Core Web Vitals Assessment: Failed when i get 100% at all

42 views
Skip to first unread message

Egphp

unread,
Jul 12, 2023, 2:09:33 AM7/12/23
to Chrome UX Report (Discussions)
0

PageSpeed Insights that shows you the experience of real users. Now I checked one of my websites with this feature on mobile and I got this message

enter image description here

even i got all 100% at all

enter image description here

url of test https://pagespeed.web.dev/analysis/https-egy-news-net/kv5lrh69pa?form_factor=mobile

Barry Pollard

unread,
Jul 12, 2023, 5:16:39 AM7/12/23
to Egphp, Chrome UX Report (Discussions)
Lighthouse does a simple load of the page under predefined conditions as an audit of a site to suggest best practices. The Core Web Vitals metrics are measured based on real users, which may be using the site very differently than Lighthouse. Lighthouse is often fairly representative but it has limitations.

You can see more information on why these differences happen here: https://web.dev/lab-and-field-data-differences/

In particular I note the following differences:
  • The site does not look like it's fully loaded in the Lighthouse screenshots. Is it missing content that was somehow excluded from Lighthouse?
  • A Lighthouse score of 100 is very difficult to achieve for a site of any complexity so I'm always somewhat suspicious when I see this, and drastically different field data. Many performance plugs claim to optimize site performance by serving different content to Lighthouse. This may trick Lighthouse, but does not trick real users and could be one reason for the difference.
  • Using the site locally a cookie banner is present, but this did not appear in Lighthouse. This may explain the much lower LCP time for real user.
  • Does the cookie banner load different content for real users, that makes the site slower on next visit?
  • For CLS differences, this is often caused by scrolling (which Lighthouse does not do). More details here: https://web.dev/optimize-cls/#identifying-post-load-cls-issues
  • For interactivity (FID and INP) these are based on users interacting with the website. As Lighthouse just does a load, without interactivity there are often differences here. See more information here: https://web.dev/diagnose-slow-interactions-in-the-lab/

--
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/7883b2a1-765e-4def-be4a-b0fb8b5c8335n%40chromium.org.
Reply all
Reply to author
Forward
0 new messages