Traffic falls, visibility rises: how to read the numbers now

Fewer visits and more search impressions can happen at the same time. The numbers describe different parts of the journey. Before attributing the gap to AI answers, establish which queries, pages and audiences changed.

Separate the measurements

Search impressions record appearances under the search engine’s counting rules. Clicks record interactions with a result. Analytics sessions record visits your tracking system observed. Enquiries, purchases and subscriptions are business outcomes.

AI citations add another measure: whether an assistant referenced a page in the answers you sampled. A citation is not a visit, and a crawler request is not a citation. Keep those distinctions visible in reports rather than combining them into one visibility score.

Investigate where the gap opened

Compare the same date ranges and segment by page, device, country and branded versus non-branded queries. Check ranking changes and whether the site began appearing for a broader set of queries with lower click-through rates.

Review the results pages for representative queries. AI answers, ads and other features may change how people interact with results, but a chart alone cannot identify the cause. Seasonality, changes in demand and competitors can matter too.

If Search Console clicks and analytics visits diverge, inspect tracking and consent changes before drawing a marketing conclusion. The two systems use different definitions and will not match exactly, but a sudden break deserves investigation.

Add AI evidence without overreading it

Where available, Google’s AI performance report can show impressions in AI Overviews and AI Mode. It cannot supply all the missing explanations about clicks or customer intent. See what that report measures.

For other assistants, repeat a defined set of questions and retain the answers and sources. Citation monitoring measures the chosen sample, not every conversation customers had. A bot visiting a page proves access; it does not prove the page was used or recommended.

Report the business effect

A useful client report might say: visits fell on these pages, impressions rose for these queries, and enquiries remained stable. It should also say which part remains unexplained.

Annotate site releases, tracking changes and major campaigns. Compare relevant page groups over time, and be cautious about attributing an improvement to a single edit when several things changed together. Reversible deployments help recover from problems; they do not turn an ordinary before-and-after comparison into a controlled experiment.