LCP Field Data Anomaly

Question

I have an oddity with my LCP field data score - which currently stands at a dire 3.2s - according to https://pagespeed.web.dev/

However, according to every test I run the score is between 0.6s and 1s

I've tried Chrome Lighthouse, online tools like https://www.webpagetest.org/ - and I've tried logging my LCP data to Google Analytics and analysing using https://web-vitals-report.web.app/ - all show an LCP score of less than 1s

I've looked at this: https://web.dev/debug-web-vitals-in-the-field/#the-web-vitals-report-tool

But there doesn't seem to be a way to show which element is causing the LCP delay in the field?

Any advice gratefully received!

Answer 1

That post you linked details how to gather the LCP element just above the section you linked from: https://web.dev/debug-web-vitals-in-the-field/#usage-with-the-web-vitals-javascript-library .

The next version of web-vitals.js library will include this code in the library itself in an Attribution build that makes it easy to report back elements to the likes of Google Analytics.

If you are not seeing slow LCP in your field data read by web-vitals.js then that is promising and suggests the issue may be fixed. The CrUX data measures over 28-days and it could be just an old issue.

However, it could still be an issue, that is just not showing up. Either because you occasionally get traffic from slower devices and connections, Or because there is something the web-vitals library cannot measure. In particular JavaScript solutions cannot always measure cross-site content accurately (eg iframes or images from other domains). More detail in this post: https://web.dev/crux-and-rum-differences/

Unfortunately without some more details (is it origin-level LCP that's an issue or URL-level? What is your LCP element typically? Are the amount of pages passing as given by the CrUX API increasing each day?) it is difficult to give more advice than that.