Why is my organic traffic falling in 2026 when my rankings have not changed?

Because AI-generated summaries now answer many searches on the results page, so people read the answer without clicking through. Pew Research Center found the click-through rate to traditional results roughly halved when an AI summary was present. Your ranking can hold while the click falls, because the AI Overview sits above the first organic result and settles the question before the reader reaches you. The effect is heaviest on informational queries and much lighter on commercial ones.

Do AI Overviews actually cause the lost clicks, or is it just correlation?

There is causal evidence. A randomized experiment from the Indian School of Business and Carnegie Mellon hid AI Overviews for a random group of users and measured the difference. With the summaries hidden, clicks to outside sites rose by roughly 60% and the zero-click rate fell sharply. Random assignment is what lets the study show cause rather than coincidence.

No, but its job has narrowed. Ranking still earns clicks on the commercial and transactional queries that Overviews summarise less often, and those pages are holding up. What has thinned is top-of-funnel informational traffic, where the answer now appears in place. The work shifts from ranking a page to being a source the AI answer names, and from counting clicks alone to measuring whether you appear in the answer.

How do I prove it is AI Overviews specifically and not something else?

Put impressions next to clicks in Google Search Console for your worst-hit pages over the last year. Steady or rising impressions with falling clicks is the signature of the click being absorbed before it reaches you. If impressions fell too, the cause is more likely an ordinary ranking or relevance issue, which is the first thing to rule out. Checking whether the loss concentrates on informational rather than commercial pages will confirm the pattern.

Google said the Pew study was flawed. How do I hold the line if someone raises that?

Google pushed back, mainly on sample size. The way to hold the line is not to lean on Pew alone. The randomized experiment from the Indian School of Business and Carnegie Mellon reached the same conclusion by a method that removes the sampling objection, and UK regulator Ofcom has separately reported that AI summaries now appear on a substantial share of UK searches. When a behavioural panel, a controlled trial and a regulator point the same way, the consistent direction is the point, not any single percentage.

Should I just publish more to win the traffic back?

Publishing more undifferentiated explainer content tends to deepen the problem, because it feeds the summaries more material to answer questions in place while competing for a click that is leaving. The better response is to work to be cited inside AI answers, protect the commercial pages that still convert, and measure presence in the answer so that being named counts even when no click follows.

SEO Archives - Ridley Digital

We use cookies to run this site, understand how it is used and improve it. Essential cookies are always on. You can accept everything, reject the rest or choose what runs. Read our cookie policy.

Essential

Needed for the site to work, including security and your cookie choice. These cannot be switched off.

Analytics

Helps us understand which pages people read so we can write better ones. Anonymous and aggregated.

Marketing

Used to measure campaigns and show relevant content across other platforms.

Search

From Rankings to Presence: How AI Search Is Changing SEO Strategy

SEO for Financial Services: The 2026 Playbook for Regulated Firms

Essential

Analytics

Marketing