Comparisons
Search Console vs PageSpeed Insights: Field Data or Lab Data
They disagree because they measure different things: one reports what real visitors experienced over 28 days, the other simulates one load on a throttled device.
Run a fresh DomainLens audit and use the report as your priority list.
Why the numbers differ
Search Console reports field data — real Chrome users, aggregated over a rolling 28-day window, grouped by URL pattern. PageSpeed Insights shows both: the same field data at the top, and beneath it a Lighthouse lab run performed on demand against a simulated mid-range phone on a throttled connection.
A single lab run on one device with one network profile will not match an average of thousands of real sessions across every device and connection your visitors have. The disagreement is the tools working correctly, not one of them being wrong.
| Search Console | PageSpeed Insights | |
|---|---|---|
| Data | Field only | Field and lab |
| Window | Rolling 28 days | Field: 28 days. Lab: right now |
| Scope | Groups of similar URLs | One URL |
| Availability | Needs enough traffic | Lab always; field if data exists |
| Used for ranking | Yes — this is the data | Only its field half |
| Best for | Which pages have a problem | Why that page is slow |
Only field data counts for ranking
Google has been explicit: the page experience signal uses Core Web Vitals from real users, taken from the Chrome UX Report. The Lighthouse performance score — the 0-to-100 number everyone screenshots — is not used for ranking at all.
This inverts how most teams use the tools. Chasing a perfect Lighthouse score optimises a simulation. The number that matters is whether your field metrics clear the thresholds, and a site can pass those comfortably while scoring 70 in the lab.
Using each for what it is good at
That lag is the part teams find hardest. A fix deployed today will not show fully in field data for weeks, and the temptation is to keep changing things in the meantime. Resist it, or you will not know which change worked.
- Start in Search Console: it groups URLs, so it tells you which templates fail rather than which single page you happened to test.
- Take one failing URL from that group into PageSpeed Insights for the lab trace.
- Read the lab diagnostics — the LCP element, blocking resources, main-thread work — to find the cause.
- Fix at the template level, since Search Console already told you the group shares it.
- Verify in the lab immediately, then wait for field data to confirm; the 28-day window means real confirmation is slow.
When field data is missing entirely
A URL with too few Chrome visitors gets no field data, and PageSpeed Insights falls back to origin-level data or shows none at all. That is common for new pages and small sites, and it is not an error.
In that situation the lab run is all you have. Use it, but treat it as a plausibility check on a simulated device rather than a measurement of your visitors, who may be on better or far worse connections.
How DomainLens contributes
DomainLens pulls both alongside the rest of the audit: field metrics where they exist and Lighthouse lab results where they do not, so a performance finding sits next to the indexability and on-page findings instead of in a separate tool. For choosing between all three data sources see PageSpeed Insights vs Lighthouse vs CrUX , and for the metrics themselves see the Core Web Vitals guide .
- Why do Search Console and PageSpeed Insights show different numbers?
- Search Console reports 28 days of real-user data grouped by URL pattern; the PageSpeed lab section simulates one load on a throttled device. Different measurements, different answers.
- Which one does Google use for ranking?
- Field data — real Chrome user measurements. The Lighthouse performance score is a diagnostic and is not a ranking input.
- Why does my page have no field data?
- Too few Chrome visitors to reach the reporting threshold. That is normal for new or low-traffic pages, and the lab run is your only signal until traffic grows.
- How long before a fix shows up in Search Console?
- The window is 28 days, so expect several weeks before the improvement is fully reflected. Verify in the lab immediately and be patient with the field.