A local business owner looks at their reports and finds something that should not be possible: the rankings are the same as last year, and the traffic has halved. Nothing was penalised. Nothing was redesigned. The positions held. The visits did not.
The usual reaction is to assume something is broken and start changing things. Almost always, nothing is broken. What changed is that ranking and traffic have stopped being the same measurement.
What actually changed
A large and growing share of searches now end without anybody clicking anything. Published analyses through 2025 and 2026 put roughly six in ten searches in that category, and when an AI-generated summary appears at the top of the results the figure reported is far higher again — north of eight in ten.
If your page is the source that summary was built from, you are still doing the work and still getting the credit. You are simply not getting the visit. Position one now means “the answer came from you”, which is not nothing, but it is not a session in your analytics either.
The part that is genuinely good news
The clicks that still arrive are better. The same analyses find surviving clicks convert meaningfully better than before — people who read a summary and then click anyway are self-selecting for real intent. They have already got the definition. They are coming for a decision.
So a business can lose a third of its sessions and hold its enquiries, and if it is only watching sessions it will spend the year trying to fix something that is working.
What to measure instead
- Enquiries per session, not sessions. If sessions fall and enquiries hold, your traffic got better, not worse.
- Which pages still earn clicks. Informational pages are hit hardest; pages answering a decision hold up. That tells you what to write next.
- Whether you are being cited at all. Ranking reports were built for a world where ranking meant traffic. They quietly stopped answering the question.
Where it depends on your business
How much of this applies depends heavily on what your pages are for. A site built on how-to explainers is directly exposed — those are exactly the queries that now get answered in place. A site built on local, specific, decision-shaped pages is far less so, because “which of these applies to my situation in my city” is not a question a two-sentence summary can close.
Working out which category you are actually in requires looking at your own data, and the answer frequently differs page by page within the same site. That is the analysis, and it is not one you can do from a benchmark.
How we would walk you through it
We pull your last twelve to sixteen months and split pages by what they are trying to do rather than by how they perform. Then we look at impressions against clicks per group. The pattern is usually stark: one group has lost clicks while holding impressions, and another has barely moved.
What comes back is a view of which pages are still earning visits, which are now earning citations instead, and what that implies for what to publish next. Often the recommendation is to write less and more specifically.
That analysis is technical SEO work inside CAO Technologies, and the reporting side of it usually ends up with digital and operations.
Tell us whether your enquiries fell as much as your sessions did. If they did not, you have a measurement problem rather than a marketing one, and that is a much cheaper problem to have.
A question worth answering
Over the last year, did your enquiry count fall by the same proportion as your traffic? If not, which number have you been reporting on?
