Key takeaways
- Clicks and business value differ.
- Investigate by page and intent.
- Do not assume AI caused every decline.
Compare equivalent periods
Segment by landing page, device and query type. A sitewide total can hide an informational guide gaining impressions while a commercial page loses relevant clicks. Check brand demand, campaigns and seasonal changes.
Separate visibility from outcomes
Rising impressions with flat clicks can reflect a different query mix or result layout. Falling conversions can come from a broken form. Inspect the full journey before deciding a page needs rewriting.
Record the visible search experience
Review a small set of affected queries in the relevant market. Save dated observations of result types. Do not extrapolate a single screenshot to all users, devices and locations.
Protect the useful next step
Give readers scope, examples, limitations, pricing context and a clear contact or signup route. Fewer visits can produce more qualified enquiries, but they can also mean lost business. Measure the outcome rather than assuming it.
Create a balanced review
Use a table with the page, search trend, observed result changes, conversion trend and proposed action. If evidence is inconclusive, say so. Continue monitoring rather than manufacturing a causal explanation.
References and further reading
Primary documentation and product references. Practical checklists and worked examples above are editorial guidance, not claims about private ranking algorithms.