Treat broad traffic-impact claims as directional until the query set, period, geography, and counterfactual are visible.
- Inspect the query sample
- Separate presence from causation
- Measure business exposure, not headlines
01
Start with the sample
Ask which queries were tracked, how they were selected, and whether the same result feature appeared throughout the measurement window.
02
Presence is not causation
An AI result can appear alongside seasonality, ranking movement, SERP redesigns, demand changes, and competitor gains. A correlation is a lead for investigation, not a finished causal claim.
03
Translate exposure into decisions
Segment queries by commercial role and page type. A change to low-value informational clicks may demand a different response than a change to qualified product discovery.
This article is a decision aid, not a universal ranking rule. Verify the conditions on your own site and preserve a baseline before changing them.