Zero-Click Search

Zero-click search is a search in which the user gets what they need from the results page or an AI-generated answer and never visits a website. The term predates generative AI, since featured snippets and knowledge panels had the same effect, but AI Overviews and conversational engines have made it the central measurement problem in AI search.

The evidence supports a clear statement and a qualified one. Clearly: when an AI answer appears, fewer people click organic results. Less clearly: how large the loss is, whether it is still growing, and how much of the missing traffic returns later through other channels. The two best public datasets disagree on the first two points.

The Click-Through Evidence

Ahrefs (February 2026). Using aggregated Google Search Console data for 300,000 keywords, half with an AI Overview and half without, and comparing December 2023 with December 2025, Ahrefs found that the presence of an AI Overview is associated with a 58% lower click-through rate for the first organic result. Its April 2025 study had put the figure at 34.5%. On this reading the effect is large and deepening.

Seer Interactive (April 2026). Seer tracked 53 brands, 5.47 million queries and 2.43 billion organic impressions from January 2025 to February 2026. Organic click-through on queries showing an AI Overview fell from 3.19% in January 2025 to a floor of 1.31% in December 2025, then recovered to 2.36% in February 2026. On queries without an AI Overview it was 3.82% in February 2026. Paid click-through shows the same gap: 16.21% with an AI Overview against 21.85% without.

Why the Studies Conflict

Seer's February 2026 figures imply organic click-through roughly 38% lower on AI Overview queries, a smaller gap than Ahrefs' 58%, and Seer shows a rebound where Ahrefs shows acceleration. Neither is wrong on its own terms. Ahrefs isolates position one, the result with the most to lose, and compares two Decembers two years apart. Seer measures every position its clients hold, month by month, on a panel of 53 brands that is not a random sample of the web. Ahrefs' endpoint also coincides with Seer's December 2025 low point, so the Ahrefs snapshot may have caught the trough.

Both are observational. AI Overviews appear mostly on informational and question-style queries, which have different click behaviour from transactional ones regardless of any AI answer, so a simple with-and-without comparison mixes the effect of the feature with the effect of the query type. Ahrefs itself describes its finding as a correlation and cites other estimates ranging from 47.5% to 90%. The honest range for planning is wide.

One finding is consistent: citation softens the loss. Seer reports that a brand cited in the AI Overview earns 120% more organic clicks per impression than a brand that is not cited, though still fewer than on queries with no AI Overview at all. Google, for its part, says in its Search documentation that clicks from results pages with AI Overviews are "higher quality", meaning users are more likely to spend more time on the site.

The Hidden Branded-Search Effect

Zero-click does not mean zero effect. Similarweb data reported by Search Engine Journal in June 2026 found that people shown a brand in a ChatGPT recommendation were 2.5 times more likely to visit that brand's site within seven days, and that 55.9% of those visits arrived through branded search rather than a click from the AI answer. The user saw the name in one place and typed it into another.

The study covers US desktop activity in finance, travel and beauty only, and it shows correlation rather than causation; people who ask about a category may already be inclined to visit its brands. With those limits, it is the best public evidence that part of what looks like lost traffic is displaced traffic, credited by analytics to brand search or direct visits. The implications are developed under AI referral traffic.

What It Means for Measurement

Sessions and click-through rate, taken alone, now understate reach. A page can inform many answers and record few visits. Three adjustments follow.

First, treat presence in the answer as an outcome in its own right. Whether a brand is cited or mentioned has to be sampled by running prompts repeatedly, because citations vary from day to day; this is the role of AI visibility measurement and of tools such as LLM Optimizer. Second, watch branded search volume and direct traffic as lagging indicators of AI exposure, while accepting that they cannot be attributed cleanly. Third, segment click-through analysis by whether an AI Overview appeared and whether the site was cited, since the three cases behave differently. Google reports AI feature traffic inside ordinary Search Console totals, so this segmentation requires third-party data.

The strategic shift is from counting visits to measuring influence, which is harder to do and easier to overclaim. Any single headline figure for zero-click behaviour, including widely circulated ones with no traceable source, deserves scepticism.