See exactly how each metric contributes to your performance score.
Lighthouse performance score is a weighted average of these five key metrics. Improving any of them will raise your overall score.
Time until first text or image is painted
How quickly content visually populates
Time until largest content element is visible
Sum of blocking time for long tasks
Movement of visible elements during load
Lighthouse calculates performance score as a weighted average of six metrics: First Contentful Paint (10%), Speed Index (10%), Largest Contentful Paint (25%), Total Blocking Time (30%), and Cumulative Layout Shift (25%). Each metric is scored 0-100 using log-normal curves based on real-world data from HTTPArchive.
Lighthouse scores can vary 5-10 points between runs due to: network conditions, server response time variations, third-party script loading, background processes on your device, and variability in JavaScript execution. Run multiple tests and use the median score for accuracy.
A score of 90-100 is considered good (green), 50-89 needs improvement (orange), and 0-49 is poor (red). Focus on reaching 90+ but don't obsess over perfect 100—the last few points have diminishing returns and normal variance makes it impractical.
Total Blocking Time (30%) and Largest Contentful Paint (25%) together account for 55% of your score. CLS adds another 25%. To improve your score quickly, focus on reducing JavaScript execution time (TBT) and optimizing your largest above-the-fold element (LCP).
PageSpeed Insights tests from Google's servers with a simulated mobile device, while Chrome DevTools uses your local machine and network. PSI also shows field data from real Chrome users (CrUX) which reflects actual user experience and may differ significantly from lab results.
Lighthouse is a lab test. For production sites, use our free tools to check your real Chrome User Experience data.