Lighthouse Score Test: How to Fix Your Site's Speed

A Lighthouse score test runs a simulated page load in your browser and returns six scores covering speed, search-engine visibility, accessibility, best practices, and progressive-web-app readiness. It is a controlled lab measurement that isolates one variable at a time to show which optimization improves the score before spending hours debugging live traffic.

Run the test against your own site, a staging build, or a competitor's page. Compare the results in the same session. The scores are comparable because the same engine runs across all of them. The article also explains when to use PageSpeed Insights instead, how Core Web Vitals relate to the full Lighthouse report, and a simple framework for deciding how often to retest as your site changes.

What is a Lighthouse score test?

Lighthouse is a free, open-source tool originally built by Google and maintained by the Chrome team. It audits a single page against a set of automated checks and returns results on five categories: Performance, SEO, Accessibility, Best Practices, and Progressive Web App (PWA). Each category receives a score, where the highest band falls in the green band, the middle band in orange, and the lowest band in red. The test runs a simulated page load inside a headless browser, intercepts every network request, measures timing events, and evaluates the rendered content against accessibility rules and web standards.

The six scores come from five categories because PWA is its own row. The overall Performance score is derived from four Core Web Vitals — Largest Contentful Paint, Interaction to Next Paint, Cumulative Layout Shift, and Largest Contentful Paint — weighted together into one number. The other four categories are independent checklists: SEO looks at meta tags, heading structure, link attributes, and crawlability; Accessibility checks contrast ratios, form labels, alt text, and keyboard navigation; Best Practices covers HTTPS usage, CSP headers, deprecated APIs, and console errors; PWA tests for service workers, manifest files, and offline fallback.

A Lighthouse score measures how a page performs in a controlled test environment. It does not account for the network conditions a visitor encounters, the server response time or database query latency that affect load time, or which queries from Search Console bring visitors to your pages. A low score on Performance means the simulated page loaded slowly under controlled conditions. It does not tell you whether your actual visitors bounce or stay. Google uses Core Web Vitals as a ranking signal, but Lighthouse itself is only a measurement tool. It is, however, a practical proxy for page experience because it isolates one variable at a time. You can fix an image format, rerun the test, and see exactly how that change moves the score before applying the same fix to every other page.

The test works on any public URL. You can run it against your live site, a staging environment, a local development copy served through a tunnel, or even a competitor's page to benchmark your own results. The same engine runs across all of them, so the scores are comparable within a single audit session.

How to run a Lighthouse score test

You can run a Lighthouse score test in three ways that cover every stage of the team: a browser-based tool for business owners, a command-line runner for developers, and a staging setup for testing before going live. Pick the method that matches your access level.

1. Open the page you want to audit in Chrome or Edge.
2. Press F12 or Ctrl+Shift+I (Cmd+Option+I on Mac) to open DevTools.
3. Click the Lighthouse tab at the top of the browser's developer tools. If you do not see it, click the two-dot menu → More tools → Lighthouse.
4. Select the categories to audit. For speed work, keep only Performance checked. For a full health check, check all five.
5. Choose the device profile — Desktop or Mobile. Mobile uses a slower network throttling profile and a smaller viewport, which is usually the one that matters for your visitors.
6. Click Analyze page load.
7. Review the results. Each metric shows a numeric value, a pass/fail icon, and a suggestion for improvement.

This method requires no software beyond a recent browser. It works on any public URL, including a staging site or a local build served through a tunnel.

Using PageSpeed Insights (free online tool)

PageSpeed Insights runs a Lighthouse score test under the hood and then adds real-world field data from the Chrome User Experience Report. You can use it when you want to see both lab scores and actual visitor measurements in one report.

1. Go to pagespeed.web.dev or search "PageSpeed Insights" in your browser.
2. Paste the full URL of the page you want to test.
3. Choose Mobile or Desktop in the top-right toggle.
4. Click Analyze.
5. Read the two main sections: Opportunities (what the lab test found) and Field Data (what real users experience). The field data shows Core Web Vitals broken down by percentile — the standard percentile is the one Google uses for its ranking threshold.

This tool is free, requires no developer account, and gives you the same Lighthouse engine that a browser's developer tools uses, plus the added layer of real-user data.

Using the Lighthouse CLI (for developers)

The command-line interface lets you run audits in CI pipelines, bulk-test pages at scale, and integrate results into dashboards. It is the method used by teams that need automation rather than manual checks.

1. Install a JavaScript runtime if you do not have it.
2. Run npm install -g lighthouse in your terminal.
3. Run lighthouse https://your-page.com --output html --output-path ./report.html to generate a full report file.
4. For performance-only audits, add --only-categories=performance. For mobile throttling, add --preset=mobile.
5. For bulk testing, pipe a list of URLs into a script or use the Lighthouse Metrics API (available at lighthouse-metrics.com) to run tests programmatically and export scores into a spreadsheet.

This approach is the most flexible but requires technical setup. Use it when you manage more than a handful of pages or when the test needs to run as part of a build pipeline.

Running on a staging or local site

To test before going live, serve your staging environment through a public URL. Tools like ngrok, a tunneling service, or a platform's preview deployment expose a local build to the internet temporarily. Run the Lighthouse test against that staging URL exactly as you would against production. The scores will reflect the same issues you are about to push live, so you can verify fixes before they go public.

Running for a mobile device

Always audit mobile first if your traffic skews toward phones. Mobile uses slower network throttling, a smaller viewport that triggers layout shifts differently, and a CPU throttling profile that matches mid-range devices. A page that scores 85 on Desktop may score 42 on Mobile because the simulated network slows rendering enough to push Core Web Vitals out of the green band. Run both profiles side by side and compare the delta. </step-by-step>

What each Lighthouse score means for your business

A low score in any category translates into a concrete business outcome. You do not need to read the technical metrics to see what is at stake — each score maps to a specific type of visitor loss, compliance gap, or trust signal.

Performance matters most for conversion. A low score means slower pages, which means more visitors leave before seeing your offer. The mechanism is simple: every second of delay increases bounce probability. A slow page buries your call-to-action below the fold longer than the visitor's patience allows.

SEO score matters because it covers what Google needs to understand and index your content. A low score means your page is not being found for the keywords your customers search. Missing title tags, unstructured headings, or broken internal links are common causes. These are fixes that any editor can apply without touching code.

Accessibility score covers legal risk as much as it covers user experience. A low score means you are excluding customers with disabilities. Screen readers cannot interpret unlabeled images, keyboard-only users cannot navigate forms, and low-contrast text is unreadable on certain devices. This is also a legal exposure in many jurisdictions. Fixing accessibility issues improves SEO at the same time because structured markup and proper headings serve both audiences.

Best Practices score is about trust and security. A low score means your site does not follow modern web standards. It may serve content over HTTP instead of HTTPS, omit a Content Security Policy, or trigger console errors that browsers flag. Visitors who see a browser security warning before they read your content have already decided the site is untrustworthy.

PWA score matters for mobile engagement. A low PWA score means your site is not optimized for an app-like experience. On repeat visits, visitors expect instant loads, offline availability, and push notifications. Without those, mobile traffic treats your site as disposable — open it once, close it, return nowhere.

The overall Lighthouse score is a weighted average of these categories. You do not need to chase a perfect score across all five. Prioritize Performance first, then SEO, then the others in order of your business model.

Lighthouse vs PageSpeed Insights

PageSpeed Insights runs a Lighthouse score test under the hood and adds real-world data from the Chrome User Experience Report. The difference between the two tools comes down to one question: do you want a lab measurement or a field measurement?

Lab data versus field data is the core distinction. A lab score measures the page in isolation: fixed network speed, fixed CPU throttling, fixed viewport. It tells you what the page can do. Field data measures what actual visitors experience across devices, networks, and locations. A page might score poorly on lab Performance but show a high percentile Largest Contentful Paint for real users because the server response time is slow and no CDN is configured. The lab test measures how fast the browser renders after the first byte arrives, not the server response time. PageSpeed Insights shows the field data that fills that gap.

Core Web Vitals are the only metrics that appear in both reports with the same definition — Largest Contentful Paint, Interaction to Next Paint, and Cumulative Layout Shift. Lighthouse measures them in the lab profile; PageSpeed Insights shows the field value alongside. When field Core Web Vitals fall outside the green band while lab scores remain green, the issue usually lies in infrastructure or server-side configuration rather than front-end code. You would not fix JavaScript by rewriting it; you would add a CDN or tune the database. That diagnosis only comes from the field data layer.

PageSpeed Insights is the best free tool for business owners who want to understand their site's speed without installing software or reading technical reports. Run it monthly as a health check. Use a browser's developer tools or the CLI when you have specific issues to fix and need the detailed opportunity list with before-and-after numbers.

Lighthouse vs Core Web Vitals

Core Web Vitals are three specific metrics that Google uses to rank pages. They are a subset of the Lighthouse score test. A Lighthouse report measures Core Web Vitals, but it also measures many other things: accessibility rules, SEO structure, security headers, and service-worker configuration. Core Web Vitals are not more important than other page experience factors for ranking because other page experience aspects do not directly help a site rank higher in search results. The other four categories — SEO, Accessibility, Best Practices, PWA — matter for user experience and compliance, but they do not appear in Google's ranking algorithm in the same way.

The Lighthouse Performance score for that page might be low. The three Core Web Vitals breakdown looks like this:

In this scenario, the lab test catches the LCP problem clearly. The field data confirms it is even worse for real visitors because the server response time adds nearly half a second on top of the render delay. The CLS value is close to the green band in the lab but crosses it in the field — meaning layout shifts are happening intermittently under real network conditions that the fixed lab profile does not reproduce.

The fix path here is sequential. First, optimize the hero image — convert to a modern image format, apply responsive sizing so mobile loads a smaller file. Second, defer non-critical CSS and JavaScript so the above-the-fold content renders before the rest of the page. Third, add a CDN cache header for static assets so repeat visits can skip the download. Each step can be rerun through a browser developer tool to see the delta before pushing live.

Core Web Vitals are not the only thing Lighthouse measures. The SEO category alone checks over 80 rules: title tag length, meta description presence, heading hierarchy, link text quality, structured data validity, and crawlability signals. A page can pass all three Core Web Vitals and still score 40 on SEO because the title is missing or the H1 is empty. That is why the full Lighthouse report exists — it catches the problems that Core Web Vitals alone miss.

How to fix a low Lighthouse score

When the test returns red scores, you do not need to rewrite the entire site. Follow this prioritized list in order. Each fix addresses the largest measurable impact first.

1. Enable compression. Verify that your server serves compression for HTML, CSS, and JavaScript files. A typical uncompressed HTML file can be reduced significantly with compression. Check the response headers in a browser developer tool Network tab; look for the content-encoding header. If the header is missing, the web server or CDN is not compressing.

2. Reduce JavaScript. Identify scripts that block rendering. Defer non-critical JS with the defer or async attribute. Remove unused libraries — a full jQuery bundle on a page that only needs one DOM interaction adds unnecessary bytes. Split large bundles using code splitting so each route loads only what it needs.

3. Use a CDN. A content delivery network caches your static assets at edge locations near your visitors. This reduces Time to First Byte and improves repeat-visit load time dramatically because the browser skips the origin server entirely for cached resources. Most modern hosting platforms include a free CDN layer.

4. Use a caching plugin or cache headers. Set aggressive cache-control headers on static assets (images, fonts, CSS) so browsers store them locally. A well-configured site should return a 200 status from cache on second visits, not fetch the resource again. For WordPress sites, a caching plugin handles this automatically; for other platforms, set the headers through your hosting control panel or CDN dashboard.

   </checklist>

After applying them, rerun the Lighthouse test against the same URL in a browser developer tool or PageSpeed Insights. Compare the before-and-after numbers in the Performance tab. If the score moves into the green band, you have validated the fix. If it stays orange, move to CSS optimization: minify stylesheets, remove unused rules, and inline critical above-the-fold CSS so the first paint is not blocked by a stylesheet download. Then check HTML: verify that meta tags are present, headings follow a logical hierarchy (one H1, then H2s under it), and internal links use descriptive anchor text instead of "click here."

How Often to Run the Test

A Lighthouse score test is only useful if you run it consistently. The frequency depends on how often your site changes and how much traffic it receives. Use this framework to decide.

High-traffic sites should be tested weekly because small regressions accumulate — a new plugin adds a JavaScript file, a theme update shifts a CSS rule, a CDN cache purge forces a full reload for a week. Each of these pushes Core Web Vitals one notch lower. A weekly cadence catches the drift before it becomes a visible traffic problem.

Medium-traffic sites need testing after every major site update and monthly as a baseline check. If no update happened that month, one test is enough. The delta between last month's report and this month's tells you whether something drifted.

Low-traffic sites — under a few thousand visits per month — can run monthly and after any significant content change. The cost of running the test is near zero (a few minutes in a browser developer tool), so there is no reason to skip it. The real question is whether you act on the results.

The test should also run after every new page is published if that page uses a different template or layout than your existing pages. A new template may introduce unoptimized images, missing meta tags, or an unstructured heading hierarchy that skews the SEO score.

Using Lighthouse for SEO

A Lighthouse score test does not directly affect SEO. It does not submit pages to Google, it does not adjust your search ranking, and it does not replace Search Console data. What it does is give you a structured checklist of the technical signals Google needs to understand and index your content. Improving the SEO score improves your page experience for crawlers and for visitors at the same time.

The Lighthouse SEO report flags these specific issues:

Each of these fixes is something an editor or content manager can apply without developer help. The SEO score measures structural integrity, not the quality of the copy. Google's crawler reads your HTML in a linear pass. If the heading hierarchy is out of order, the crawler cannot determine which topic is primary. If the title tag is too long, it may truncate before the keyword appears. If the meta description is missing entirely, Google generates its own snippet from the page body — often poorly. These are not ranking guarantees, but they are the signals that determine whether your page appears in a useful position when someone searches for what you offer.

The SEO category also checks crawlability: whether your robots.txt allows indexing, whether your canonical tags point to the right URL, and whether your internal links form a logical silo structure. A site with hundreds of orphan pages — pages no other page links to — will never earn authority because Google's crawler has no path to discover them. Fixing the SEO score is a prerequisite to any content strategy. You can write the best article in your industry, but if the page is structurally invisible to crawlers, it does not exist in search.

Using Lighthouse for Accessibility

A Lighthouse score test does not directly affect accessibility. It does not replace a manual screen-reader audit or a legal compliance review. What it does is flag the most common automated failures that prevent people with disabilities from using your site. Improving the accessibility score improves your reach and reduces legal exposure at the same time.

The accessibility category covers over 150 individual checks. The most impactful ones are the ones that affect both accessibility and SEO simultaneously. A missing alt attribute on an image excludes a screen-reader user and also removes a text signal Google uses to understand the image's context. An unlabeled form field excludes a keyboard-only user and also removes a structured data point that helps search engines parse your content. Heading hierarchy matters for the same reason: it structures the page for crawlers and for assistive technology at once.

Fixing these issues requires an editor to add alt text, a developer to set contrast ratios above the minimum threshold, and a QA tester to verify that every interactive element is reachable via Tab key. These are low-cost fixes that compound across hundreds of pages. A site with a low score on Accessibility is operating with a structural gap that affects both compliance and conversion.

Using Lighthouse for Best Practices

A Lighthouse score test does not directly affect best practices. It does not replace a security audit or a code review. What it does is flag the most common web-standard violations that browsers and visitors notice immediately. Improving the best-practices score improves trust, security, and browser compatibility at once.

The Best Practices category is the one most owners ignore because the failures do not look like failures to them. A visitor who sees a browser security warning — a red triangle, a crossed-out lock icon — has already decided the site is untrustworthy before they read a single word. That trust loss is immediate and irreversible. The fix is usually one setting in your hosting or CDN panel.

Security headers are the highest-impact Best Practices fixes. A Content Security Policy tells the browser which domains your site can load scripts from. Without it, any compromised third-party script can execute arbitrary code on your page. This is the mechanism behind most web-based attacks. Setting a CSP is a single line of configuration that protects every visitor simultaneously.

HTTPS is the baseline. A site served over HTTP triggers browser warnings that block some functionality entirely — geolocation, push notifications, and service workers will not work without it. Modern browsers also downgrade pages served over HTTP in search results relative to HTTPS equivalents. The fix is a decision to enable SSL through your hosting provider or a free certificate from a certificate authority.

These five categories — Performance, SEO, Accessibility, Best Practices, PWA — form the full Lighthouse report. You do not need to chase perfection across all five categories. You need to prioritize by business model: e-commerce sites start with Performance and SEO; B2B service sites start with SEO and Best Practices; mobile-first apps start with PWA and Performance. The test gives you the map, and you decide the route.

Get a second opinion on your site

If your Lighthouse scores have been low for months and no one on your team has the time to run the test, compare your results against what a trained audit catches. If you want to see how a professional writes and publishes content that moves rankings instead of just chasing numbers, a free sample article lets you judge the quality before committing to anything.

The way in is to talk to us about becoming a partner at /contact?track=partner.

Related reading: Lighthouse Performance Optimization Checklist · How Is Lighthouse Score Calculated · SEO Audit for Small Business

Every post here was written by Vannor. Researched, written and reviewed by the Vannor Engine on our own infrastructure.How the engine works →