Web Vitals and candidate metrics & focus areas ahead

Hey folks, happy friday!

Last week we (Google) announced the Web Vitals initiative
<https://blog.chromium.org/2020/05/introducing-web-vitals-essential-metrics.html>.
As a tl;dr, there are two big components: a program and commitment+promise
to better align all of our various tools and recommendations moving
forward, and a set of “core vital” metrics (LCP, FID, CLS) that we intend
to highlight and help site owners optimize for. Plus, predictable update
cycle, etc.

First off, big thank you to everyone in the group for all the help in the
incubation process of these metrics! While these metrics are not “new” to
anyone here, the announcement definitely drew a good amount of attention
and it’s been exciting to see tools, services, and frameworks integrating
them into their products: WebPageTest WPT
<https://twitter.com/patmeenan/status/1260294231589113859>, SpeedCurve
<https://speedcurve.com/blog/web-vitals-user-experience/>, Next.js
<https://nextjs.org/blog/next-9-4#integrated-web-vitals-reporting>, Gatsby
<https://github.com/bejamas/gatsby-plugin-web-vitals>, ... and there are
many more in the pipeline.

Celebrations aside though, there are many aspects both to the existing
metrics we’re highlighting, as well as broader aspects of user experience,
that we’d love to continue exploring. For example, for input we need to
expand coverage to all input not just the first (Event Timing will help us
address this; blink i2s went out today
<https://groups.google.com/a/chromium.org/g/blink-dev/c/1UQqjz1CR-k>), but
we also need to get better at measuring the end-to-end input instead of
just focusing on the initial event delay. Similarly, while LCP is a good
alternative to speed index for RUM
<https://datastudio.google.com/c/u/0/reporting/1sCHMXg3hgza3vqNVhDFAftYRgI94s2da/page/QDj4>,
we need to think about beyond initial page load: smoothness, frame
throughput, etc. There is also a host of open questions to answer on how
back-forward cache should affect all of our existing metrics, how we
normalize cumulative metrics like CLS, and the list goes on!

I would love to review some of these with more depth in one of our upcoming
design calls, and get input and feedback on priority areas that we could
explore as a group.

Nic, Yoav: could we reserve some time in one of the upcoming design calls?
🙏

ig