[Layout Shift Metrics] Higher scores on lower volume pages

56 views
Skip to first unread message

Chris Graybeal

unread,
Apr 15, 2021, 10:58:41 AM4/15/21
to web-vitals-feedback
Hey Google,

I'm trying to understand the discrepancy in CLS scores I am seeing for a particular set of urls. Take this an example:

This url's desktop view appears to have a score of 0.59 according to Google Search Console
image (15).png

PageSpeed Insights lab data has it at 0.  Origin data is 0.15 and there is not enough field data to measure.  I'm wondering if the low volume of traffic to the page affects how its score is measured in GSC.


image (16).png


Other tools show a relatively low score:

image (17).png

Chrome's Core Web Vital's Plugin does pick up some shifts after pageload. However these are still minimal. 0.008
image (18).png


Speedcurve Synthetic 75th percentile tests also score it at 0.0005
image (19).png


Speedcurve LUX tests 75th percentile of the page type are 0.0003
image (20).png


We have addressed several of the CLS issues on the page that we have found through this investigation, but none of them ever came close to giving us a score above 0.15, let alone 0.59.  My hypothesis is it has something to do with the lower volume of traffic this page gets, since field data is unable to be pulled in PageSpeedInsights. 

Can you provide any insight as to why we are seeing this discrepancy?

Best,
Chris Graybeal

Chris Graybeal

unread,
Apr 19, 2021, 9:14:11 AM4/19/21
to web-vitals-feedback
I have just completed a second validation in Google Search Console and it is still declaring this group of urls as 'poor' with an Aggregate CLS of 0.57.  However none of the other tools available to me can tell me what layout shift is occurring.  Can you please provide some guidance?

Chris Graybeal

Philip Walton

unread,
Apr 20, 2021, 1:55:45 AM4/20/21
to Chris Graybeal, web-vitals-feedback
Hey Chris, I'm not sure I have a specific answer for you, but I did want to point out a few things to make sure they're clear.

Of the screenshots you've shown above, I believe all of those tools (with the exception of Search Console and Pagespeed Insights) only measure CLS up until the page has finished loading. However, CLS is supposed to be measured throughout the entire lifecycle of the page, so that leads me to believe the shifts you're seeing reported in Search Console and PSI are occurring after the user scrolls down or interacts with the page in some way.

Related to what I said above, if you're seeing layout shifts reported in PSI but you don't know what's causing it, here are a few resources we've published recently to help you track them down:


I hope that helps!


--
You received this message because you are subscribed to the Google Groups "web-vitals-feedback" group.
To unsubscribe from this group and stop receiving emails from it, send an email to web-vitals-feed...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/web-vitals-feedback/79d71e93-7e86-462f-a48a-9bfd90482788n%40googlegroups.com.

Chris Graybeal

unread,
Apr 20, 2021, 9:32:07 AM4/20/21
to web-vitals-feedback
Hey Philip,

I appreciate your feedback.  We are currently adding some additional metrics to track CLS shifts that happen after pageload.  However, I believe Chrome's Web Vital's plugin should measure shifts after pageload as well, no?

Best,
Chris Graybeal

Philip Walton

unread,
Apr 22, 2021, 2:56:58 PM4/22/21
to Chris Graybeal, web-vitals-feedback
I'm not sure what, specifically, you're referring to by the "Chrome's Web Vital's plugin", but I'm sure the answer is yes because all Chrome tools measure CLS the entire time the page is open (including the web-vitals JS library, the Web Vitals Chrome extension, CrUX, Pagespeed Insights, and Lighthouse).

The only catch to what I said above is Lighthouse will close a page as soon as its finished loading and all audits are done running. So while it does capture any layout shifts that occur while the page is open, it only has the page open for a very short window of time.




Reply all
Reply to author
Forward
0 new messages