WPT chooses different LCP than Pagespeed Insights


On article pages, WPT often selects the title as the LCP rather than the hero image. Google’s pagespeed test always selects the image as LCP so I would like WPT to do the same, since the LCP time is being underestimated when the title is used. Is there any way to do this? I am emulating mobile on a Samsung Galaxy S8, which I believe has the largest device screen of the available Android devices.

Example: https://pocketnow.com/iphone-13-colors

Perhaps this will help…

Thanks @RogerM, I read the thread you linked but it doesn’t address my question, I understand that Lighthouse and PSI are different tests from WPT. I’m asking if there’s a way to force the image to be chosen as LCP, or rather, I want to know if there’s a way to track the time at which the hero image is fully rendered.

The WPT tool is obviously a much more robust testing tool for debugging Core Web Vitals than Google’s tools. But, it’s far less useful to me if I can’t test the timing of the hero image.

Are you able to add Element Timing to the image? (See more here: September WebPageTest Roundup - WebPageTest Blog)

If so, my recommendation would be to add it to both the image and the h1. If lab tools are sometimes showing you one versus the other, it’s likely real user data and Chrome User Experience Report data are also going to be occasionally reporting one or the other, so monitoring and measuring both would be important.

If you use Element Timing, you can grab it in RUM data and WebPageTest will also report the timing data alongside your LCP metric.