Hi all,
I took a look at what the usage of our prioritized features is in HTTP Archive, which is per page in the index in contrast to Use Counters which are per page view in Chrome.
I averaged the HTTP Archive usage between desktop and mobile as they didn't vary as much as I thought, and here's the comparison:
Feature | Use Counter | HTTP Archive |
Flexbox | 75% | 68% |
Grid | 8.5% | 6.2% |
position: sticky | 8.3% | 2.8% |
aspect-ratio | 2.9% | 0.2% |
Transforms | 80% | 86% |
The large difference for position:sticky and aspect-ratio is interesting, but makes sense. It's enough for some actively maintained sites with a lot of users to adopt a new feature to drive up the use counter, but in HTTP Archive each of the 5-7M sites count equally.
If both measures of usage go up and the signals of developer pain (bugs, surveys) don't increase then we're on the right track.
Best regards,
Philip
P.S. the query was:
SELECT
client,
feature,
ROUND(100*pct_urls, 2) AS usage
FROM
`httparchive.blink_features.usage`
WHERE
yyyymmdd = '20210201' AND feature IN ('CSSFlexibleBox', 'CSSGridLayout', 'PositionSticky', 'CSSFeature_657', 'CSSPropertyTransform')
ORDER BY
feature