CookieYes? No thanks!

I just reviewed the performance of a customer’s WordPress site. Things got a lot worse he wrote and he assumed Autoptimize (he was a AOPro user) wasn’t working any more and asked me to guide him to fix the issue.

Instead it turns out he installed CookieYes, which adds tons of JS (part of which is render-blocking), taking 3.5s of main thread work and (fasten your seat-belts) which somehow seems to increase DOM size to over 60K elements in this case.

I knew cookie consent solutions can have a performance impact, but this? Thanks but no thanks!

Spectra 2.12 performance regression (update: due to AI assistant)

So I like and use (aka “ultimate addons for gutenberg”) on some of my WordPress sites.

If you use Spectra as well, you might want to be know that Spectra 2.12 has a significant performance regression, sucking in almost all of wp-includes/js/dist.

Here’s hoping this was not intentional and that it’ll get fixed soon 🙂

More info in https://wordpress.org/support/topic/2-12-significant-performance-regression/

Update Feb. 15th: the regression is due to the new AI features, as a workaround go to Spectra -> AI features -> Manage AI features and disable “AI Assitant” (and “AI Copilot” if you don’t need/ want that either).

AOPro 2.3: delay all JavaScript

As of Autoptimize Pro 2.3 there is an option to delay all JavaScript. Delaying can have a bigger positive performance impact then asyncing or deferring because AOPro will only load the delayed JS after a to be defined delay or (better even) only at the moment user interaction (such as swiping/ scrolling or mouse movement) is noticed.

Obviously when delaying all JS you might want to be able to exclude and that is possible as well; you might want add data-cfasync there for example as contrary to Autoptimize’s normal behavior for deferring, delaying ignores attributes like data-cfasync, meaning that if you want to leave such JS untouched you will have to add it to the exclusion list yourself.

Lastly; if your site’s above the content depends on JavaScript to render correctly (which has serious performance impacts), delaying all JS would result in the page not rendered entirely, so test and exclude as required.

AOPro 2.2: delaying iframes (and other improvements)

I released AOPro 2.2 today and the main new feature is the ability to delay the loading/ rendering of iFrames. Despite the fact iFrames can be lazy-loaded (using browser-native loading=lazy attribute), in reality this does not always work in WordPress, hence the new “delay iFrames” option on the Pro Boosters tab.

  • How it works: a delayed iFrame has no src but a data-src attribute. Upon user interaction (or a timeout as also specified on the Boosters settings) the src will be set. If the iFrame is a YouTube, Vimeo or Dailymotion video, AOPro will load the video thumbnail as a placeholder.
  • How you can activate it: by default no iframe is delayed, if you want to have e.g. YouTube video’s delayed add youtube.com to the “delay iFrames” option.
  • What’s next: this is a “minimal viable product”, but depending on user feedback I might incorporate some of WP YouTube Lyte’s functionality for video iFrames. Tell me what you think in the comments!

Apart from delaying iFrames, AOPro 2.2 also has;

  • new under the hood logic to improve the speed at which Critical CSS rules are generated and visible
  • if you are “removing CSS/ JS” then the logic change this at a per page/ post level has changed; if you want one page NOT to remove CSS/ JS, just enter none in the “Autoptimize this page” metabox. If you leave the field empty the default “removals” apply. If you put something else in there then that overrides the default removals.
  • The ShortPixel Adaptive Images JavaScript component was updated to version 1.1.

HTTP/2; do not aggregate? Well, maybe …

Ever since the advent of HTTP/2 people asked if they still needed JS/ CSS files to be combined and indeed by default aggregation is off in Autoptimize. But based on this interesting article by Harry Roberts, “bundling is here to stay for a while”. So when in doubt; test performance with and without aggregating CSS/ JS ideally both on mobile and desktop bandwidth profiles and learn which works best for your site! 🙂

When lazyloading iframes does (not) work (automatically)

WordPress has made some good progress to speed up site rendering the last couple of years and part of this is thanks to images and iframes having the “loading” attribute with value “lazy”, telling browsers these can be loaded later allowing other (more important) assets to take priority. Especially iframes can consume a lot of bandwidth (think embedded Google Maps or one or more YouTube videos), so the performance impact of lazyloading those can be very significant.

Unfortunately one cannot always rely on WordPress core to automatically make sure there is no performance penalty from stuffing your site with iframes. Here is a non-exhaustive list of when iframes will still delay your site:

  1. WordPress core does not always add the loading="lazy" attribute;
    1. if loading="eager" is set (which means load asap)
    2. if no width & height are set (as lazyloading iframes without those could cause layout shifts)
  2. Firefox (and some less important browsers) does not support lazyloading iframes even if loading="lazy" is set
  3. iframes in or near the “above the fold” part of a page are loaded immediately, even if loading="lazy" is set

Conclusion; show restraint when adding iframes; adding an image of Google Maps which links to (a separate page with) Google Maps is almost always as informative and the performance benefit of using an image instead of an GMaps iframe is huge. And when using iframes then consider using alternative solutions to avoid the performance impact (for YouTube you might want to give WP YouTube Lyte a try).

Autoptimize Pro: page caching (finally) in beta

I just released Autoptimize Pro 2.0 as beta and now AO(Pro) finally includes page caching. If you’re on Cloudflare (APO), if your host offers page caching at server level (nginx/ varnish or similar) or if you already have a page caching plugin then you don’t need page caching in AOPro, but based on discussions with current customers page caching is still an important missing piece on a lot of WordPress sites, so for those page caching in AOPro will be a big help to get better performance.

If you’re on AOPro already you can switch to the beta-channel yourself (Settings -> Autoptimize Pro -> Account) or you can ask me to enroll you in the Beta. If you’re not using AOPro yet maybe this is the moment to give it a go? You can use the nowwithpagecache coupon code to get a 15% discount 🙂

Want to know some Autoptimize secrets?

There’s a first for everything and so last week I did a presentation at a WordPress Meetup in Antwerp titled “Autoptimize: 5 secrets and an intermezzo” which at least I had fun with.

You can find a PDF export of the presentation here.

Questions go below, in the comments (or in the form on the contact page).

Autoptimize Pro 1.3; instant.page

New booster in Autoptimize Pro 1.3: instant.page, a 3rd party JS component that can significantly improve performance for visitors going from one page to another on your site by preloading a page based on visitor behavior.

Do take into account that it could increase the number of page requests as the preloaded page might end up not being requested after all.

More info on https://instant.page.

AOPro 1.2; delay JS, CSS & HTML as long as you want

Autoptimize Pro has “Boosters” to delay JavaScript (esp. interesting for external resources), CSS and HTML (which can be very impactful for long and/ or complex pages). Up until today’s release the delay was either until user interaction OR a set timeout of 5 seconds, but now you can choose the delay time yourself, setting it to e.g. 20s or 3s or -and that’s where things get a teeny bit shady- 0s to disable the time-out, waiting for user-interaction to load the delayed resources.

Setting the delay to 0 is a bit shady because at that point you are hiding those assets for performance tests which -although artificial- is likely to improve (lab test) performance scores. If you do use the 0s delay then do take into account that *real* users will still need to load/ render those assets and that that still may be a sub-optimal experience.