Groups keyboard shortcuts have been updated
Dismiss
See shortcuts

lighthouse LCP longger than performace LCP

57 views
Skip to first unread message

zeren zhang

unread,
Mar 7, 2025, 5:57:08 AMMar 7
to web-vitals-feedback
My application uses React streaming rendering. The data obtained from the Performance panel in Chrome DevTools is basically consistent with the results measured by the Performance API and the Web Vitals API.
1.png
3.png

However, when using Lighthouse to measure LCP, I found that the time consumed is significantly higher than the aforementioned data.
2.png

Barry Pollard

unread,
Mar 7, 2025, 6:06:29 AMMar 7
to zeren zhang, web-vitals-feedback
Hi, you haven't said where you're running Lighthouse (in DevTools, in PageSpeed Insights, or in CI), but let's assume DevTools for now.

Lighthouse uses particular settings for it's run. In particular in throttles CPU, and network on mobile. You can see this when you hover over the Device and Network settings in the middle column at the bottom:

image.png

For mobile it uses 4x slowdown, so it's unsurprising that LCP is 4x slower. The is based on the fact that most mobile devices are considerably slower than most developers desktop machines.

For desktop Lighthouse tests, there is no throttling, so they are likely to be more inline with your unthrottled tests.

For PageSpeed Insights (PSI) and CI, the Lighthouse figures will depend on the hardware it's run on. You can see in the above screenshot that my M1 Macbook has a base rating of 3079 but for PSI the servers are much more constrained by that.

When using Lighthouse it's better to think of them as just one example load of what the page might be like under those particular conditions. For CI you can tweak the throttling settings to better match your userbase if you know the types of devices your users use. But personally I think it's better to just use the same Lighthouse in before and after runs to measure the impact of any improvements rather than worry about calibrating them too much.


--
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 visit https://groups.google.com/d/msgid/web-vitals-feedback/389086c3-d54a-4757-bf9a-59b6963f2639n%40googlegroups.com.
Reply all
Reply to author
Forward
0 new messages