Two of the largest published studies on Google’s AI Overviews both measure the same headline question — what AI Overviews do to clicks — and both land in the same rough neighborhood. Ahrefs found a 58% drop in position-1 click-through. Seer Interactive found a 61% drop in organic CTR on AI-Overview queries. Close enough to sound like agreement. They are not measuring the same thing the same way, and the differences between the methods are more useful to you than either number.

The studies, precisely

Ahrefs (February 2026)

Ahrefs re-ran a study it first published in April 2025, using data from December 2025. The setup (Ahrefs, February 2026: https://ahrefs.com/blog/ai-overviews-reduce-clicks-update/):

  • 300,000 keywords from its Keywords Explorer database: 150,000 with an AI Overview present, and 150,000 informational-intent keywords with no AI Overview.
  • Aggregated Google Search Console data across its user base: average desktop CTR per month (sum of clicks / sum of impressions).
  • Comparison of December 2023 (pre-rollout) against December 2025.

Findings: the presence of an AI Overview correlates with a 58% lower average CTR at position 1, 50.8% lower at position 2, and 46.4% lower at position 3. Ahrefs derived the position-1 figure by forecasting what the CTR would have been without the AI Overview (using the non-AI-Overview set as the baseline) and measuring the gap between forecast and actual.

Three method notes worth keeping:

  1. This is observational, not experimental. Ahrefs’ own wording is that AI Overview presence “correlates” with the lower CTR. No control experiment was run.
  2. “AI Overview keywords” are defined by a snapshot: keywords that triggered an AI Overview in December 2025.
  3. It is desktop-only and drawn from Ahrefs’ aggregated client base — a global, cross-vertical mix of keywords, weighted toward whatever that user base queries.

Seer Interactive (published November 2025)

Seer, a marketing agency, published its analysis of AI Overviews on client account data (Seer Interactive, via PPC Land, November 2025: https://ppc.land/google-ai-overviews-reduce-organic-ctr-61-paid-traffic-68/):

  • 3,119 informational queries across 42 organizations, covering 25.1 million organic impressions.
  • Trend measured since mid-2024, the period when AI Overviews began rolling out.

Findings: organic CTR on AI-Overview queries is down 61% since mid-2024; paid CTR on the same queries is down 68%. Brands that got cited in the AI Overviews earned 35% more organic clicks than those that were not cited — a pattern the researchers explicitly describe as observational: they state they cannot prove that citation causes the higher click-through.

The key method notes here:

  1. The sample is a portfolio of real marketing accounts (42 organizations’ queries), not a broad keyword-database sample. It is weighted toward queries that marketing teams actually invest in.
  2. It measures a moving trend since mid-2024, not two point-in-time snapshots.
  3. It separates organic and paid — Ahrefs measures organic only.
  4. It is likewise observational before/after, not an experiment.

Why the numbers differ

58% and 61% look almost interchangeable, but they answer subtly different questions, and the methods explain the gap:

  • Sample. Ahrefs: 300,000 keywords aggregated from a global tool user base. Seer: 3,119 queries from 42 client organizations. One is a broad cross-section; the other is a portfolio skewed toward commercially managed queries. Different samples, different baselines.
  • Window. Ahrefs compares December 2023 with December 2025 — two fixed months, two years apart. Seer tracks a trend since mid-2024. A before/after snapshot and a rolling trend have different denominators: the snapshot captures the total change between two points, the trend captures the average decline across the period, including its early months.
  • Definition of “AIO shown.” Ahrefs fixes the set by a December 2025 snapshot — a keyword is “AIO” if it triggered an Overview that month. Seer looks at informational queries featuring AI Overviews across the measurement window, where coverage can shift as Google expands and contracts which queries get Overviews.
  • CTR construction. Ahrefs is desktop-only, clicks divided by impressions from aggregated Search Console data. Seer’s paid and organic figures come from client account data with its own device and market mix.
  • Causal framing. Neither study ran an experiment. Both are correlations with a constructed counterfactual (Ahrefs) or a before/after comparison (Seer). Neither number can be read as “if Google had not launched AI Overviews, your CTR would be exactly X% higher.”

That is the whole story of the disagreement: different samples, different windows, different definitions of the treatment. The two studies are estimates of related-but-not-identical quantities.

What is robust across both

Strip away the method differences and a few claims survive in both studies:

  1. Direction. AI Overview presence is associated with meaningfully lower organic click-through on the affected queries, in both studies, by a wide margin. Nobody with access to this data is seeing the opposite.
  2. Magnitude at the top. Both land in the same order of magnitude — roughly a 50–60% reduction in the leading positions — which matters more than the exact figure for planning.
  3. The cited-brand premium. Seer’s finding that cited brands earned 35% more organic clicks is the single most actionable number in either study — and also the weakest on its own: one source, observational, with the authors themselves declining to claim causation. Cited brands may earn the extra clicks because the Overview links them, or because brands citable enough to appear were stronger to begin with. Treat it as a strong signal that citation and CTR move together, not as proof that citation causes the difference.

What is not robust: the exact percentages (the next study will produce a third number), whether the decline is still continuing or has stabilized, and how much of the drop is AI Overviews versus the longer “law of diminishing clickthrough” that has hit search for years.

What this means for a B2B SaaS site

The honest framing for a B2B SaaS site is: being un-cited is an active loss, and the CTR data is the evidence.

Two mechanisms are at work. First, where an AI answer is served on your query, clicks to the top results fall sharply — so a query where you used to capture clicks at position 1 is now a smaller prize even with no change in your ranking. Second, if AI answers cite competing brands instead of yours, the cited-brand pattern in the Seer data suggests citation and click share travel together. For a B2B SaaS product, the exposed queries are category queries (“best [category] for [use case]”), pain-point queries, and comparison queries — exactly the informational-commercial hybrids where AI Overviews and chat answers are forming. If the answer to one of those queries is served without you in it, the buyer may never reach your comparison page at all.

Two caveats keep this honest. The studies are informational-query-heavy; purely transactional queries likely behave differently. And these are Google Search results — the chat assistants your buyers also use (ChatGPT, Perplexity, Gemini, Claude) are separate surfaces where the same logic applies — citation presence determines whether you are in the answer — but with no published CTR studies to lean on. Your own measurements are the only data that will cover that gap.

How to check your own site

You do not need to wait for the next published study:

  1. Open Search Console’s generative AI performance report. It shows how many impressions your pages earn in Google’s AI features (AI Overviews and AI Mode), by page, country, device and date. It is impressions only — no clicks, no citations — but it tells you which of your pages the AI answers are actually featuring. (It counts appearances in Google’s own AI features; it says nothing about ChatGPT or any other assistant.)
  2. Watch CTR on your money queries in the normal performance report. Identify the queries where you rank in the top three, check where AI Overviews trigger on them, and record your CTR month over month. You will have your own version of the Ahrefs before/after.
  3. Run your money queries in the AI surfaces yourself. ChatGPT, Perplexity, Gemini, Claude, on the same 20–50 queries, on a monthly cadence. Record who is cited, who is mentioned without a link, and who is absent. Capture the answers verbatim.
  4. Measure your own citation premium. Among the queries where you do get cited, compare click and referral behavior against months and queries where you were not. Your own data is the only signal with a chance of being causal for you.

That last loop — run, capture, diff, act — is what we document in our methodology.

Try the free quick check

If you would rather see the third step run for you: we take your money queries, run them across ChatGPT, Perplexity, Gemini and Claude, and send back a snapshot scorecard — cited, mentioned, or absent, with what each answer actually said. No meeting required. Get the free quick check.