RankfinityDocs Open the app

DocsSite

Core Web Vitals

How to check page speed with real-user and lab data, what LCP, INP and CLS mean, and how the weekly check and alerts work.

4 min read·Markdown

Ask an AI about this page

ChatGPT and Claude open with the question copied — paste it with Ctrl+V (⌘V).

What Core Web Vitals are#

Core Web Vitals are Google's three measures of how a page feels to a visitor:

MetricMeasuresGoodPoor
LCP — Largest Contentful PaintHow fast the main content appears2.5 s or lessover 4 s
INP — Interaction to Next PaintHow fast the page responds to a click or tap200 ms or lessover 500 ms
CLS — Cumulative Layout ShiftHow much the layout jumps while loading0.1 or lessover 0.25

Between Good and Poor is Needs improvement. Google uses these signals in ranking, and slow or jumpy pages lose visitors.

Supporting metrics shown as well: FCP (First Contentful Paint — first text or image) and TTFB (Time to First Byte — server response time).

Field data and lab data#

Rankfinity shows two kinds of numbers:

  • Field data — measured from real Chrome users visiting the page over the last 28 days, reported as the 75th percentile (p75): 75% of visits were this fast or faster. Field data is what Google uses. Small sites often have no field data because there are not enough visits; the screen says "not measured" in that case, never zero.
  • Lab data — a test run on demand, on a simulated mobile phone and on desktop. It always exists, so it is useful for diagnosing, but it is one test, not real users. INP needs real interaction, so it is "not measured in a lab test".

Running a check#

  1. Open Site audit › Core Web Vitals.
  2. Click the run button. The check is free within daily limits.
  3. Rankfinity tests the most important pages from your last crawl (by default five), plus the site as a whole.

A second click within 24 hours shows the stored result rather than testing again. If a daily limit is reached, the screen says when you can run again.

Reading the results#

The site#

Field data for the whole origin (all pages together): p75 LCP, INP and CLS with a Good / Needs improvement / Poor badge, and the share of visits that were good, needing improvement and poor.

Each page#

  • Lab mobile and Lab desktop — LCP, CLS, FCP, TTFB and the overall performance score.
  • Field — p75 values with badges, if the page has enough traffic.
  • History — a line of field data for about six months, when available.
  • Top opportunities — the five biggest suggested improvements with the estimated saving (for example "Serve images in modern formats").

Pages with poor results also appear as a core_web_vitals issue in the site audit's Action Plan and in Copy fix prompt for your AI.

Weekly check#

Click Weekly check: on to test the project automatically once a week. It is free, off by default and uses the same limits. If a limit is reached or there is no crawl yet, it tries again later.

Alerts on a drop#

When a page that was Good becomes Poor on the same type of measurement (field against field, or mobile lab against mobile lab), you get an alert email together with the site audit alerts. A move from Good to Needs improvement does not send an email.

The Site Audit report template includes a "Page speed · Core Web Vitals" section, and the client share link shows the same section, using stored results only. See Client reports and sharing.

Improving the numbers#

  • LCP: compress and resize the hero image, serve images in WebP or AVIF, use a faster host or a CDN, avoid render-blocking scripts and fonts.
  • INP: reduce heavy JavaScript, split long tasks, delay third-party scripts (chat widgets, trackers) until needed.
  • CLS: give images and embeds a width and height, reserve space for banners and ads, avoid inserting content above what the visitor is reading.
  • TTFB: enable server caching, check slow plugins or database queries.

Test again after changes. Field data needs up to 28 days to reflect fixes, because it is a rolling window of real visits.

Common questions#

Why is field data "not measured"? The page or site has too few Chrome visitors for Google to publish data. Use lab data until traffic grows.

Why are lab and field numbers different? Lab is one simulated visit; field is thousands of real visits on real devices and networks. Google ranks on field data.

Does this cost credits? No, within the daily limits. The separate Performance tab is a paid lab test per URL.

Can a client run the check? Clients can view results. Running checks and the weekly schedule are for the agency.

Which pages are tested? The most important pages from your last crawl, chosen by the same ranking as the Action Plan.

Still stuck? Write to office@rankfinity.app.