Google PageSpeed Insights Algorithm Changes Since May

[JesterMagic]

I check a number of different pages every once in a while in PageSpeed insights just to make sure the scores are in the green to help with user experience and SEO.

I notice for the same pages I tested back in May (after upgrading my website) for Optimization I was getting mid to high 80s for mobile and low to mid 90's for desktop. I have made very few changes to the template since then and I ran the tests again and notice on my page, on average mobile has droped about 10 points while desktop numbers have increased slightly (1-3 points). I even rolled back a few changes to see if it made the difference (it did not).

Optimization suggestions are the same as before and the stats for them have not changed since May.

Reduce Server Response Time
Eliminate render-blocking JavaScript and CSS in above-the-fold content (1 blocking CSS resources - All my CSS is in one file and minified)
Leverage browser caching (just for the Google Analytics file)

I am not worried about getting 100 or anything as I think it is more important to spend time on content but I would rather keep all pages above 80.

Anyone have any idea what Google change in the PageSpeed tool algo which change how Google views optimizations for mobile?

[keyplyr]

Anyone have any idea what Google change in the PageSpeed tool algo which change how Google views optimizations for mobile?

Nothing specific, beyond the consistent increased performance demands over the years.

We're still experiencing the Mobile-first (re)Index & just went through one of the Core updates. With AI playing a more significant role in the algo, I would expect an influence on the ranking factors reported in some of the Google tools.

I have seen much the same as you report with PageSpeed Insights.

[JesterMagic]

Yeah I have had a few mobile-first index notifications but not for my major site yet.

I may play around with added some css styles for the header directly into my templates to see if that ups the score a bit since it will be reducing the use of render-blocking css classes but I use the uikit framework and it may make more of a mess than I want. I understand the reason to inline the css but once the css file has been downloaded and the visitor visits another page it makes this optimization moot, actually it makes it a negative since it has added extra text to the page which has to be downloaded every time now instead of once with the css file (which is then cached).

[robzilla]

Data from the Chrome User Experience Report is updated daily, but you seem to be referring to the optimization score. The score is just a measure of how well you're following best practices, it has no direct relation to page speed; a page with a score of 40 can be faster than a page with a perfect score. Perhaps they changed the scoring, or you changed a small thing that triggers another optimization suggestion. If the measured page speed is good, I wouldn't worry about it. (Though reducing server response time is always a good idea.)

[keyplyr]

@JesterMagic - I agree with your statement about caching the CSS file. Google's tools (all of them) don't consider metrics beyond what they are set up to measure.

In fact, Google's tools don't deal with caching well at all. I use a strong cache at the server level and Google tools cite an error because they don't find it at the page level.

At some point, I would think the AI infused algo will consider some niches deal with a high level of return visitors. Pages within those niches should justly benefit from cached files