I've had several clients come to me asking to help with their website performance, insisting that they must get a PageSpeed Insights score of 100.
I have quite a reserved opinion on this. Yes, Google PageSpeed Insights is incredible useful and lists a lot of easy performance gains but it's a waste of time trying to maximise this arbitrary score. The focus should be on things that improve the speed, and I often find these things aren't even mentioned.
Every website that has Google Analytics will see a message saying they should have a better caching policy on some resources; ie. the Google Analytics script. We don't even have control over that and they don't recommend inlining the contents of the script because it's likely to change. Nice one.
I've found their scoring system is very inconsistent. When testing on my own website, I noticed that "Desktop" lists a 0.04s TTFB as an issue but "Mobile" doesn't have any such issue listed. Sure, maybe the standards are different but a 40ms delay from the server is good. The "Desktop" tab reports a score of 100 but has some recommendations with red warning triangles. What does that mean? Why am I not deducted? This is where you start to spend a lot of time going down rabbit holes.
It's important to remember that it's just one of the many testing tools out there. Ok sure, Google rules the web and they may look favorably on sites with a higher performance score, but throw your site into one of the many others like Pingdom or GTmetrix and you'll get a new set of issues to spend time fixing. Then you spend precious time fixing those issues only to go back to the first site and see they've changed their scoring criteria.
In summary, I agree with what WP-Rocket says on the matter:
Google PageSpeed can be helpful as long as you don’t treat it as the be-all, end-all.