The 202206 dataset is live

303 views
Skip to first unread message

❄ Johannes Henkel

unread,
Jul 13, 2022, 9:31:39 PM7/13/22
to '❄ Johannes Henkel' via Chrome UX Report (Announcements)

Hi CrUX users,

This is your monthly announcement that the latest dataset has been published to BigQuery.

The 202206 (June 2022) dataset is now available and it covers 16,230,572 origins, an increase of 47.2% over last month. Let’s take a look at origins' Core Web Vitals performance this month:

  • 54.9% of origins had good LCP

  • 94.3% of origins had good FID

  • 72.3% of origins had good CLS

  • 41.3% of origins had good LCP, FID, and CLS

As mentioned in the release notes for the the previous month, the size increase of the dataset is due to including origins that were previously excluded because all dimensions (form factor and effective connection type) were required in the BigQuery schema; as we’ve made these dimensions optional, allowing for them to be NULL values in the BigQuery column, we instead report overall stats for origins that can’t be split in this way. That is, NULL in the effective_connection_type column stands for “all effective connection types” and NULL in the form_factor column stands for “all form factors”. We hope you enjoy the expanded coverage in this dataset as much as we do.

The significantly extended coverage makes it so that naively comparing the percentage of origins in the bullet points above with the previous month is problematic. To see how origins on the internet are changing with respect to the Core Web Vitals metrics, it’s useful to segment the discussion by popularity of the origins.

For the top 1k origins by rank magnitude (experimental.popularity.rank), we see 74.7% of origins with good LCP, 94.4% with good FID, 68.4% with good CLS and 53.2% with good LCP, FID, and CLS; so there’s an overall improvement compared to the prior month, with LCP improving most. It’s similar for the top 10k, for which we see both LCP and CLS improvements. The less popular segments of the web have the most churn in our dataset relative to the prior months, so a comparison is not useful at this moment. We’ll look at the trends again next month when we expect the dataset to remain more stable.

If you have any questions about CrUX, feel free to reach out to us on any of these channels:

Cheers,

Johannes and Rick

Reply all
Reply to author
Forward
0 new messages