Methodology
How we measure website performance
What we measure
- TTFB Time to First Byte — how long until the server starts sending a response. This is the primary speed metric we rank by. Lower is better.
- Total time Total elapsed time for the HTTP response to complete, including all headers and body content received.
- Uptime % Percentage of checks where the site returned a successful HTTP response (2xx or 3xx status code) within the timeout window.
How we measure
Each check is an HTTPS GET request made by a Cloudflare Worker to the domain's root URL (https://domain.com/). We follow up to one redirect.
What we don't measure
- Full page render time (no browser, no JavaScript execution)
- DNS resolution time (not broken out separately)
- CDN cache hit/miss status
- Asset load times (images, CSS, JS files)
- Mobile vs. desktop performance
- Geographic variation (single Cloudflare edge region per check)
- Core Web Vitals (LCP, CLS, FID)
Limitations & caveats
Bot blocking: Some companies detect and block automated requests. A site appearing "down" may simply be blocking our monitor rather than being genuinely unavailable. We can't distinguish between these cases.
CDN caching: Many large sites serve cached responses from CDN edge nodes, which may appear faster than their actual origin server performance.
Redirects: If a domain redirects to a different URL (e.g., www. subdomain), the redirect time is included in our measurement.
Single region: Currently all checks originate from a single Cloudflare region. Companies with stronger East Coast vs. West Coast infrastructure may appear faster or slower depending on routing.
S&P 500 inclusion: We use the S&P 500 constituent list at the time of our data collection. Companies added or removed from the index may lag behind real-time changes.
Future plans
- Multi-region checks (US East, US West, Europe, Asia)
- DNS resolution timing broken out separately
- Historical trend data over weeks and months
- Public API for accessing raw data
- Alerting / status subscriptions