Search visibility and search traffic used to move together. A page that climbed the results page got more visitors, and the reporting worked because the two numbers pointed the same way. That relationship has weakened, and a lot of SEO budgets are still being justified against the older version of it.
The bigger change is not just another ranking-factor update. It is that more results pages now answer the question before anyone leaves them.
The Click Gap
Pew Research Center tracked the browsing of 900 US adults across roughly 69,000 Google searches in March 2025. AI-generated summaries appeared on 18% of those searches, and 58% of participants saw at least one.
On results pages carrying an AI summary, users clicked a traditional search result in 8% of visits. On pages without one, they clicked nearly twice as often, in 15% of visits.
Links inside the summaries themselves were clicked in 1% of visits. Users were also more likely to end the browsing session entirely after a page with a summary, 26% of the time against 16%.
The study covered US users across a single month, so it is a snapshot rather than a trend line. It tracked what people did rather than what they said they did.
A business can maintain its position while the share of searches producing a click declines, without anything appearing wrong in a ranking report.
Google’s Own Position
The commercial response has been to sell a new discipline, badged AEO or GEO. Google’s guidance does not support the framing.
Its documentation states that from Search’s perspective, improving visibility in generative AI features is the same work as improving visibility in Search, and remains SEO. It also says there are no additional requirements to appear in AI Overviews or AI Mode, and no special techniques necessary for them.
That matters for budgeting. The problem is real, but Google does not require a separate set of AEO or GEO tactics to address it.
Where the Clicks Still Go
Query intent is doing the sorting. An AI summary can answer a question. It cannot take a booking, quote a job or dispatch a technician.
The distinction shows in the query itself. A question about how a roof gets replaced can be summarized. A search for someone to replace one cannot be. Agencies built around that kind of query organize their work accordingly, as www.hqdm.io/seo-services-miami illustrates.
Searches where someone is trying to reach a business are different. They are about contact rather than comprehension. Those results still lead to links, listings and map placements, where the goal is often to prompt a click, call, visit or other action.
The split is not good content against bad, but between queries a summary can close and queries it cannot.
What to Measure Instead
Reporting built on position and impressions can look healthy while the commercial result decays. A short set of changes covers most of it:
- Clicks and calls by query type, separating informational from transactional
- Position paired with click-through rate, rather than position on its own
- Share of target queries where a summary appears at all
- Conversions attributed to search, tracked independently of traffic volume
Much of that can be measured with existing analytics and search data, although the setup depends on the business and its conversion path. The harder part is reading the reports differently, and being willing to report a decline that the position column is hiding.
The Number That Moved
Rankings did not stop mattering. They stopped being a reliable proxy for the thing companies are buying, which is contact with a customer.
The practical consequence is a harder conversation about what a retainer covers. Work that once justified itself through traffic growth now has to justify itself through contact volume, and those two lines have come apart.
For businesses whose search demand is local and transactional, less has changed than the headlines suggest. Where demand is informational, clicks can be harder to come by, though that is a shift in how search traffic behaves rather than proof that the work has stopped paying off.
Article received via email
