Page Speed & Core Web Vitals
Google PageSpeed Insights: what it does, what it costs
Lab data and real-user data on the same screen, which is why the two disagree.
Google PageSpeed Insights at a glance
Every row is a verified field or an honest absence. A row is dropped rather than filled with an em dash, so the length of this table says how complete the record is.
| What it is | Lab data and real-user data on the same screen, which is why the two disagree. |
|---|---|
| Category | Page Speed & Core Web Vitals |
| Best suited for | Establishing whether a page passes Core Web Vitals, from real-user data |
| Starting price | Free |
| Free plan | Yes |
| Free trial | None offered |
| Main strength | Field and lab data on one screen, free, with no account required |
| Main limitation | Field data is a rolling window, so it lags a fix by weeks |
| Evidence | Primary-source researched |
| Closest alternatives | Google Search Console, Sitebulb, Semrush |
| Pricing checked | 19 August 2026 |
| Record checked | 19 August 2026 |
What Google PageSpeed Insights is
PageSpeed Insights reports the performance of a page on mobile and desktop, and it does so twice over. Field data comes from the Chrome User Experience Report — Largest Contentful Paint, Cumulative Layout Shift, Interaction to Next Paint and First Contentful Paint, measured from real Chrome users. Lab data comes from Lighthouse, which simulates a single page load in a controlled environment and produces the diagnostics.
Knowing which half to act on is the entire skill. Field data is what gets assessed against the Core Web Vitals thresholds, bucketed as Good, Needs Improvement or Poor — it is the real-world truth, and it lags because it is a 28-day rolling window. Lab data is a simulation under fixed conditions, which makes it reproducible and therefore useful for debugging, and unrepresentative of your actual visitors.
The common failure is chasing the Lighthouse score. It is a weighted composite of lab metrics on a throttled simulated device, so it moves when nothing a user experiences has changed. If your field data says Good and your score says 62, your site is fine and your score is a lab artefact.
What you can use it for
- Check whether a page passes Core Web Vitals using real-user data rather than a lab score
- Get prioritised diagnostics for a slow page, with the estimated saving for each fix
- Compare mobile and desktop experience for the same URL
- Establish which of LCP, CLS or INP is actually failing before optimising anything
The features that matter for SEO
Each one says what it does and the SEO job it serves. A vendor feature list without that second half tells you nothing you could not read on their own site.
Field data from the Chrome User Experience Report
Reports LCP, CLS, INP and FCP from real Chrome users, bucketed Good, Needs Improvement or Poor.
This is the half that is assessed against Core Web Vitals. It is the only data here that reflects what your visitors experience.
A rolling window, so it lags a fix by weeks. Low-traffic pages may have no field data at all.
Lighthouse lab diagnostics
Simulates a page load under fixed throttled conditions and reports performance, accessibility, best practices and SEO diagnostics.
Reproducible, which is what makes it useful for debugging — you can change one thing and re-run. It tells you what to fix; field data tells you whether it mattered.
A simulation on a throttled device. The score moves for reasons no real user would notice.
Core Web Vitals assessment
Passes or fails the page on LCP, CLS and INP against Google's published thresholds.
The single binary anyone reporting on page experience actually needs. Everything else on the report is diagnosis.
Opportunities and diagnostics with estimated savings
Lists fixes ordered by estimated time saving.
Turns a score into a work queue, and the estimates are good enough to prioritise with.
Mobile and desktop reported separately
Runs the whole analysis twice, once per form factor.
Mobile is usually the failing half and the one that is assessed, so a desktop-only check hides the problem.
API and DevTools access
The same analysis is available programmatically and inside Chrome DevTools.
The API makes performance monitoring across many URLs possible without a paid tool.
Pricing and limits
Starting price: Free Checked 19 August 2026 Vendor’s own pricing page
| Plan | Price | What it includes |
|---|---|---|
| PageSpeed Insights | $0 /mo |
|
Strengths and limitations
Strengths
- Field and lab data on one screen, free, with no account required
- The only free source of real-user Core Web Vitals for an arbitrary URL
- Diagnostics come with estimated savings, so they order themselves into a work queue
- Available as an API, which turns it into a monitoring tool at no cost
- Mobile and desktop assessed separately, which is how Google assesses them
Limitations
- Field data is a rolling window, so it lags a fix by weeks
- Pages with too little traffic have no field data, leaving only the simulation
- The Lighthouse score is a lab composite and moves without any real-world change
- One URL at a time in the interface — site-wide analysis needs the API or a crawler
- Diagnoses symptoms, not causes; the fix is usually in a template or a third-party script
Who it is for, and who should skip it
Worth it if
- Establishing whether a page passes Core Web Vitals, from real-user data
- Debugging a specific slow page with reproducible diagnostics
- Anyone who needs performance data and has no budget
Skip it if
- You need site-wide performance data — use the API or a crawler that includes it
- You are optimising for the score rather than for the field metrics
Alternatives to Google PageSpeed Insights
Each one answers when it is the better choice, which is the only thing a list of competitors is useful for.
- Google Search Console Choose it when you need Core Web Vitals grouped across the whole site rather than one URL at a time.
- Sitebulb Choose it when you want performance auditing across every page in a crawl, with the issues explained.
- Semrush Choose it when you want scheduled site-wide performance monitoring inside a platform you already pay for.
Sources and verification
Last verified 19 August 2026 1 day ago
Pricing 19 August 2026 Features 19 August 2026
- About PageSpeed Insights Documentation verifies features, limits, pricing checked 19 August 2026
- Introduction to Lighthouse Documentation verifies features checked 19 August 2026
- PageSpeed Insights Official site verifies availability checked 19 August 2026
What changed
- 19 August 2026 — Listed as one product with Lighthouse recorded as the engine behind its lab data, not as a separate listing. "Lighthouse by Google" and "Google Lighthouse" are the same thing and must never appear as their own entries.
Questions
Why is my PageSpeed score low but my Core Web Vitals pass?
Because they measure different things. The score is a weighted composite of Lighthouse lab metrics taken on a throttled simulated device; Core Web Vitals are assessed from real Chrome users over a rolling window. If field data says Good, your visitors are fine and the score is a lab artefact. Act on the field data.
Why does PageSpeed Insights show no field data for my page?
Field data comes from the Chrome User Experience Report, which needs enough real-user traffic to report on. Low-traffic pages have none, leaving only the Lighthouse simulation — which is the least representative half of the report.
Is Lighthouse the same as PageSpeed Insights?
Lighthouse is the engine that produces the lab half of a PageSpeed Insights report. PSI adds the real-user field data from CrUX around it. Lighthouse also runs standalone in Chrome DevTools, as a CLI and as a Node module.