Measures unexpected visual movement using layout-shift scores. A good result is 0.1 or less.
Cumulative Layout Shift measures visual instability: content unexpectedly changes position while someone is using a page. It is a unitless score, not a number of seconds. The current metric considers clusters of unexpected shifts rather than simply adding every movement over an entire visit. Not every animation or user-requested layout change is treated as an unexpected shift.
A good CLS is 0.1 or less at the 75th percentile of page loads for the relevant device category. The practical concern is whether visitors lose their place, misread information or try to select a control that moves. A technically completed download is not a good experience if the interface keeps rearranging itself afterward.
Investigation should identify what caused the movement, not only what moved. A heading may shift because an image above it appears without reserved space, or a third-party component may expand after loading. The visible symptom and the responsible element can therefore be different. Record the sequence as it occurs rather than relying only on a final screenshot.
Takeaway: Keep page layouts predictable as resources and widgets load. Reserve appropriate space and verify the final design across realistic loading conditions rather than assuming a stable editor preview reflects public behavior.
Example
A homeware store places a shipping notice above its product controls after an external widget loads. The controls jump just as a customer reaches for the quantity selector. The team investigates that late insertion and the space allocated to it, rather than changing the wording of the quantity label.
What to check
Test a fresh visit as well as a repeat visit, including the period after the initial screen appears. Previously cached resources may hide timing differences. Verify that a layout fix does not cover content or leave an unusable blank area on smaller screens.
For more detail, see Web.dev CLS guide.