difference in LCP score in Search Console and PSI

[Tanya Tritelnitskaya]

Hello!

Could you please help me to understand the difference between LCP score in Google Search Console and LCP for real-user experience in PSI for the desktop group. As far as I read the source for the both reports is CrUX. For our site in general we have 79% good score and 2.1s in PSI, as you can see in the screenshot. But in the Console we have 4.1s for the our blog posts group. The point is that it is the pages of this group are the most visited on our site, that is, it seems to me that it is they who create not all, but the main part of these 79% good score. Obviously, I'm misunderstanding something, and I'd be very grateful if you could help me to clarify why we have such a difference. Thank you very much in advance! 

[Barry Pollard]

I'm not sure what URL you've put into PageSpeed Insights but think you have just put home page in, as when I do that, I get similar figures to you:

https://pagespeed.web.dev/report?url=https%3A%2F%2Fexistek.com%2F&form_factor=desktop

Note also it has shown the whole origin score (i.e. all the pages on the site) because it does not have data for the home page.

However, when I put the example blog post that GSC highlights it gets a much worse LCP score - and that is at page-level:

https://pagespeed.web.dev/report?url=https%3A%2F%2Fexistek.com%2Fblog%2Fsdlc-models%2F&form_factor=desktop

This suggests there is another group of pages, visited as often or even perhaps more often than your blog posts, that is pulling your origin level score back into the passing category. Does Google Search Console show any Good or Needs Improvement groups as well as the Poor blog group? Alternatively you can use Google Analytics to see the most visited pages on desktop for Chrome users to see what CrUX sees.

Thanks,

Barry

On Mon, 1 Aug 2022 at 08:39, 'Tanya Tritelnitskaya' via Chrome UX Report (Discussions) <chrome-u...@chromium.org> wrote:

Hello!

Could you please help me to understand the difference between LCP score in Google Search Console and LCP for real-user experience in PSI for the desktop group. As far as I read the source for the both reports is CrUX. For our site in general we have 79% good score and 2.1s in PSI, as you can see in the screenshot. But in the Console we have 4.1s for the our blog posts group. The point is that it is the pages of this group are the most visited on our site, that is, it seems to me that it is they who create not all, but the main part of these 79% good score. Obviously, I'm misunderstanding something, and I'd be very grateful if you could help me to clarify why we have such a difference. Thank you very much in advance! 

--
You received this message because you are subscribed to the Google Groups "Chrome UX Report (Discussions)" group.
To unsubscribe from this group and stop receiving emails from it, send an email to chrome-ux-repo...@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chrome-ux-report/7b07033e-5d54-4b2f-aab5-c58cce1e28fdn%40chromium.org.

[Tanya Tritelnitskaya]

Hi, Barry!

Thank you very much for your fast and clear reply!

Indeed we have another group of pages, that have green score, but the visit to these pages are much less, that's why I had a question. 

So I am right in assuming that PSI shows the average score for the entire site, even though the blog pages still have significantly more traffic?

Thanks!

[Barry Pollard]

PSI in origin view will show the visits to all page views on that site - weighted by number of page views.

This suggests your blog pages do not have significantly more traffic after all, as you seem to think.

Or alternatively some of your blog pages have good Core Web Vitals and are being grouped into the other pages?

Or your blog pages DO get most of your visits overall, but less so on desktop. Which is why only desktop has this overall passing score.

Thanks,

Barry

[Tanya Tritelnitskaya]

OK, thank you so much for you clear and detailed explanation, Barry!

Have a nice day!

[Tanya Tritelnitskaya]

Hello, Barry!

Sorry for bothering you again, but I'd be very grateful if you could clarify to me the difference in LCP score at GSC and PSI for the one of our URLs with LCP usage data, namely https://existek.com/blog/hospital-managment-system/. We get different data for the desktop as you can see at the table(where I put the results from PSI) and at the screenshot. At mobile the data is the same. FYI - the date at the table is the date, when I add the new line. Thanks a lot in advance!

   

[Barry Pollard]

Hi Tanya,

I'm afraid I can't talk to what Search shows in Google Search Console so that would need to be directed to that team. I would expect it to be very similar now they are exposing URL-level data in Search Console (as opposed to just the page grouping level data they used to show before) so not sure why they differ here, but this feature is pretty new.

Regardless of the slight differences, both show the LCP of that page is in the "poor" region above 4.0 seconds so both indicate there is a problem - even if one shows it as slightly worse than the other.

Thanks,

Barry

[Tanya Tritelnitskaya]

OK,  thank you very much for your time and attention, Barry!

[Tanya Tritelnitskaya]

Hello, Barry!

I would really appreciate it if you could take a little more time and help me clarify if I understand the process of improving performance after fixing the issue correctly. Here is a link https://docs.google.com/spreadsheets/d/13Fe1xtKyPM0rgGzMboJnw_5eU4Ot_6Han7Vg1r5DMTM/edit#gid=2146955330 where you could check my calculations. And do I understand correctly that during the 28 days we can observe the progress anyway if the issue was really fixed, or we will see it after the 75% threshold?

Thank you so much in advance!

[Barry Pollard]

The progress tab of the spreadsheet looks about right as a guess of the fixes.

It looks like from that, as of today you're expecting to have 32% of "good" pages, and 68% of "bad" pages is that right?

Looking at the URL below in PageSpeed Insights, and expanding the view we see this for mobile:

And this for Desktop:

So does indeed look like about 30% of good, which seems to match your expected calcs.

If you continue to check that over the next few days, hopefully you will see the percentage of "Good" pages continue to climb and once approximately 75% of your page views are in the "Good" category, I would expect your p75 score to also be <= 2.5 seconds and change to green. Until then you may not see any improvements on your p75 score, as it shows the value that 75% of your pages are equal to or better than, and some of your older page views likely got this 4.1 or 4.6 second values.

Thanks,

Barry

[Tanya Tritelnitskaya]

You are absolutely right about your guesses, Barry, and I'm very grateful for your taking time to check and your clear explanation!

Thanks again! Keep well!