Ranking first in Google search results used to mean a predictable share of clicks. It still means a rank. It no longer means the same clicks. Ahrefs' own February 2026 update, built on 300,000 keywords and aggregated Google Search Console data, found that position-1 click-through on queries where an AI Overview appears fell from 7.3% in December 2023 to 1.6% in December 2025 — 59% below the 3.9% click-through that equivalent, non-AI-Overview queries still saw in the same period.[1] Every query's click-through has eroded over two years, AI Overview or not — but the erosion is roughly twice as steep when an AI-generated summary sits above the results. A separate, earlier finding still widely cited in 2026 coverage — Pew Research's mid-2025 tracking of real user browsing — found people click through at roughly half the rate (8% vs. 15%) when an AI Overview is present, and only 1% click the specific sources the summary cites.[2] The diagnostic finding is narrow and precise: the rank is the same currency it always was. What it buys has changed.
Search click-through has been eroding for years as Google adds more on-page answers, but Ahrefs' February 2026 update isolates the specific effect of AI Overviews, the AI-generated summaries Google now places above traditional results. Tracking 300,000 keywords — 150,000 with AI Overviews present, 150,000 comparable informational queries without — against two years of aggregated Search Console data, position-1 click-through on AI-Overview queries fell from 7.3% (Dec 2023) to 1.6% (Dec 2025).[1]
The comparison that matters most isn't the trend over time — every query type lost clicks over those two years as Google's result pages got denser. It's the same-period gap: by December 2025, a position-1 result on a non-AI-Overview query still captured 3.9% click-through. The equivalent position-1 result on an AI-Overview query captured 1.6% — 59% lower, in the same month, for the same rank.[1] Ranking first still means ranking first. It buys roughly a third of the traffic it used to.
This finding sits alongside an earlier, separately-conducted study still widely cited in 2026 coverage: Pew Research Center tracked real browsing behavior from 900 U.S. adults across 68,879 actual Google searches in March 2025, and found people clicked through at 8% when an AI Overview appeared versus 15% when it didn't — and clicked one of the summary's own cited sources only 1% of the time.[2] The two studies use different methods and different time windows and shouldn't be blended into one number, but they point the same direction independently: an AI Overview doesn't just add a box to the page, it measurably redirects where attention goes.
The honest limit of this case: it documents a real, reproducible click-through gap. It does not, by itself, prove total traffic collapse for every publisher or site — aggregate effects vary by query type, industry, and how much of an answer the AI Overview actually gives versus how much still requires a click to resolve. What the data supports is narrower and still consequential: the unit of value search marketers have optimized for — the position-1 rank — now converts to meaningfully less traffic than it did two years ago, on the specific queries where an AI Overview appears.
The same rank, two years apart, converting to a fifth of the clicks — and 59% below what an equivalent non-AI-Overview position-1 result still earned in the same month.[1]
How two years of AI Overviews changed what a top search ranking is worth.
Position-1 click-through on what would become AI-Overview queries stands at 7.3%, roughly in line with non-AI-Overview queries at 7.6% — a normal, small gap.[1]
Baseline900 U.S. adults, 68,879 actual Google searches: 8% click-through when an AI Overview appears versus 15% when it doesn't — real behavior, not modeled estimates.[2]
Independent CheckPosition-1 CTR on AI-Overview queries falls to 1.6%, versus 3.9% for non-AI-Overview queries at the same rank, same period — a 59% shortfall that didn't exist two years earlier.[1]
The GapThe 300,000-keyword study, methodology fully disclosed, becomes the primary reference point for the AI-Overview click-through effect.[1]
The DisclosureAs of this writing, no newer large-scale CTR study has updated these figures — the next data point is open.
UnresolvedPosition-1 CTR on AI Overview queries fell to 1.6% — 59% lower than equivalent non-AI-Overview queries. — Ahrefs, AI Overviews click-through update, February 2026
| Dimension | Evidence |
|---|---|
| Customer (D1) Origin · 86 | The lever is a measured, reproducible shift in where searcher attention goes once an AI-generated summary sits above the results — confirmed independently by two differently-sourced studies.[1][2] D1 is the origin because everything else in this case follows from what searchers actually do, not what platforms claim.Attention Redirected |
| Revenue (D2) L1 · 80 | Publisher and marketer revenue models built around position-1 traffic now face a materially smaller conversion from rank to visit.[1] D2 amplifies from D1 because the behavior shift only matters commercially because real revenue was built on the old conversion rate.The Traffic Publishers Priced In |
| Operational (D6) L1 · 76 | SEO and content strategy built around achieving position 1 now targets a materially smaller reward for the same rank, without the strategy itself having changed.[1] D6 amplifies alongside D2 as the operational consequence of the same shift.Optimizing for a Smaller Prize |
| Quality (D5) L2 · 58 | Whether an AI Overview genuinely resolves a searcher's intent (a real quality improvement) or simply discourages a click that would have found a better answer is an open question this case's CTR data can't settle alone.[2] D5 sits here as the unresolved quality dimension underneath the behavior shift. |
| Regulatory (D4) 48 | The regulatory backdrop — an antitrust remedy and an open EU probe, both documented in this cluster's other cases — touches how AI-generated answers are distributed and disclosed, though neither directly regulates click-through economics yet. D4 sits at moderate weight as adjacent, not central, pressure. |
| Employee (D3) 30 | Deliberately the thinnest dimension. This is a search-behavior and publisher-economics cascade; no comparable workforce-level finding exists in the research. |
The cascade originates in D1 — Customer — because the lever is a real, measured shift in searcher behavior: where attention goes once an AI-generated answer sits above the results.[1][2] From D1 it amplifies into D2 (publisher and marketer revenue, which depends on the click-through this data shows eroding) and D6 (the operational reality that ranking strategy, built around position-1 as the goal, now targets a smaller prize).[1] It then reaches D5 (the quality question — is an AI Overview answer actually replacing the need for a click, or just discouraging one) and D4 (the regulatory backdrop this cluster's other cases document — an antitrust remedy and a live EU probe both touch how AI-generated answers are built and disclosed). D3 is deliberately thin — this is a search-behavior and publisher-economics cascade, not a workforce one. Cross-references: [UC-278] is the regulatory constraint on the same distribution mechanism, now under appeal; [UC-279] is the counterexample showing AI adoption doesn't guarantee a platform wins the resulting attention; [UC-280] scoreboards whether disclosure rules and the appeal reach this mechanism before the next data point lands.
-- UC-277: The Click Nobody Sees: 6D Diagnostic Cascade
-- AI Overviews cut position-1 CTR to 1.6pct, 59pct below non-AIO queries, Dec 2025 (cluster: UC-278/279/280)
FORAGE click_nobody_sees
WHERE ai_overview_ctr_measured = true
AND ctr_gap_vs_non_aio_confirmed = true
AND independent_corroborating_study_exists = true
ACROSS D1, D2, D6, D5, D4, D3
DEPTH 3
SURFACE click_nobody_sees
DIVE INTO rank_versus_reach
WHEN position_unchanged = true
AND clickthrough_collapsed = true
TRACE attention_redirection_cascade
EMIT ai_overview_ctr_signal
DRIFT click_nobody_sees
METHODOLOGY 90
PERFORMANCE 45
FETCH click_nobody_sees
THRESHOLD 1000
ON MONITOR CHIRP high 'Ahrefs Feb 2026: 300K keywords, GSC data. Position-1 CTR on AI-Overview queries fell 7.3pct (Dec 2023) to 1.6pct (Dec 2025), 59pct below the 3.9pct CTR non-AIO position-1 queries saw same period. Pew 2025: 900 US adults, 68,879 searches, 8pct CTR with AI summary vs 15pct without, 1pct clicked cited sources. Same direction, independent methods'
SURFACE analysis AS jsonRuntime: @stratiqx/cal-runtime · Spec: cal.semanticintent.dev · DOI: 10.5281/zenodo.18905193
Position 1 in December 2025 is still position 1. It converts to roughly a fifth of the click-through it did two years earlier on the same query type — the currency didn't change, its exchange rate did.[1]
Every query's CTR eroded over the two-year window as result pages got denser. What's distinct about AI-Overview queries is the size of the erosion relative to non-AI-Overview queries at the identical rank — 59% lower, same period.[1]
Pew found only 1% of users click through to a source the AI summary itself cites — meaning the summary isn't primarily routing traffic to sources, it's substituting for the need to visit them.[2]
The data is specific and well-sourced on click-through rate. It does not, on its own, quantify total revenue impact across every publisher business model — that effect varies by site and is the open question the rest of this cluster's regulatory and competitive cases sit alongside.
Two sources: Ahrefs' own disclosed-methodology click-through study, the primary dataset behind this case, and Pew Research's independent tracked-browsing study, an earlier but still widely-cited corroborating finding.
Two independent studies, two different methods, the same direction: an AI Overview doesn't add to the page. It redirects where the click goes.