Bounce-back rate (AI surface)

What it is

Bounce-back rate on an AI surface is the rate at which users return to the results page or refine their prompt after seeing an AI Overview, rather than acting on the answer. It reflects answer dissatisfaction at the surface level.

Why it matters

High bounce-back is associated with Overviews being silently removed for that query cluster, so it is an early signal that an AI surface may stop appearing for those queries, changing the visibility landscape.

How it works

Practitioners infer bounce-back from query refinement and return-to-results behaviour, then watch clusters with high rates as candidates for Overview removal or for content that better resolves intent.

When it applies

Watch it for query clusters where AI Overviews appear and you want to anticipate whether they will persist.

Examples

• Users re-querying immediately after an Overview that did not answer their question
• A query cluster where Overviews disappear after sustained high bounce-back
• A team flagging high bounce-back clusters as unstable for planning

How it is measured

• Share of sessions returning to results after an Overview
• Prompt refinement rate following an AI answer
• Bounce-back rate by query cluster
• Correlation between bounce-back and Overview disappearance