speedlify2

Lighthouse lab metrics, Axe, and Chrome UX Report field data for a list of sites recorded and compared over time. Follow us on Mastodon or Bluesky.

Charts Compare Measurement runs Embed a score Open to Work

Charts

Pulls the most recent measurement of every site (excluding starter projects). Tooltips are only shown on green (well scoring) data points.

Page Weight vs LCP 7 sites

3 within 2.5 s 0 to 4 s 4 over 4 s Hover a green dot to see its site.

JavaScript Weight vs LCP 7 sites

3 within 2.5 s 0 to 4 s 4 over 4 s Hover a green dot to see its site.

TTFB vs LCP 7 sites

3 within 2.5 s 0 to 4 s 4 over 4 s Hover a green dot to see its site.

Web Font Weight vs CLS 7 sites

6 within 0.1 0 to 0.25 1 over 0.25 Hover a green dot to see its site.

Performance vs Third Party 7 sites

Lighthouse’s Performance score weighed against third-party weight — usually things like analytics, embeds, fonts, tag managers, ads.

3 scoring 90 and above 1 scoring 50 to 90 3 scoring under 50 Hover a green dot to see its site.

Performance vs DOM Elements 7 sites

Lighthouse’s Performance score against DOM element count. A big document is paid for in style and layout work on every frame.

3 scoring 90 and above 1 scoring 50 to 90 3 scoring under 50 Hover a green dot to see its site.

Client Rendering by Generator 7 sites

Sites that painted nothing with scripts disabled, by the tool their page reports building it. One row per generator: a dot for how many of its sites were judged, a dot for how many came up blank, and a line joining them — so the length of the row is the size of the sample the finding came out of.

1 empty pre-JavaScript blank pages from 7 measured sites.

0 1 2 3 Sites Next.js 1 of 3
Sites judged Drew nothing without JavaScript Scroll sideways for the rest of the scale.

The other 1 generator flagged nothing across 4 judged sites between them — mostly on samples of a handful of sites each.

Field vs Lab LCP 7 sites

One row per site, ordered by our throttled lab run: a dot for what the lab measured, a dot for what real users actually saw (Chrome’s field p75) and a line joining them. Sorting by the lab figure straightens it into a curve, and the field dots scatter across it.

The lab is slower than reality on 5 of these and faster on 2.

Field Lab, within 2.5 s Lab, to 4 s Lab, over 4 s