What Speed Index measures
Speed Index captures how fast the page looks complete, not just when the first or largest element appears. Lighthouse records a video of the page loading, measures how visually complete each frame is, and turns that progression into one number. A page that shows most of its content early scores better than one that stays blank and then appears all at once, even if both finish at the same moment.
It is one of five metrics in the Performance section of a Lighthouse report, next to FCP, LCP, TBT and CLS. In Lighthouse 10 it carries 10% of the performance score. It isn't one of the Core Web Vitals.
Speed Index thresholds
| Rating | Mobile | Desktop |
|---|---|---|
| Fast (green) | 0–3.4 s | 0–1.3 s |
| Moderate (orange) | 3.4–5.8 s | 1.3–2.3 s |
| Slow (red) | Over 5.8 s | Over 2.3 s |
Lighthouse sets these by comparing your result with the Speed Index of real websites in the HTTP Archive.
Example
Two versions of a landing page both finish rendering at 4 s in a mobile Lighthouse run. Version A shows the header, headline and hero image by 1.5 s and fills in the rest later; its Speed Index is 2.2 s. Version B waits on a font and a client-side render, then paints everything at 3.8 s; its Speed Index is 3.9 s, moderate. Same finish time, different experience.
How to measure Speed Index
Run Lighthouse in Chrome DevTools or PageSpeed Insights' lab section. WebPageTest reports the original version of the metric that Lighthouse's is based on. Because Speed Index needs a video of the load, it is a lab-only metric: there's no browser API to measure it for real visitors. For field data, FCP and LCP cover the same ground.
Improving it is the same work as improving load speed in general: less main-thread work, less JavaScript before first render, fonts that show text while they load, and the most important content in the initial HTML.
Common Speed Index mistakes
- Treating it as a ranking metric. It isn't a Core Web Vital; Google's assessment uses LCP, INP and CLS from field data.
- Comparing mobile and desktop runs. The thresholds differ by a factor of about 2.5.
- Trusting one run. Lab results vary between runs; test several times and compare the median.
How MetricRaven handles Speed Index
MetricRaven doesn't report Speed Index, because it can't be measured in a real visitor's browser. Its Page speed feature records the field metrics instead: LCP, INP, CLS, FCP and TTFB from each real page load. Pages show LCP and INP at the 75th percentile next to the purchase rate, with the others on hover, so you can see whether slow pages are the ones that sell less, which a lab score can't tell you. See Page speed vs conversion and compare with your conversion rate.
Questions
- What is a good Speed Index?
- In Lighthouse, 3.4 seconds or less on mobile and 1.3 seconds or less on desktop is fast. Over 5.8 seconds on mobile, or over 2.3 seconds on desktop, is slow.
- Is Speed Index a Core Web Vital?
- No. The Core Web Vitals are LCP, INP and CLS. Speed Index is a lab metric that counts for 10% of the Lighthouse 10 performance score.
- How is Speed Index different from LCP?
- LCP marks one moment: when the largest element appeared. Speed Index scores the whole visual progression, so a page that fills in gradually beats one that appears all at once. LCP can also be measured for real visitors; Speed Index can't.