See exactly how each metric contributes to your performance score.
Lighthouse calculates your performance score using a weighted average of five key metrics. Each metric is scored from 0-100 using a log-normal distribution based on real-world performance data from the HTTP Archive.
The scoring algorithm converts raw metric values (like milliseconds or unitless shift scores) into a 0-100 score using statistical curves. A metric value at the 10th percentile of all sites scores 90 (good), while the median scores 50. This means your score reflects how your site compares to others on the web.
80% of your score comes from TBT, LCP, and CLS.
Lighthouse doesn't use simple linear scoring. Instead, it applies a log-normal distribution to convert raw metric values into scores. This statistical approach means:
Each metric has two key control points: the p10 value (scores 90) and the median (scores 50).
Google's essential metrics for page experience ranking
Core Web Vitals are Google's official metrics that affect your search rankings. As of March 2024, they consist of LCP, CLS, and INP (which replaced FID). To pass, 75% of page visits must meet the "good" threshold for all three.
Time for the largest visible element to render—when users perceive the page as "loaded."
Visual stability—how much elements move unexpectedly as the page loads.
Responsiveness—delay between user interactions and visual response. Lighthouse uses TBT as a proxy.
Additional metrics that affect your Lighthouse score
Beyond Core Web Vitals, Lighthouse measures additional metrics. TBT carries the highest weight (30%) because it directly correlates with how responsive your page feels.
Main thread blocking time >50ms. Lab proxy for INP.
First text or image rendered.
How quickly visible content is populated.
Server response time—not scored but delays all metrics.
Lighthouse assigns color-coded ratings based on your final weighted score.
Your homepage might score 100, but what about the rest? Scan every page to find issues site-wide.