CrUX, Lighthouse lab scores, and real-browser tools measure related ideas on different clocks. Agencies get into trouble when they merge those clocks into one “PageSpeed number.”
Three instruments, three jobs
Chrome UX Report (CrUX) summarises real Chrome user experiences for eligible origins and URLs, with privacy thresholds and lag. Lighthouse lab scores (including PageSpeed Insights lab) run a simulated load under controlled conditions. Real-browser tools such as GTmetrix observe a page in Chrome from chosen locations and expose waterfalls suited to debugging.
Each instrument can be “right” while disagreeing with the others. That is expected. Treating disagreement as a tool defect usually means the wrong clock was used for the decision.
How Watcher and GTmetrix sit on those clocks
Apogee Watcher schedules PSI: lab Lighthouse metrics for regression history, plus CrUX when PSI returns field data. GTmetrix emphasises observed Chrome diagnostics and location choice. Neither replaces CrUX’s field role, and neither should be sold as a single source of truth for ranking.
Rules we use in client conversations
- Never average lab and field into one score.
- Label every chart with the clock and the date range.
- Use diagnostics when you need element-level proof; use schedules when you need history.
- Say “CrUX not available yet” instead of inventing confidence.
If you want the longer treatment we use on the Watcher blog, start from lab-versus-field explainers on apogeewatcher.com and keep this page as the short agency rulebook.