Skip to content

Lighthouse (lab CWV)

The Lighthouse view runs Google Lighthouse lab tests on your pages and sets each result beside the real-user field data for the same page. Lab and field answer two different questions. Lab is a single, reproducible test in a controlled environment — the right tool for debugging a regression or proving a fix before and after. Field is what real Chrome users actually experienced over the trailing 28 days, and it is the source of truth Google uses for ranking. Seeing both side by side is the point: when lab looks great but field looks poor, real users are hitting conditions your clean test never simulated. The data comes from VisibilityIQ.

  • A project with crawled pages to test.
  • Available monthly data budget — queuing Lighthouse runs consumes from the project’s included budget.

Open the project, go to Site Audit → Lighthouse, and click Run Lighthouse. VisibilityIQ queues a lab test for each page; rows move from pending to done as results land. Mobile and desktop are tested and reported separately, because the same page can pass on one and fail on the other. When the runs complete, each page shows its lab metrics next to the matching field metrics and the month that field data covers.

For each page and device you get the lab measurement, the Lighthouse performance score, and the field measurement beside it.

MetricTargetWhat it is
LCP< 2.5sLargest Contentful Paint — when the main content finishes rendering
CLS< 0.1Cumulative Layout Shift — how much the page jumps as it loads
INP< 200msInteraction to Next Paint — the field interactivity metric Google ranks on
FCPsupportingFirst Contentful Paint — when the first content appears; a supporting signal
TBTlab proxyTotal Blocking Time — a lab stand-in for INP, since INP needs real interactions

Lab columns (lab LCP, CLS, FCP, TBT, and the performance score) come from the controlled test. Field columns (field LCP, CLS, FCP, and the field data month) come from real users. Read them together. When lab and field agree, the lab test is a faithful debugging surface — fix the page, re-run, and watch the lab number move. When lab is green but field is red, trust the field: real devices, real networks, and real third-party scripts are producing an experience your test environment didn’t reproduce. TBT is your lab clue toward an INP problem you can’t measure in a single synthetic run, because INP only exists once a real person interacts.

Click Run Lighthouse again to re-queue tests; runs move from pending to done as they finish. A run consumes from the project’s monthly budget, so re-run after you ship a performance change you want to verify rather than on a fixed cadence. Field data refreshes on its own 28-day trailing window and lags a deploy — lab will move the moment you re-run, field will catch up over the following weeks.