CLS not capped correctly

[Fabrizio Ferrari - VSM]

Hello.

I have been trying to report this to you and your team via Twitter in the past month or so, but I am not sure you got it.

CLS is still not correctly measured, try to browse the sheet music displayed on this page by clicking the left-right arrow on the sides of the displayed sheet music and you'll see what I mean:

https://www.virtualsheetmusic.com/score/LargoFREE.html

If you keep turning pages, CLS keeps accumulating even after minutes. I am attaching a screenshot of what I got by using your “Experimental Web Vitals” Chrome plug-in.

I don’t think that’s what you want CLS to account for.

Thank you for listening!

All the best,

Fabrizio Ferrari

Virtual Sheet Music, Inc.

https://www.virtualsheetmusic.com

[Rick Viscomi]

Hi Fabrizio,

Based on the screenshot it looks like you're using the Web Vitals extension for Chrome. The extension is in the process of updating to the latest version of web-vitals.js, so once those changes go live in the Chrome Web Store you should start to see the new capped CLS behavior. You'll know when you're using the updated version of the extension if you go to chrome://extensions/ and it's listed as Web Vitals 1.0.0.

Apologies for the confusion and for not seeing your messages on Twitter sooner!

Rick

[Meera Preety]

What is each post price?

waiting..

--
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/34e48369-85dd-46cc-87f1-dc3634a13543n%40googlegroups.com.

[Dmitriy Yusupov]

Hi, Google Search Console shows high CLS values even for pages with very simple restraint. For example, you can use a page with one word content: https://poncy.ru/tst/cls.html ( https://developers.google.com/speed/pagespeed/insights/?url=https%3A%2F%2Fponcy.ru%2Ftst%2Fcls.html&hl=ru&tab=desktop ), the page has neither CSS nor JS, and the analyzer shows that the Cumulative Layout Shift (CLS) value is 0.13 for the desktop. How can this be fixed? Because of this, the google search console marks all pages with this fake metric. Many thanks.

понедельник, 28 июня 2021 г. в 10:59:11 UTC+3, meerafrien...@gmail.com:

[Philip Walton]

Did you recently change this page from something else to just displaying one word?

If so, keep in mind that Google tools that show field data use a 28-day aggregation window, which means if you made the change less than 28 days ago, data from your older page is still contributing to the score.

To view this discussion on the web visit https://groups.google.com/d/msgid/web-vitals-feedback/4fa281bd-9a91-4a2a-8a8f-9cf1468545ddn%40googlegroups.com.

[Dmitriy Yusupov]

Hello, Philip. Thanks for the answer. No, I specifically created this page today for a test to find places that google doesn't like.

You do not want to say that this indicator is the same for all pages of the site?

четверг, 15 июля 2021 г. в 05:38:15 UTC+3, Philip Walton:

[Dmitriy Yusupov]

Moreover, this CLS value (0.13) is obtained for all my pages that have completely different templates and were made a very long time ago, for example, this page: https://poncy.ru/news_opensite.html. It looks as if there is no influence of the page content at all, so I created an example of a simple page to show that the erroneous cls hangs on everything.

четверг, 15 июля 2021 г. в 05:45:59 UTC+3, Dmitriy Yusupov:

[Dmitriy Yusupov]

And this value of 0.13 is everywhere, new pages or old ones does not matter, the template also does not matter (another example: https://developers.google.com/speed/pagespeed/insights/?url=https%3A%2F%2Fponcy.ru%2Ferror_pages%2F403.html&hl=ru&tab=desktop)

CLS value of 0.07 for mobile is as stable as 0.13 for desktop

четверг, 15 июля 2021 г. в 05:52:25 UTC+3, Dmitriy Yusupov:

[Philip Walton]

Ahh, sorry, I didn't notice before that the PageSpeed Insight link you sent was showing origin data.

Given that this is a brand new page, it makes sense that this isn't enough real-world data to show for it, so instead PSI is showing you the data for all pages on your origin. 

Search Console is doing a similar thing, it groups pages with not enough data together based on similar HTML structure, and if there isn't enough data even for pages in a group, it will fall back to using origin data.

I would recommend figuring out which pages on your site have high CLS and fixing those pages, then that should improve the numbers you see in PSI (after 28 days).

To view this discussion on the web visit https://groups.google.com/d/msgid/web-vitals-feedback/d3aacc82-1ab9-4d86-b7ae-5f7a2cd09cb2n%40googlegroups.com.

[Dmitriy Yusupov]

Thanks, Philip, but how do you know which pages have a high CLS score if they all have a value of 0.13? I have been looking closely at the shift of the template and cannot find what google dislikes. Maybe you know exactly how to determine (perhaps tools) or from what such a high CLS value is shown, because the pages are quite stable. Thanks.

четверг, 15 июля 2021 г. в 06:50:02 UTC+3, Philip Walton:

[Philip Walton]

Sorry, if none of your individual pages has sufficient field data, then I don't think any of our tools will show you which one (or ones) is causing the issue.

Search Console generally orders pages in it's list by popularity, so one suggestion would be to start with the first one it lists.

Another option is to use analytics to measure the field data yourself, which should help give you more insight. I wrote a post recently about how to do that:

https://web.dev/debug-web-vitals-in-the-field/

To view this discussion on the web visit https://groups.google.com/d/msgid/web-vitals-feedback/75895b07-6695-4c40-bcb9-2ac49ad627abn%40googlegroups.com.

[Dmitriy Yusupov]

Thanks Philip. I think I understand - above these are the values for the entire site, and for a specific page in the "imitation of page loading" block. But it's still not clear what exactly to fix if the pages listed in the google search console have values significantly lower than 0.13, which are indicated for correction in the top block.

четверг, 15 июля 2021 г. в 08:10:56 UTC+3, Philip Walton:

[Philip Walton]

Yeah, I understand this can be confusing, and unfortunately the tool interface does not do a good job of clarifying what the difference between these two numbers is.

The most likely explanation for the difference in your case is that layout shifts are happening, but they're happening lower down on the page (outside of the initial viewport), and they're only visible if the user scrolls down.

These kinds of shifts show up in "field data" because field data measures real user interactions and page visits. "Lab data" on the other hand only measures the page while it's loaded, and it doesn't account for user interactions or scrolls.

Again, I understand this can be confusing. We're planning to publish some content on web.dev soon that outlines all the reasons lab and field data can be different.

To view this discussion on the web visit https://groups.google.com/d/msgid/web-vitals-feedback/f74258c5-78bc-453e-826a-43200baf6d0fn%40googlegroups.com.

[Dmitriy Yusupov]

Thanks Philip, you explained well. The only pity is that I do not know how to fix this, but it is necessary, since this indicator seems to have affected the visibility of the site in google search.

четверг, 15 июля 2021 г. в 08:53:54 UTC+3, Philip Walton:

[Dmitriy Yusupov]

Philip, please tell me, the when way the user interacts with the content in this way (example of the page: https://poncy.ru/crossword/?mask=-----&desc=%D0%B7%D0%B0%D1%8F%D1%86) can cause high cls values? Those. you can click on words in the list, words expand into descriptions and shift the content below them, can this shift cause a bad rating? If so, is that right?

четверг, 15 июля 2021 г. в 09:04:46 UTC+3, Dmitriy Yusupov:

[Dmitriy Yusupov]

Philip, please tell - how then, can the CLS values for the pages be improved if the "user interaction" data is lefted behinde the scenes,  E. how to understand what exactly this analyzer does not like?

четверг, 15 июля 2021 г. в 09:44:38 UTC+3, Dmitriy Yusupov:

[Philip Walton]

Hi Dmitriy, the best suggestion I can offer is the one I gave above, that's explained in this post:

https://web.dev/debug-web-vitals-in-the-field/

The techniques described in the post will help you determine what elements are shifting for the users of your site.

To view this discussion on the web visit https://groups.google.com/d/msgid/web-vitals-feedback/5966bd20-f22e-46fc-87a2-c7dce2bd661en%40googlegroups.com.

[Dmitriy Yusupov]

Philip, thank you very much for your answer. I will carefully study this information.

четверг, 15 июля 2021 г. в 19:30:17 UTC+3, Philip Walton: