PageSpeed Insights number of distinct samples to show data for a URL logic

Question

I'm reading the PageSpeed Insights documentation and am wondering if anyone knows how Google is determining what is considered a sufficient number of distinct samples per this FAQ:

Why is the real-world Chrome User Experience Report speed data not available for a URL?

Chrome User Experience Report aggregates real-world speed data from opted-in users and requires that a URL must be public (crawlable and indexable) and have sufficient number of distinct samples that provide a representative, anonymized view of performance of the URL.

I'm building a report centered around Core Web Vitals data and realizing some URLs have few data points with CWV timings, and I'm curious exactly how Google is handling these situations. I've been searching through docs and articles, haven't found anything with a specific reference.

Answer 1

The exact threshold is kept secret, so that's why you won't find it documented anywhere. However, as a site owner there are a few things you can do to work around a URL not having sufficient data:

• Use the Core Web Vitals report by Search Console, which groups similar pages together, making them more likely to collectively exceed that threshold.
• Look at origin-level aggregations in PSI or the CrUX API. These include user experiences from all pages on the origin, so it's much less granular, but it gives you a sense of typical experiences overall.
• Instrument your site with your own first-party Core Web Vitals monitoring. web-vitals.js can be integrated with your existing analytics provider to track vitals on all of your pages. If you're integrating with Google Analytics, you can link your data with the Web Vitals Report to see how your site is doing.
• Use your site with the Web Vitals extension enabled to see the Core Web Vitals data for your own experience. Your local experiences may not be representative of most users, but this can be a great tool for validating expectations vs reality.
• Use lab data as a proxy. For example, lab data from Lighthouse in PSI can tell you how a mobile user on a slow connection might experience your page. This should really only be used as a last resort when no other field data is available.

Answer 2

It's pretty crazy the difference using the Web Vitals extension.

I thought my website was super slow because of the Page Speed Insights, but turns out just fine using the extension.