Lighthouse alternatives for measuring page speed

Three complaints, and all three are correct.

• The score moves between runs of the same page, because a simulated load is not deterministic. A performance score is a reading, not a fact.
• It is lab data. Google's page experience assessment uses field data: the 75th percentile of real Chrome users' loads over 28 days, per URL or per origin.
• It cannot interact with the page, so it cannot measure Interaction to Next Paint, which replaced First Input Delay as a Core Web Vital in March 2024. It reports Total Blocking Time as a lab proxy instead.

• Lab or field: a simulated load, or what your real visitors experienced.
• One page or the whole site: a single URL, or every URL grouped.
• Depth: a verdict, or the request-by-request evidence.
• Watching: a run you trigger, or a schedule with alerts.
• Interaction: can it measure what happens after a click.

1. 1

   The verdict, free: PageSpeed Insights

   It runs Lighthouse and shows the Chrome User Experience Report field data above it, so one page gives you both the judgement and the diagnosis. Read the field section first; the lab score below it is the explanation, not the grade.

2. 2

   Field data at scale: the Chrome UX Report

   The same real-user dataset, available beyond one URL: origin-level figures, history, and comparisons between sites. It answers "is my whole site slow" rather than "is this page slow".

3. 3

   The whole site as Google groups it: Search Console

   The Core Web Vitals report groups your URLs into good, needs improvement and poor, by device, from the same field data. It is the shortest path from a speed problem to a list of templates to fix.

4. 4

   Depth on one load: WebPageTest

   Filmstrips, waterfalls, many locations and devices, first and repeat views, and runs you tune one variable at a time. When the question is why, this is the instrument.

5. 5

   A readable report you keep: GTmetrix

   Built on Lighthouse, with a history of runs, test locations and monitoring on paid plans. It is the tool for watching a page over months rather than debugging one load.

6. 6

   Your own browsing: DevTools and the Web Vitals extension

   The Lighthouse and Performance panels in Chrome DevTools for a page you are working on, and the Web Vitals extension to see the metrics as you click through your own site.

• Before and after one change, on your machine, with everything else held still.
• On a page with no field data at all, which is every new page and every low-traffic page. A lab run is the only reading you can get.
• For the other three categories: accessibility, best practices and SEO audits, each scored 0 to 100, which the field data says nothing about.
• In continuous integration, where a budget can fail a build when a bundle grows.
• As a list of causes: the opportunities and diagnostics are the most actionable part of the report, whatever the score says.

Used that way, the varying score stops mattering. You are reading the list, not the number.

• PageSpeed Insights: one URL at a time, and no field section when the page has too few visits.
• Chrome UX Report: no advice, only measurements, and only for sites with enough Chrome traffic to report.
• Search Console: grouped URLs, so you fix a template rather than a page, and a lag while the rolling window catches up.
• WebPageTest: no verdict, and a learning curve; free tests come with limits.
• GTmetrix: lab only, so a good score is not a passing assessment; monitoring sits on paid plans.
• DevTools and the extension: your machine, your connection, your one visit.

1. 1

   Monthly: Search Console

   Open the Core Web Vitals report. If every group is good on mobile, speed is not your problem this month.

2. 2

   When a group fails: PageSpeed Insights

   Take one URL from the failing group. Read the field data, then the largest saving Lighthouse names underneath it.

3. 3

   When the cause is not obvious: WebPageTest

   Run the same URL on a mid-range phone over a slow connection. The filmstrip shows when the page looks finished, and the waterfall shows what arrived after that.

4. 4

   Fix one thing and prove it

   Rerun the lab test for a quick signal, then wait for the field data, because it is measured over a rolling 28-day window of real visits.

That set costs nothing and covers a product site of any size a founder is likely to have.