Lighthouse audits, explained

Plain-English deep-dives on the Lighthouse audits that fail most often, with concrete fixes and code you can copy. Every page covers one audit ID, what it measures, why Google cares, common causes, and prescriptive fixes, written for developers solving real problems, not SEO marketers.

Why these pages exist

Google's Lighthouse audit produces ~100 distinct checks. The official documentation is technically correct but optimized for completeness, not for "I have 30 minutes to fix this before deploy." These pages strip the audits down to:

1. What the audit actually measures
2. What the official "Good" / "Needs improvement" / "Poor" thresholds are
3. The 3–5 fixes that resolve 90% of real-world failures, with code
4. How to verify the fix without re-running the entire Lighthouse suite

Each page is also mirrored as raw markdown at /audits/<slug>.md for AI agents and LLM-based search to consume cleanly.

Performance audits

These directly affect Core Web Vitals and your Google ranking.

• Largest Contentful Paint (LCP), how fast the biggest visible element renders. Core Web Vital, target under 2.5 s.
• Cumulative Layout Shift (CLS), unexpected layout movement. Core Web Vital, target under 0.1.
• Total Blocking Time (TBT), main-thread blocking during load. Predicts how janky your page feels.
• First Contentful Paint (FCP), when the first text or image paints. Target under 1.8 s.
• Reduce unused JavaScript, bytes shipped but never executed. The most common bundle-bloat killer.
• Reduce unused CSS, stylesheets with selectors the page never uses. Strip with PurgeCSS or Tailwind JIT, inline critical, defer the rest.
• Eliminate render-blocking resources, CSS and synchronous scripts that delay first paint.
• Serve images in next-gen formats, JPEG/PNG that would be smaller as WebP or AVIF.
• Enable text compression, text assets served without gzip or Brotli.
• Ensure text remains visible during webfont load, web fonts without font-display hide your text while they download. One CSS descriptor fixes it.
• Speed Index, how quickly the page becomes visually complete. Target under 3.4 s; fix render-blocking, fonts, and hero images.
• Serve static assets with an efficient cache policy, short or missing Cache-Control on static files, so repeat visitors re-download them. The most-failed audit in our Shopify study at 91.4%.
• Avoid multiple page redirects, redirect chains that cost a round trip each before your HTML arrives. Also covers too many redirects loops.

Accessibility audits

These affect screen reader users and your Lighthouse Accessibility score.

• Document does not have a main landmark, the page lacks a <main> element. Screen readers can't jump to primary content.
• Heading elements not in sequentially-descending order, headings skip levels (h1 → h3). Breaks the document outline.
• Background and foreground colors do not have sufficient contrast, text fails WCAG AA contrast requirements.
• Image elements do not have alt attributes, missing alt on <img>. Breaks accessibility AND image SEO.
• Tap targets are not sized appropriately, retired from Lighthouse, but WCAG 2.5.8 still requires 24px and 44px is the practical target.

SEO audits

These directly affect how Google indexes and ranks your page.

• Document does not have a meta description, missing <meta name="description">. Tanks SERP click-through rate.
• Document does not have a valid rel=canonical, also known as a "non-canonical URL" or "canonical missing" error in other SEO tools. Add one self-referential canonical per page.
• Document does not have a title element, the strongest on-page SEO signal and the SERP headline.
• Document does not have a valid hreflang, missing or malformed hreflang for multi-language sites. Self-reference, x-default, reciprocal return links.

Best Practices audits

These affect security, trust, and your Lighthouse Best Practices score.

• Does not use HTTPS, page served over HTTP or with mixed content.
• Browser errors were logged to the console, failed requests and uncaught exceptions during load. Failed by 90.3% of the stores in our Shopify study.

Coming soon

Pages in the pipeline (likely shipping this month):

• preload-lcp-image, preloading the LCP element specifically
• uses-responsive-images, srcset and image CDNs
• efficient-animated-content, replace GIFs with MP4/WebM
• link-text, descriptive anchor text
• meta-viewport, the mobile viewport tag

How to use this library

If you're staring at a failing Lighthouse audit, find the matching page above and follow its fix steps. Each page is self-contained, you don't need to read them in order.

If you want a tailored brief for your specific URL, paste it at lighthouse-md.com, we'll generate a CLAUDE.md covering every failing audit on your page with ranked offenders and code-level fixes, ready to drop into your repo or open Claude Code with one click.

If you run a Shopify store, the same audit lives inside your admin as the Lighthouse-md Shopify app: weekly Core Web Vitals tracking, CLAUDE.md fix plans for your theme, and answers about your store's speed in Sidekick. 3 audits free.

Data studies

• We ran Lighthouse on 269 Shopify stores: median mobile performance was 48, median LCP 10 seconds, and not one recognizable brand cracked 50. The ranked list of audits nearly every store fails, with a downloadable dataset.