. You can’t cache these assets. They don’t belong to you. They reside on other servers.
The more third-party assets you dump on your site – the worse it gets. They are unpredictable for delays. These include:
- Google Analytics
- Google Maps
- HTTPS / SSL server handshaking
- any offsite signup like MailChimp or iContact
- adding Google reCaptcha to forms or logins
- Google YouTube videos
- podcast services
- many popups like OptinMonster and ChatBox
- any API
- on and on
You have to figure out what you need for survival and dump the rest – or use some creative trickery. That’s value analysis. It’s an industrial discipline that applies to page speed optimization. It includes:
The biggest thing affecting speed is deciding up front what you’re not placing on your site. Then disciplining your team to stick to the goals. A performance budget must be set in advance. Can you simplify?
Image optimization opportunities are small.
Sometimes there’s an opportunity in converting non-transparent PNG images to JPEG format. You can compress JPEGs at a quality between 70 and 82. 82Q is the WordPress default. It’s the same as a Photoshop compression of 50Q. That “grade” passes the WebPagetest.org criteria of “goodness.” It’s a good benchmark. This will reduce page weight. Page weight reduction isn’t always proportionate to speed improvement. We measure speed in milliseconds.
Convert non-transparent PNGs to JPEGs to the WordPress 82-quality setting default.
Converting non-transparent PNGs to JPEGs (lossy-quality setting of 82) reduces the page weight. This classic error of choosing the wrong image file format is novice. But it happens all the time. PNG to JPEG image optimization changed a 6.5M, 15-second page into a 2.6M, 8.5-second page. Is that good? Hardly! Not when your goal is 2 seconds.
This is a free plugin that needs cautious use or you can ruin many things. Our advice is: after using the plugin on the media library, disable it. That will prevent any slow down. This is the same thing to do for any plugin that’s used for maintenance. They don’t have to be chugging server resources all the time.
Of the page weight, often 88 percent is images. All sites are image intensive. The average is 55 percent of weight is images. So you’d think, “Let’s optimize images and get better results.” What if those are already optimized images using a plugin service like Imagify.
Optimizing them more with “Ultra” settings only ruins them by making them fuzzy. We’ve tested it. It was worth a shot. But it failed to produce good-enough quality. It dropped the page weight but ruined the images. Did Ultra make a speed difference? No. The needle wouldn’t move. Image weight becomes transparent and gets a boost from browser parallel loading.
One goal is identifying the few plugins that add 80 percent to the total plugin load time. These are plugins causing *site drag*. Remember: Site drag is global loading of assets whether used or not on a page or post. Not all plugins do this but it seems heavy and popular ones always do.
Usually the more popular the plugin the worse it is for site drag. Why? We can only guess. It’s a correlation that is real and holds up time after time. *Popular* means an easy selection and fast-choice for functionality. It doesn’t translate into speed.
A fast (normal) plugin loads in one millisecond or less and doesn’t cause site drag. A heavy plugin loads in 50 milliseconds. 50 times slower. Some are much worse. It’s not the number of plugins that slow down a site but the quality. Imagine 100 bad plugins!
Remember it’s not quantity – but quality.
Ignore speed test scores. Watch milliseconds of load time. Accepting average 8-millisecond speed is not a failure. Excellence (perfection?) sometimes has too high of price tag.
WordPress core loads in about 500 milliseconds. The paid Humanity theme averages 1.1 second load time. Yes. Over one second. All stripped-down free themes load in about 50 milliseconds. Humanity theme is 20 times slower than a theme built for speed. The Divi theme which is notorious for slow loading beats that with a 679 millisecond load time. So not so bad, but still 10 times heavier than a “speed theme.”
Webfonts are heavy. A fast site may have 50k of fonts or best case *zero* (web-safe fonts). But some sites range around 250k to 300k in webfonts. Disabling these make a difference in speed. On a lightweight 1M site, we’d have lost 25 percent of the weight by pulling the fonts. But on a site with a 4M page weight, other page assets overwhelmed the font gain. It won’t matter.
The same goes for 250k weight of Google Maps. On a 1M page, it would be significant to remove it. But not on a 4M to 6M page.
HTTPS / SSL adds 500 milliseconds of site drag in the Time To First Byte (TTFB). Another slowdown.
The average web page today weighs 2.3 to 3 megabytes. Average-pages load in 8 seconds on desktop. A site weighing 4.6 megabytes is double the average.
Will CDN help? Not a free CDN like Cloudflare anyway.
Speed excellence is beyond the reach of misfit sites. They’re constructed with heavy themes, slow page builders, and third-party APIs. It’s not one thing or even a few things. It’s everything. Too many things.
What drastic changes would make a difference in speed? Complete site rebuilds. Is it worth it? Nope. You made a big investment getting these sites to where they are today. You must have a return on investment before you can overhaul the sites. They aren’t broken.
It’d be a rebuild from the ground up. That means:
- starting fresh with a host with a TTFB of 100 to 300 milliseconds
- avoiding HTTPS/SSL if possible
- selecting fast themes
- using fewer images
- using Google Maps tricks
- not using Google Fonts
- lazy loading YouTube videos
- no CDN
- no sliders
- focusing on written content for SEO
- no live counters showing social popularity
It would be Spartan but it would be fast.
That’s too big of a sacrifice. You don’t want to go there and we don’t blame you.
Our pragmatic recommendation is don’t add any more features and let the site make money. Only make big changes if you start losing money. Mediocre speed or even average is good enough if you’re making money. The most important metric is sales. It’s quality of traffic, not sheer volume, that translates into profits.
Nowadays, SEO is about 2 things:
- relevant content
- page titles that arouse curiosity
If you use WordPress, don’t worry about:
- machine SEO
- tweaking snippets
- keyword usage
Imagine people aren’t looking for what you offer. They don’t care. No amount of keywords or manipulation will change anything. SEO is about motive. What are the users motive to visit your site?
Market positioning is a creative communication strategy. It serves as a shortcut to the buyers motive. They won’t wade through junk trying to figure out why you are valuable or why they should care about what you do. They don’t have the patience. It’s a much more intolerant world.
Essential or core UX is also simple. UX is about overcoming three critical things for quality first impression (aka credibility):
Speed being prime. Why? If you can’t get past this hurdle the user won’t hang around to even see your cool presentation or offer.
2Next attractive aesthetics. We have an emotional reaction to what we see and determine in an instant if a site is “good” or “bad.” We base “stay-or-go” decisions on a 50-millisecond visceral design perception.
3And last, readability and findability (like navigation and text size, etc). Websites are still about reading content (or skimming at the least). People are foraging for entertainment or problem-solving. Pictures are nice. But it’s words that still communicate to humans and are also machine readable.
UX is that simple. Three helpful things – not thousands of tricks. And metrics (big data) are a tiny part of the evaluation. UX is about *feeling right* and being polite. What meter do you use for measuring that?
“Page speed” is a phrase searched for 10-times more than “performance optimization.” And mobile speed is synonymous to page speed in world-wide interest. For mobile users, page speed is a competitive differentiating factor.
Why the pain today? Because mobile devices flood the earth. Massive adoption is changing how people work and play. In remote third-world countries, smartphones are the main computer technology and communications devices. Don’t ignore small screens if you want your business to make money in new global markets.
Origin optimization is cumulative. 500 milliseconds here and 1 second there all add up.
To be in the top 1 percent of fastest sites requires drastic changes and self-discipline.
PagePipe contains many speed ideas. If you’d like us to help you on future web construction, we’re willing and (usually) available. Contact us.