Lighthouse scores are cited in countless SEO audits as ranking factors, with 90+ positioned as the target for “good SEO performance.” Google has explicitly stated that Lighthouse scores do not influence search rankings. The page experience ranking signal uses CrUX field data — real user performance from Chrome sessions — not lab scores from Lighthouse. A site with a Lighthouse score of 55 that passes CrUX at the 75th percentile has better page experience ranking signals than a site with a Lighthouse score of 98 that fails CrUX. Optimizing for Lighthouse score instead of CrUX data is optimizing for the wrong metric.
Lighthouse Scores Are Not a Ranking Factor and Why the Confusion Persists
Google’s documentation on page experience signals specifies that Core Web Vitals assessment for ranking purposes comes from the Chrome User Experience Report (CrUX), not from Lighthouse. John Mueller stated directly: “Google doesn’t use the X/100 lighthouse score for search, we use the core web vitals separately.” Martin Splitt has reinforced this position in developer advocacy content, consistently distinguishing between Lighthouse as a diagnostic tool and CrUX as the ranking data source.
PageSpeed Insights (PSI), the Google tool most associated with performance evaluation, displays both Lighthouse lab data and CrUX field data on the same page. The CrUX field data section is labeled “Discover what your real users are experiencing” and appears at the top, while the Lighthouse lab data section appears below with the label “Diagnose performance issues.” The visual hierarchy signals which data Google considers authoritative. Despite this clear labeling, the Lighthouse performance score (0-100) dominates attention because it provides a single, easy-to-understand number.
The position confidence on this distinction is confirmed through Google’s documentation, John Mueller’s public statements, and the architectural separation between Lighthouse (a lab diagnostic tool) and CrUX (the field data source for ranking signals). Google has made this distinction explicitly and repeatedly.
The confusion between Lighthouse scores and ranking factors persists through several reinforcing mechanisms:
PageSpeed Insights as a Google tool: because PSI is a Google product that prominently displays Lighthouse scores, practitioners assume Google uses those scores for ranking. The tool’s Google branding implies the score has ranking significance. In reality, PSI displays Lighthouse data for diagnostic purposes and CrUX data for ranking relevance — two distinct functions within one interface.
SEO tool vendor integration: commercial SEO platforms (Semrush, Ahrefs, Screaming Frog, SE Ranking) integrate Lighthouse scores into their dashboards alongside ranking data. This co-presentation creates an implicit association between Lighthouse scores and ranking outcomes. The tools display Lighthouse scores because they are easy to collect programmatically via the Lighthouse API, while CrUX data requires meeting traffic thresholds and is only available for origins with sufficient Chrome users.
Single-number simplicity: the Lighthouse performance score is a single integer from 0-100 that maps intuitively to letter grades (90+ = good, 50-89 = moderate, 0-49 = poor). CrUX data requires understanding metric-specific thresholds (LCP under 2.5s, CLS under 0.1, INP under 200ms) evaluated at the 75th percentile across a 28-day rolling window. The simpler metric naturally dominates in reports, presentations, and client communications.
Early CWV commentary: when Google announced Core Web Vitals as a ranking factor in 2020, early industry commentary conflated lab metrics with field metrics before Google clarified the distinction. Articles titled “How to improve your Lighthouse score for SEO” proliferated and continue to rank well in search results, perpetuating the misconception.
The Structural Problem with Targeting a Lab Score
The Lighthouse performance score is a weighted composite of six metrics: First Contentful Paint (FCP), Speed Index (SI), Largest Contentful Paint (LCP), Total Blocking Time (TBT), Cumulative Layout Shift (CLS), and Speed Index. The weights are calibrated for lab-based user experience assessment, not for predicting CrUX outcomes or matching Google’s ranking evaluation.
Specific structural disconnects between the Lighthouse score and ranking-relevant CrUX data:
TBT versus INP: Total Blocking Time receives 30% of the Lighthouse performance score weight, making it the most influential metric in the score. TBT measures the total milliseconds of main-thread blocking during page load. The CWV ranking signal uses INP (Interaction to Next Paint), which measures responsiveness to user interactions throughout the page session, not just during load. A change that reduces TBT (improving the Lighthouse score) may not improve INP if the blocking occurs during load and user interactions happen after load. Conversely, a change that improves INP (improving the ranking signal) may not change TBT if the interaction bottleneck is a slow event handler rather than a blocking script during load.
Score non-linearity: the Lighthouse score uses a logarithmic curve where improvements at the high end (moving from 90 to 95) require disproportionately larger metric improvements than at the low end (moving from 50 to 55). This curve shape incentivizes diminishing-returns optimization at the top of the scale. For CrUX, the evaluation is binary: each metric passes or fails at its threshold. There is no incremental ranking benefit from improving CrUX LCP from 2.0s to 1.0s — both pass equally.
Metric inclusion mismatch: Lighthouse includes Speed Index and First Contentful Paint in its composite score. Neither is a CWV metric used for ranking. Optimizing these metrics improves the Lighthouse score without directly improving the ranking signal.
What Lighthouse Scores Are Actually Useful For
Lighthouse serves two legitimate and valuable functions in SEO performance work that do not involve targeting a specific score:
Regression detection: Lighthouse CI integrated into the CI/CD pipeline detects performance regressions introduced by code changes before they reach production. If a code deployment drops the Lighthouse performance score by 10 or more points, something performance-relevant changed. The score change triggers investigation into what changed and whether it will affect CrUX metrics in the field. This early-warning function is valuable regardless of the absolute score.
Diagnostic recommendations: the Lighthouse audit list identifies specific, actionable issues: render-blocking resources, improperly sized images, unused JavaScript, missing text compression, unoptimized font loading, and dozens of other detectable problems. Each audit provides a description of the issue, an estimated impact, and guidance for remediation. These diagnostics are useful development feedback independent of the composite score.
The correct workflow uses Lighthouse as a diagnostic and regression-detection tool while targeting CrUX “good” status as the success metric. Use Lighthouse to find and fix issues. Use CrUX to verify that the fixes produce real-world improvement. Do not target a specific Lighthouse score number — it does not correspond to the ranking signal, and the score’s composite weighting may drive engineering effort toward metrics that do not affect CrUX outcomes.
The Target That Matters: CrUX “Good” at the 75th Percentile
The actual SEO performance target is having each Core Web Vital at or below the “good” threshold at the 75th percentile in CrUX field data:
- LCP: under 2.5 seconds
- CLS: under 0.1
- INP: under 200 milliseconds
This assessment is binary for ranking purposes. A page that barely passes all three (LCP at 2.4s, CLS at 0.09, INP at 195ms) receives the same positive page experience signal as a page with exceptional performance (LCP at 0.8s, CLS at 0.01, INP at 50ms). There is no incremental ranking benefit from exceeding the thresholds by a wider margin. The ranking system evaluates pass/fail, not a performance score.
This binary nature has a specific strategic implication: once CrUX passes, further performance investment produces user experience and business metric returns but no additional ranking signal. Engineering resources spent moving LCP from 1.5s to 1.0s do not improve rankings. Those same resources invested in content quality, internal linking, or authority building produce ranking improvements that CWV optimization beyond the threshold cannot.
Monitor CrUX data through Search Console’s Core Web Vitals report (updated daily, covering the 28-day rolling window) and PageSpeed Insights (URL-level CrUX data). Set alerts when any metric approaches its threshold from below (indicating regression risk) rather than targeting continuous improvement in the score.
Has Google ever confirmed that Lighthouse scores affect rankings?
No. Google has explicitly stated that Lighthouse scores do not influence search rankings. John Mueller and other Google representatives have clarified that Google uses CrUX field data for the page experience ranking signal, not Lighthouse lab data. The confusion persists because PageSpeed Insights displays both Lighthouse and CrUX data on the same page, and many SEO tools report Lighthouse scores as performance indicators.
Is a Lighthouse score of 100 necessary for good SEO performance?
No. A perfect Lighthouse score indicates that a page performs well under specific lab conditions, but it does not guarantee passing CWV in the field. Conversely, pages with Lighthouse scores below 100 frequently pass CrUX assessment because real users on fast devices and connections experience better performance than the throttled lab simulation. The SEO target is CrUX “good” status, not a Lighthouse score threshold.
Can Lighthouse accessibility and best practices scores affect SEO?
Not through any direct ranking mechanism. Google’s page experience system evaluates only the three Core Web Vitals (LCP, INP, CLS) plus HTTPS and mobile-friendliness. Lighthouse’s accessibility score measures WCAG compliance, and the best practices score evaluates security and development quality. These are valuable for user experience but are not connected to Google’s ranking algorithms.
Sources
- https://www.nostra.ai/blogs-collection/myth-lighthouse-scores-are-used-by-google-for-seo
- https://www.erwinhofman.com/blog/lighthouse-scores-do-not-affect-google-search-seo/
- https://www.searchenginejournal.com/do-google-lighthouse-scores-affect-seo/425347/
- https://vercel.com/blog/how-core-web-vitals-affect-seo
- https://www.debugbear.com/docs/core-web-vitals-ranking-factor
