The evidence that Google AI Overviews displace organic clicks is no longer a debate. Seer Interactive measured a 65% drop in organic click-through on queries that trigger an AI Overview (1.76% to 0.61% over its measured window) and a 68% drop in paid click-through; Ahrefs found a position-one page 58% less likely to be clicked when an overview is present, up from a 34.5% reduction the previous April; a randomized field experiment from Carnegie Mellon and the Indian School of Business measured a 39.8% cut in outbound clicks on affected queries, with BrightEdge at 30% and Pew measuring an 8% versus 15% click rate. These are dated, independent sources — Seer and Ahrefs from late 2025 to early 2026, the field experiment posted to SSRN, and all of it consistent. And the December 2025 to September 2026 direction, with Google expanding overviews into full-page answers, is toward more displacement, not less.
So the useful question is not whether the clicks are gone. It is how to defend a strategy whose underlying currency just changed. A defensible plan has five parts, and every part starts from the same reality: optimizing a page to rank is no longer the same as optimizing a page to be cited, and you should forecast and budget for the difference.
Part 1: Re-forecast traffic per query segment, not per page
Before you cut anything, model your own exposure. Take last quarter's keyword set and split it four ways by intent, because the hit is wildly uneven. Informational queries are the critical risk — an AI Overview displaces 70 to 80% of their clicks. Commercial queries are moderate, at 35 to 45%. Transactional queries hold up best, at 20 to 30%. And branded queries are the one segment that gains, with click-through up about 18.7% when an overview appears. Build the re-forecast by segment, not by total reach, and you will know roughly where the next quarter is going before you spend against it.
Part 2: Move spend from the ten-blue-link page to the citable page
Where the budget used to go to the content that ranks, shift a meaningful share to the content that gets cited. The research is explicit: a brand cited inside an overview sees 35% more organic clicks and 91% more paid clicks than a brand left out. That is the return you are now buying. Invest in the pages an engine can quote — a data-backed explainer, a clearly scoped comparison, a genuine table or definition block — and in the citation signals that make them eligible. This is a re-allocation, not an add-on; it is the organic equivalent of shifting a line item from a channel that is decelerating to one that is still paying.
Part 3: Reserve paid budget where organic can no longer deliver
The paid side is not immune — Seer measured paid click-through down 68% on AI-overview queries — but paid has a different job now. On queries where the overview owns the answer and organic cannot win the citation, a paid slot is increasingly the only way to get in front of the user at the moment of intent. Reserve budget there deliberately, and buy it on the queries where your organic gap is widest, rather than spreading it across keywords that organic still handles. The defensible pattern is to let organic carry the queries it can still win, and pay for the ones the answer surface is taking away.
Part 4: Stop treating the ranking report as the report
A strategy built on a rank checker is measuring a surface that is losing relevance. Rank your pages still matters, but it is no longer the signal that predicts traffic. Re-base the scorecard on what the answer actually says: which of your money queries produce an answer that names you, with a source, and which produce a silence or a competitor. That citation scorecard is what you report to leadership, because it is the number that moved when the feature launched and the number that still predicts the 35% and 91% uplifts. If you report clicks only, you will spend the next two quarters explaining a decline you cannot defend.
Part 5: Build the buffer that lets you absorb the hit
The final part of a defense plan is the part that has nothing to do with ranking. The Daily Mail case is the cleanest example in the public record: publishers with strong direct and branded traffic absorbed an 80 to 90% click-through crash on AI-overview keywords with a "very, very low single-digit" total-traffic impact, because they were not renting every visit from the search box. Build that buffer deliberately — brand the product, grow email and direct habits, and protect branded query health. It is the only move that makes a citable page and a good ad buy into a survivable mix rather than a losing one.
The honest limit
Every figure above is directional. Two of the studies are vendor analyses, one is a randomized field experiment that is still a working paper, and none of them can tell you your exact loss — only your own segment-split re-forecast can. Build the forecast with your own data, re-baseline the KPI to citation, and treat every number here as the reason to start measuring rather than as a number to copy. The thesis that survives is the practical one: the currency changed, so the strategy and the budget have to change with it.
FAQ
Do AI Overviews actually cut clicks, or is this vendor hand-waving?
Multiple independent sources converge. Seer Interactive (65% organic, 68% paid), Ahrefs (58% lower position-one CTR), a randomized field experiment from Carnegie Mellon and the Indian School of Business (39.8% outbound-click drop), BrightEdge (30%), and Pew (8% versus 15%) all point the same direction. A vendor study in isolation is a claim; a stack of independent ones is a trend.
Should I just stop doing SEO because clicks are falling?
No — SEO is not dead, but its payload changed. The job is now the citable page, and brands cited inside an overview gain 35% more organic clicks and 91% more paid clicks than brands left out. Defend by re-allocating to citation, measure by citation, and keep ranking as a secondary check.
How do I model my own click loss?
Split your keywords into informational, commercial, transactional, and branded, and apply the segment deltas (70-80% / 35-45% / 20-30% / +18.7% respectively) to your own volume, then sanity-check against your GA4 and Search Console numbers. Your own data, not someone else's headline, is the source of truth.
Related reading
- How to Stay Visible in Google AI Overviews — the seven-step tactical playbook
- Which Queries Lose the Most to AI Overviews — the segment-by-segment breakdown
- How to Optimize Comparison Pages for AI Search Citations — building pages the engines cite
Author: Theo Langford, Head of Search Visibility at Auspia. Theo helps teams re-forecast and re-allocate search budgets as AI answers change how content is discovered.
