You can now more easily see what settings were used to run an experiment by clicking on the "+" icon near the top of the page.
We also now trigger two tests when running an experiment. First a baseline test with the production HTML and then another test with the experiment HTML. This avoids inaccuracies when comparing the two test results, for example due to differences in HTML download times.
Real-user monitoring improvements
We now show percentile trends for each web vitals metrics. You can easily see how fast the average page view is, as well as how the slowest 10% of page experiences are.
Different users will experience different page elements as the LCP element, for example depending on device size and the initial scroll position. Our RUM tool shows what the most common LCP elements are, including a screenshot of the element.
Chrome and Lighthouse upgrades
We now run tests using Lighthouse 10.3.0 and Chrome 114.