Methodology · Website
How the Website score is calculated
Method version performance-v3 · Last updated
In short
The Website score is the Google Lighthouse performance score of a website on a scale of 0 to 100, measured on the homepage and up to four representative template pages. The homepage counts for 40% and the other pages share the remaining 60% equally, so the score reflects the whole site rather than one page.
Formula
Website score = 0.4 × homepage score + Σ (0.6 ÷ n) × template page score n = number of template pages measured (0–4); with n = 0 the homepage counts 100%
How it is calculated, step by step
- Resolve redirects to find the final homepage URL.
- Discover the site's pages from its sitemap (or the homepage links) and pick up to four representative pages, one per page template (e.g. product, article, category), from the same language section as the homepage.
- Run Google Lighthouse through the PageSpeed Insights API on each page, separately for mobile and desktop.
- Take the weighted mean of the page scores: 40% homepage, 60% shared by the other pages. Pages that fail to load are left out and the weights are spread over the rest.
- The same weighting is applied to the other Lighthouse categories (accessibility, best practices, SEO) and to the lab metrics such as Largest Contentful Paint.
What goes into it
- Lighthouse performance score
- Google's 0–100 score per page, based on lab metrics such as Largest Contentful Paint, Total Blocking Time and Cumulative Layout Shift.
- Sampled pages
- The homepage plus up to four template pages, the same set for every brand.
- Strategy
- Mobile and desktop are measured and shown separately; the dashboard uses mobile.
Data sources
- Google PageSpeed Insights API (Lighthouse)
- The website's sitemap or homepage links for page discovery
How often it is updated
Automatically on the 1st of every month for every tracked brand, and on demand (re-tests depend on the plan).
How to read it
- 90–100 is good, 50–89 needs improvement, below 50 is poor (Google's own bands).
- Compare brands measured with the same method version only; the label above each score shows the version and the date.
- Lab scores vary by a few points between runs; look at the rank among competitors and the trend over months rather than single-point differences.
Example
A homepage scoring 60 and two template pages scoring 80 and 70 give 0.4 × 60 + 0.3 × 80 + 0.3 × 70 = 69.
Limitations
- Lab measurements simulate one device and network; real users can experience the site differently (see Core Web Vitals for real-user data).
- A site that refuses automated requests can only be measured on the pages it lets through, or not at all.
- Five pages are a sample, not the whole site.
Frequently asked questions
Why not just measure the homepage?
Visitors land on many pages; product and article templates often load very differently from the homepage. Weighting the homepage at 40% keeps it important while the rest of the site counts too.
Why does my score differ from PageSpeed Insights?
PageSpeed Insights shows one page and one run. The Website score combines up to five pages, so it differs from any single page result.
Is mobile or desktop used?
Both are measured. The dashboard and rankings use mobile, because most visitors and Google's indexing are mobile-first.
Related
- Core Web Vitals — Google's real-user experience metrics (LCP, INP, CLS) from the Chrome User Experience Report, rated good, needs improvement or poor.
- Site health score — 45 technical and content checks on a crawl of up to 200 pages, grouped into six categories; each category loses points by how often its checks fail, weighted by severity.
- Rank and fair share — Rank is your position among the tracked brands on a score (1 = best); fair share is 100% divided by the number of brands.
