Why Performance Numbers Differ
Performance is rarely a fixed property of a page. The same site can feel fast on one user's device and sluggish on another's, depending on network type, hardware, and environmental factors. A page loaded over a wired office connection will often behave very differently from one loaded over café WiFi, and even users on 4G can hit slow or intermittent connections in elevators, trains, or other locations. Because real-world performance is so variable, the tools used to measure it generally fall into one of two categories.
Lab vs. Field Data
Lab data is collected in a controlled environment with predetermined device and network settings. Because conditions are fixed, results are reproducible and easy to debug. The trade-off is that lab tests may not reflect actual user bottlenecks or connect to real-world business KPIs. To get useful lab results, you need to know your users' typical devices and connections and replicate those conditions during testing.
Field data, also called Real User Monitoring (RUM), is gathered from actual page loads by real users. It reflects genuine user experience and can be correlated with business KPIs, but it offers a narrower set of metrics and limited debugging capabilities compared to the lab environment.
Measuring in a Controlled Environment
For lab-based measurement, Lighthouse accepts a URL and runs a series of audits against the page, producing a performance report. There are several ways to run Lighthouse, including directly from Chrome DevTools for a quick audit.
Measuring Real-World Experience
On the field side, the Chrome User Experience Report (CrUX) offers metrics showing how real-world Chrome users experience popular destinations on the web. With CrUX, you can see performance data aggregated across actual user sessions.
Combining Both Approaches
PageSpeed Insights supplies both lab and field data for a given page. It relies on Lighthouse to gather and analyze lab data, while its real-world field data derives from the CrUX dataset. This combination lets you compare synthetic results against observed user experience.
For deeper investigation, Chrome Developer Tools, built into the Google Chrome browser, allows profiling of a page's runtime and helps identify and debug performance bottlenecks directly.
