Topic · 3 items
LCP
This page collects what affects LCP (Largest Contentful Paint), the Core Web Vitals metric for when the largest visible element renders during page load. Preload follows measurement changes in CrUX, Lighthouse and the web-vitals library that move a score while the site stays the same, along with the fixes that genuinely improve it. It's for anyone who has to explain a number as well as fix it.
Follow this topic By RSS, in your feed reader
CrUX finally measures soft navigations in single-page apps
Starting with the September dataset, released on October 13, CrUX adds three experimental variants of LCP, INP and CLS that count each route change in a single-page app as a navigation. They are available in BigQuery, the API, the History API and CrUX Vis, with FCP coming in November. Google aims to take them out of experimental status in May 2027, at which point Search Console would switch to this data. Across all origins, the share of good LCP goes from 68.3% to 72.1%.
Next step
On a single-page app, track both series side by side and note which dashboards and alerts only read the stable one. On a multi-page site, such as WordPress or PrestaShop without a decoupled front end, nothing changes.
A gap between the two series mostly reflects a change in what is measured, not a faster site: follow-up views reuse code that is already loaded. The JavaScript that runs on every route change still weighs on INP.
BEACON: field data from 10,000 large sites, open in BigQuery
Since late September, Cloudflare has published BEACON: anonymized field measurements from 10,000 of the largest sites on its network, updated daily in BigQuery using the RUM Archive format. Unlike CrUX, it covers WebKit on iOS, provides full distributions and breaks down LCP and INP into their subparts. For slow LCPs, render delay and load delay weigh far more than the image download itself.
Next step
Query the cf-open-web-performance.rumarchive project with the queries documented by the RUM Archive, and always filter on the date: tables are partitioned and BigQuery bills you for the bytes scanned.
A gold mine for spotting trends, not for judging a single site. The sample is limited to large sites behind Cloudflare and says nothing about your own domain.
web-vitals 6.2.3 fixes LCP attribution
The October 5 release of Google's library fixes a negative resourceLoadDuration value in LCP attribution, as well as a memory leak related to INP.
Next step
If your own monitoring reads LCP attribution, upgrade to 6.2.3 and discard any negative values you have already collected.