Thread by @__treo

📊 We analyzed Core Web Vitals based on domain popularity.

Here're the findings:
• LCP is awesome – popular websites have better LCP
• FID is irrelevant – 99% of desktop sessions are good
• CLS is problematic – poor CLS is the reason sites don't pass CWV #webperf

Thread ⤵

When CrUX publishes the monthly state of Core Web Vitals, it's based on average and represents the origins of 1m+ websites.

Popular websites have resources, and they have better metrics, except CLS, more about it in the thread.

https://twitter.com/ChromeUXReport/status/1369591476712132608

LCP is amazing ✨

After years of iterations: load event, DOMContentLoaded, speed index, first paint, first contentful paint. We have a metric that works and reflects a good loading experience.

Improving LCP will bring happier visitors to your site.

FID is the easiest Web Vital to pass.

• 99% of desktop visitors have good FID;
• 1% of mobile visitors have poor FID;
• It doesn't depend on site's popularity.

Is it still a relevant metric to consider? #webperf

CLS is rigid in the current definition.

• Sites with longer sessions (more useful) have poor CLS
• On desktops, every third session has a poor CLS
• A weird drop in the 1k-10k popularity range

There's an open discussion for better CLS: web.dev/better-layout-…

Here's a look at Core Web Vitals' performance for the top 1000 origins by country.

Pass CWV (have Good LCP, FID, and CLS):
🇰🇷Korea 39.3%
🌍Global 28.4%
🇮🇳India 24.2%
🇺🇸USA 22.7%
🇦🇺Australia 20%
🇧🇷Brazil 17.5%
🇳🇬Nigeria 9.1%

The raw data for charts: docs.google.com/spreadsheets/d…

If you want to explore Core Web Vitals for your sites, check out Treo Site Speed treo.sh/sitespeed

It's a free CWV audit that allows to:
• see historical data
• filter by devices and geography
• compare with competitors
• set custom metrics
• and more