Website review
This review may include AI-generated content and is not individually verified. Domain, certificate and other facts come from the observed data on this page.
What is Lighthouse Metrics?
Lighthouse Metrics is a web performance testing service that runs Google Lighthouse audits from multiple geographic locations and shows how your site performs in each region. Instead of testing from a single machine, it runs the same URL across a global set of regions—up to 16, according to the site—so you can see how scores and Core Web Vitals change depending on where visitors are located.
The tool reports familiar Lighthouse metrics per region. On the site, sample results show scores alongside:
- FCP (First Contentful Paint)
- LCP (Largest Contentful Paint)
- TBT (Total Blocking Time)
It also keeps a test history, so you can revisit earlier reports and compare how a page has changed over time. The site lists Monitoring, API, and Enterprise options in its navigation, alongside a docs section.
Who it's for: teams that care about performance beyond their own office network—for example, a site with users in Europe, Asia, and North America that wants to know whether a change helps one region while hurting another.
A practical next step: pick one important page and one competitor or baseline page, then run both across the same set of regions. Compare the spread between your best and worst region rather than the average score; a wide gap usually points to a hosting, CDN, or third-party script issue that a single-location test would hide. If you only need a quick local check, Google's own Lighthouse in browser DevTools may be enough. For broader comparison, see WebPageTest and PageSpeed Insights.
How does testing from multiple locations affect my website's performance scores?
Testing from multiple locations doesn't change your website's actual speed—it reveals how much that speed varies depending on where the visitor is. A single test from one region gives you one data point; a multi-region run shows the spread, which is usually what matters for real users.
The page evidence illustrates this clearly. The same site scores 99/100 in Japan with an LCP of 1.4s, but 90/100 in Finland with an LCP of 3.2s. The differences come from network distance between the test location and your hosting or CDN edge, plus regional routing and peering. First Contentful Paint and Total Blocking Time also shift by region, so a single headline score can hide a weak spot.
What the variation tells you
- Large spread between regions → likely a CDN, caching or origin-placement issue. Users far from your origin feel it most.
- Consistent scores everywhere → your delivery is well distributed; further gains probably come from front-end work rather than infrastructure.
- One region lagging → check whether that region has a local edge, and whether third-party scripts load slowly there.
A practical scenario
Suppose you run an e-commerce site hosted in the US and most of your traffic is European. A US-only test might show 98/100 and look fine. Running the same page from Germany and Finland exposes slower LCP, telling you whether to add or tune European CDN coverage before you lose conversions.
Deciding how to use it
Use multi-region testing as a diagnostic, not a score to optimize. Pick the regions where your actual audience lives, compare them, and treat the worst region as your real baseline. For a quick cross-check on how a single location's run is scored, see Chrome Developers for Lighthouse methodology, and web.dev for guidance on interpreting Core Web Vitals.
Next step: test your three highest-traffic regions, note the slowest, and investigate its cause before changing anything else.
Can I schedule automatic Lighthouse tests for my website?
Yes, Lighthouse Metrics supports scheduled testing: its "Monitoring" feature is built for repeated, automatic runs rather than one-off checks, and test history lets you compare results over time.
What the page shows
- Monitoring appears as a top-level product area alongside Measure, API, Enterprise, Pricing and Docs.
- Test History is described as a way to "access and compare previous test results" and "see how your site has changed over time," with past runs listed by age (yesterday, 2 days ago, 1 week ago).
- Measure globally runs a single test across up to 16 regions, including US West, US East, Finland, Germany, Japan and Australia.
How scheduling fits a real workflow
A typical setup: you point the tool at your production URL, pick the regions your users actually come from, and let it run on a recurring basis. Each run adds an entry to test history, so you can spot a regression — say LCP climbing in one region — rather than discovering it from a support ticket.
| Approach | Best for | Trade-off |
|---|---|---|
| Manual "Measure globally" runs | One-off audits, pre-launch checks, debugging a specific report | No alerting; you only see problems when you remember to look |
| Monitoring (recurring tests) | Ongoing regression detection and trend tracking | Needs a defined cadence and someone to act on the results |
| Programmatic runs via the API | CI pipelines, custom dashboards, tests tied to deploys | Requires engineering effort to build and maintain |
Choosing your cadence
- Match frequency to release pace. Weekly is often enough for a mostly static marketing site; daily or per-deploy suits an app that ships often.
- Test from regions that matter to you. The sample data shows regional spread — for example, Japan scoring 99 with a 900 ms FCP while Finland showed a slower 3.2 s LCP in one run — so a single-location test can hide a real problem.
- Watch variance, not just the score. Repeated runs from the same region (Finland appeared at both 90 and 98) show that a single number can mislead. Compare trends across several runs.
- Decide where alerts land. A scheduled test is only useful if the result reaches a person or a channel that will act on it.
Next step
Open Lighthouse Metrics, check the Monitoring section and Pricing page for the cadence and region limits that fit your site, then set up one recurring test on your most important page — your homepage or main landing page — before expanding to key templates. If you need tests triggered by deploys rather than a calendar, evaluate the API route instead.
How do I use the Lighthouse Metrics API to integrate performance testing into my workflow?
The API endpoint shown on the site is POST /v1/lighthouse, and the core idea is simple: you send a URL, the service runs Lighthouse for you, and you get scores back. That fits CI pipelines, scheduled checks, or a small internal dashboard where you want performance numbers without maintaining your own Lighthouse runners.
What to expect from the API
- One request, many regions. The site's headline feature is running Lighthouse from up to 16 regions in a single test, with results like FCP, LCP and TBT per region.
- Region choice matters. The sample results show the same page scoring 90 in Finland with a 3.2s LCP but 99 in Japan with a 1.4s LCP. If your audience is regional, test from where they actually are.
- History is part of the product. Test history lets you compare current runs against earlier ones, which is more useful in a workflow than a single pass/fail number.
A practical integration pattern
- Pick one or two regions that match your real users rather than all 16 — more regions means slower runs and more noise to interpret.
- Trigger the test after a deploy, or on a schedule (for example nightly), not on every commit.
- Store the returned metrics alongside the commit or release so you can see regressions over time.
- Set thresholds on the metrics that matter to you, such as LCP, and alert only when a run crosses them.
- Keep a baseline run from a known-good release to compare against.
Choosing your setup
| Situation | Sensible approach |
|---|---|
| Small site, occasional checks | Manual runs from the web UI, no API needed |
| CI on every deploy | API call in a post-deploy job, one region |
| Global audience | API call across several regions, compare per-region scores |
| Long-term tracking | API plus stored history, reviewed weekly |
The trade-off is straightforward: the API saves you from running and patching Lighthouse yourself, but you are relying on someone else's runners and network conditions, so treat scores as comparable within this service rather than as absolute truth. Start by running one test against a single page in the region closest to most of your users, then add regions once you know which numbers actually move.
For background on the metrics themselves, the definitions at web.dev are worth reading before you set thresholds.
What are the pricing options for Lighthouse Metrics?
Lighthouse Metrics does not publish its pricing tiers in the page evidence available here. The site has a dedicated Pricing page linked from the main menu, so that is the place to see current plans and any free or trial limits. The product itself is a global Lighthouse runner: it tests a URL from up to 16 regions in one run and keeps a history of past reports.
H3 What you can infer from the product
- It is aimed at teams that care about performance beyond a single location, since the same page can score very differently by region (the sample data shows scores ranging from 90 to 99 and LCP from 900 ms to 3.2 s).
- It offers measurement, monitoring and an API, which usually maps to tiered plans: a free or low-volume entry point, a paid monitoring tier, and a higher API or enterprise tier.
- The presence of an Enterprise link suggests a custom or sales-led option for larger organisations.
H3 How to decide If you only need occasional one-off tests, start with whatever free allowance the Pricing page lists. If you need scheduled checks across many regions, or API access for CI pipelines, compare the monitoring and API tiers against your expected test volume. For a rough sense of the underlying method, Google's own Chrome Developers documentation explains how Lighthouse scores are calculated, which helps you judge whether the regional spread shown here is meaningful for your audience.
Next step: open the Pricing page on the site and check three things — number of regions per test, test frequency, and whether API access is included.
How can I compare historical Lighthouse test results to track performance changes?
Lighthouse Metrics is built around exactly this need: it keeps a test history so you can view past reports and see how your site has changed over time, and it runs Lighthouse from up to 16 regions in a single test. That combination matters for historical comparison, because a score change can reflect a real regression or just a different test location.
How to make comparisons meaningful
- Keep the region constant. If you compare a US West run from last week against a Finland run today, you are measuring geography, not progress. Pick one or two reference regions and stay with them.
- Compare metrics, not only the score. The page evidence surfaces FCP, LCP and TBT alongside the score, and these often move in different directions. A stable score can hide a slower LCP.
- Test at a consistent cadence. The history view shows entries labelled by age (yesterday, 2 days ago, 1 week ago), so a regular schedule produces a clean trend line rather than a scatter of unrelated runs.
- Note what changed on your side. Deployments, new third-party scripts, image swaps and CDN changes are the usual causes; line them up against the dates in your history.
A practical scenario
You ship a front-end change and the score drops. Rather than guessing, run the same URL in the same region you used before and open the previous report side by side. If LCP rose while TBT held steady, the likely culprit is a larger hero image or a slower server response; if TBT rose, suspect added JavaScript. The regional spread also tells you whether the problem is global or local to one audience — for example, a high TBT in one region with fast times elsewhere points to a third-party script loading from a distant origin.
Choosing your approach
| Need | What to do |
|---|---|
| Track one market over time | Fix a single region and compare runs on a schedule |
| Catch region-specific regressions | Run the multi-region test, then track the worst-performing region historically |
| Feed results into your own tooling | Use the API rather than manual history browsing |
For deeper background on what each metric means and how to interpret changes, the official documentation at Chrome for Developers is a reliable reference, and web.dev covers the metric definitions in plain language.
Next step: pick your two most important audience regions, run a baseline test today, and record the date alongside your next deployment so the first historical comparison is unambiguous.
User reviews (0)