Back to Learn
Pulseperformance

User Timing Marks and Measures

What This Audit Checks

Lighthouse detects calls to performance.mark() and performance.measure() in your page and lists them in the audit results. This is an informational audit — it does not pass or fail. Instead, it surfaces the custom timing data you have instrumented so you can verify your marks are firing and review their durations alongside standard Lighthouse metrics.

Why It Matters

Standard metrics like LCP and TBT tell you about generic milestones, but every application has unique moments that matter — when the search results render, when the video player is ready, or when a critical API response arrives. The User Timing API lets you define and measure these moments precisely. Without custom instrumentation, you are guessing which parts of your code are slow. With it, you have hard numbers tied to the exact operations that affect your users.

How to Fix It

  • Add marks at key milestones. Call performance.mark('search-results-rendered') at the exact moment a critical UI update completes. Place marks at the start and end of operations you want to measure.

  • Create measures between marks. Use performance.measure('search-duration', 'search-start', 'search-results-rendered') to calculate the elapsed time between two marks. This gives you a named duration you can track.

  • Keep naming conventions consistent. Use a prefix like app: or your feature name so marks are easy to filter in DevTools and analytics. For example, app:hero-visible, app:data-fetched.

  • Send measures to your analytics. Use PerformanceObserver to listen for measure entries and forward them to your analytics pipeline. This turns lab instrumentation into real-user monitoring data.

  • Clean up in production. Remove or gate verbose debug marks behind a feature flag. A handful of meaningful measures is more useful than hundreds of noisy marks.

How Pulse Tracks This

Pulse surfaces User Timing marks and measures captured during each Lighthouse run. You can see your custom milestones alongside standard performance metrics, making it easy to correlate application-specific events with overall page performance trends.

Resources