Unexpected? Not really: new layout for PageSpeed Insights

Unexpected? Not really: new layout for PageSpeed Insights

If you're looking for PageSpeed Insights, you now get redirected to their new testing environment: The PageSpeed Insights layout has changed as of today.

And it's not really a coincidence for several reasons:

It's often in May or November

I've been tracking the PageSpeed Insights layout changes, and the bigger changes -we could say overhaul- always seems to happen in May or November. Remember the very early days of PageSpeed Insights when it wasn't using Lighthouse as testing engine yet? That was before November 2018. Then, after moving to Lighthouse, they had some changes in May.

And a new complete overhaul now once again happend in November.

Real visitors matters, so focus on field data

This might not come as a surprise for several reasons.

Lighthouse score doesn't matter for SEO

First of all, Google already told us that your Lighthouse score doesn't matter towards SEO. So, this means lab data should not always be your KPI. It can still be convenient to do lab data testing, but that's another story.

Core Web Vitals is based on real visitor experiences

Field data is representing real user experiences. Obviously only of visitors using Google Chrome. So, this might have been a hint already that real user experience data is more important. This makes sense: your very own visitors are the ones that should convert.

Just like Core Web Vitals in general, you would also see issues amongst experiences of... real users. Sure, both Core Web Vitals data as shown in PageSpeed Insights and Google Search Console data can be behind quite a bit. But data within both of these Google tools are actually coming from Chrome User Experience Report, which contains aggregated real UX data.

Complaints from the community

As a result, PageSpeed Insights received quite a bit of criticism from non-developers. For example, if field data is so important, why having the overall score -which is based on lab data- on such a prominent position within the report? This led to people not being able to correctly interpret what they were seeing.

This was food for thought and Google was already working on a new layout. Beginning of this month, Google gave us a sneak preview of how the new design would look like.

Pagespeed scores did not improve

I tried to see if scores actually improved with this change as well. It seems like nothing changed. Looking at some recent tests within the previous PageSpeed Insights versions, metrics and thus overall scores are looking quite similar (some variability is occuring though, but is quite normal when doing tests).

Moreover, there isn't any information on network throttling or connectivity being changed. But unlike before, PageSpeed Insights is actually telling is the type of network throttling. Before, they were just talking about a mobile connectivity.

For example, they are now stating:

  • Slow 4G throttling
  • Network throttling: 150 ms TCP RTT, 1,638.4 Kbps throughput (Simulated)
  • Browser location: Europe

Conclusion

In my opinion, this is confirming that non-technical stakeholders should be looking at the field data while for example developers and (technical) SEO specialists could be looking and using lab data for other purposes.

While lab data could be used to see changes between upcoming releases and deploys, field data should be used to see if expected changes actually took place for your real users as well.

This also means that faking the overall pagespeed score makes even less sense then before.