Guide
How to read a Lighthouse / Core Web Vitals report
Lighthouse and Core Web Vitals both talk about speed, but they answer different questions. Mixing them up is how teams chase a 100 lab score that real visitors never felt — or ignore a field problem because the office Wi‑Fi looked fine.
Lab vs field
Lighthouse is a lab test. On a paid WebVitalist report we request it through Google’s PageSpeed Insights API on a simulated mobile device with a throttled connection. You get a Performance score from 0–100 plus lab metrics (LCP, FCP, Total Blocking Time, CLS, Speed Index). Run it twice and the score can move a few points. It is a diagnostic, not a census of your visitors.
Core Web Vitals in the field come from the Chrome UX Report (CrUX): real Chrome users over the last 28 days, reported at the 75th percentile. Google classifies LCP, INP and CLS as good, needs improvement, or poor. Sites with little traffic often have no field data. That is expected. It does not mean the page is fine or broken.
Paid reports print both when Google has the data. The performance score we use in the overall mark is Lighthouse’s Performance category, unchanged. Details live on how we score.
The three Core Web Vitals
These thresholds are Google’s. We do not invent a fourth band or re-label a “poor” field metric as “good” because the lab run looked better.
| Metric | Good | Poor |
|---|---|---|
| Largest Contentful Paint (LCP) | ≤ 2.5 s | > 4 s |
| Interaction to Next Paint (INP) | ≤ 200 ms | > 500 ms |
| Cumulative Layout Shift (CLS) | ≤ 0.1 | > 0.25 |
Anything between the two boundaries is “needs improvement”.
- LCP (Largest Contentful Paint) — when the main content (hero image or heading block) becomes visible. Slow LCP is often a huge image, a late web font, or a server that takes too long to return HTML.
- INP (Interaction to Next Paint) — how long the page takes to respond to clicks and taps. The lab still reports Total Blocking Time as a proxy; field INP needs real users. Heavy third-party scripts are a common cause.
- CLS (Cumulative Layout Shift) — how much the layout jumps while loading. Missing image dimensions, injected banners, and web fonts that swap late are typical.
Lab targets we print next to Lighthouse numbers (informational — the score itself is still Lighthouse’s):
| Lab metric | Target |
|---|---|
| Largest Contentful Paint | < 2,500 ms |
| First Contentful Paint | < 1,800 ms |
| Total Blocking Time | < 200 ms |
| Cumulative Layout Shift | < 0.1 |
| Speed Index | < 3,400 ms |
| Server response time (TTFB) | < 800 ms |
| Total page weight | < 1.5 MB |
How to read a Lighthouse score
- 90–100 — the lab looks healthy on mobile. Still check field data and the opportunities list; a 96 with a 4 s field LCP is not “done”.
- 50–89 — there are specific, usually fixable savings (images, unused JavaScript, render-blocking CSS).
- 0–49 — something large is in the way: page weight, a slow TTFB, or a blocking script that holds the main thread.
The number is not a search ranking. Google uses field Core Web Vitals as one ranking signal among many others. A better Lighthouse score is evidence you removed lab friction. It does not guarantee positions, traffic, or revenue.
When lab and field disagree
Common reasons:
- You tested a URL users rarely visit — or the popular URL is not the one you audited.
- CrUX may cover a whole origin or a busy directory, not only the landing page you care about.
- Office Wi‑Fi is faster than the mobile throttle Lighthouse uses.
- A cookie banner, A/B test, or personalization that the lab run did not hit the same way visitors do.
When they disagree, treat field data as “what users felt” and Lighthouse as “what to try next”. If there is no field data, the lab run is still useful — just do not pretend it is a 28-day user study.
What WebVitalist actually measures
The free scan is a teaser: one public page, 19 technical SEO checks, platform detection, and the top 3 issues. Timeout 8s. It is not a Lighthouse run and it does not review your copy.
A paid Audit (from $9) adds a mobile Lighthouse run, a Security Pass (headers, cookies, leak checks), an automated accessibility audit, and an AI copy/UX review that must cite evidence from the page. Deep Audit crawls more pages; Full Test adds HTTP-level link and form checks (today that step is not a real browser — see the methodology).
Get a free technical score Start a paid audit
Want to see the output first? Sample Deep Audit of a fictional store.
What we do not claim
- No “#1 on Google”, guaranteed rankings, or promised conversion lift.
- No traffic, keyword, or backlink database — we fetch the page you give us.
- No logged-in, paywalled, or geo-restricted content. Technical checks read server-rendered HTML; Lighthouse does render the page for performance and accessibility.
- If PageSpeed Insights is unavailable, we skip performance and accessibility and say so on the report rather than inventing scores.
Weights, check ids and crawl caps: how we score. Last updated 2026-09-18.