AI Overviews and CTR: Ahrefs, Seer, and Pew Reconciled

Quick answer

The available studies generally indicate fewer clicks when an AI Overview appears, but they don't justify a universal coefficient. Ahrefs estimates a CTR roughly 58% lower in relative terms for the page in position 1 across a cohort of 300,000 keywords on desktop in December 2025. Pew observes 8% of visits clicking a classic result when an AI summary is present, versus 15% without, among 900 American adults in March 2025. Seer then observes an AIO CTR rebound from 1.3% to 2.4% between December 2025 and February 2026. These measurements answer three different questions.

Key takeaways

• A 58% relative drop is neither a 58 point drop nor a loss of 58% of all traffic.
• Pew measures behavior after a SERP visit; Ahrefs measures position 1 CTR; Seer measures a portfolio of brands and queries.
• Seer shows that being cited improves CTR within AIO SERPs, without recovering the level of queries with no AIO.
• Intent, device, period, and keyword selection strongly change the effect.
• A serious forecast applies scenarios only to the share of impressions genuinely exposed.

The model: exposure, position, citation, and click

Four events must stay separate:

1. Exposure: the query displays an AI Overview.
2. Organic presence: your URL ranks in the classic results.
3. Citation: your URL or domain is presented as a source in the answer.
4. Click: the user visits the site from one of those surfaces.

Standard organic CTR is clicks / impressions. But a Search Console impression can aggregate different interfaces, positions, and intents. An external tracker may also observe the AIO's presence at a different moment, in a different place, or with no personalization. The "AIO" status must therefore be dated and treated as a measured exposure, not as a permanent property of the keyword.

Making the main studies comparable

These studies don't cancel each other out. Ahrefs looks for the plausible effect on the first page. Pew observes whether a person clicks after the SERP. Seer examines how a commercial portfolio evolves and the relative advantage of being cited. Semrush tracks the absence of any click on keywords. Their denominators differ.

Understanding relative and absolute figures

If CTR goes from 4% to 2%, the drop is 2 percentage points and 50% in relative terms. For 100,000 impressions, clicks go from 4,000 to 2,000.

In the Ahrefs study, the counterfactual CTR is 3.734% and the observed CTR is 1.5677%. The absolute gap is:

3.734% - 1.5677% = 2.1663 points

The relative gap is:

(1.5677 - 3.734) / 3.734 = -58.02%

Saying "CTR falls by 58%" without giving both levels hides the volume. Saying "it falls by 58 points" would be false.

Eight step forecasting procedure

1. Export at least thirteen months of queries, pages, countries, and devices from Search Console.
2. Group queries by intent and by target page; keep brand terms separate.
3. Record the presence of an AIO several times a week with date, country, and device.
4. Calculate the exposed share of impressions, not just the share of keywords.
5. Measure historical exposed and unexposed CTR where the series allows it.
6. Create three relative decline assumptions: conservative, central, and severe.
7. Add a "cited" scenario distinct from the "not cited" scenario, without assuming that the citation causes the gap.
8. Compare the forecast with actuals every month and re-estimate the coefficients.

Scenario formula

Clicks = I × [E × (C × CTR_cited + (1-C) × CTR_not_cited) + (1-E) × CTR_no_AIO]

where I is the number of impressions, E the share exposed to an AIO, and C the exposed share where your brand is cited. All values are proportions between 0 and 1.

Worked example: budgeting without overpromising

A page gets 80,000 monthly impressions. Dated observations estimate that 55% are associated with an AIO. Within AIO SERPs, the site is cited on 30% of impressions. Three scenario CTRs are used: 3.6% with no AIO, 2.2% with an AIO and a citation, 1.0% with an AIO and no citation.

• AIO and cited: 80,000 × 55% × 30% × 2.2% = 290 clicks
• AIO not cited: 80,000 × 55% × 70% × 1% = 308 clicks
• No AIO: 80,000 × 45% × 3.6% = 1,296 clicks
• Total: 1,894 clicks

With no segmentation, applying 3.6% to all 80,000 impressions would have given 2,880 clicks. The scenario shortfall is 986 clicks. This isn't a causal estimate: queries with an AIO may be more informational and less clickable to begin with. The right use is budgetary, with a range and a monthly review.

What Seer's measured rebound reveals

Seer observes a floor of 1.3% organic CTR on AIO queries in December 2025, then 2.4% in February 2026, which is +1.1 points and roughly 85% in relative terms. In the same report, queries with no AIO go from 2.8% in January 2025 to 3.8% in February 2026. This suggests a dated coefficient becomes obsolete quickly.

The report also indicates, across 30,842 informational queries, that an AIO appears on 267 of the 280 comparison queries (95.4%) and 1,214 of the 1,413 queries phrased as questions (85.9%). That distribution comes from the portfolio Seer tracks; it doesn't constitute a general prevalence across the web. It does justify segmenting those formats in a B2B forecast.

The +120% gain for a cited brand, compared with a brand not cited on an AIO SERP, remains an association. A cited brand may also be better known, better ranked, or more relevant. Seer specifies that cited CTR is still 38% lower than queries with no AIO. The reasonable decision is to work on citation when the topic has commercial value, without announcing that it will restore the old traffic.

What the data proves and doesn't prove

The four studies show that the AIO context is associated with a different distribution of clicks. They also prove that the period and the query selection matter: the effect is neither fixed nor uniform.

They don't prove that adding an FAQ, markup, or a precise word count increases CTR. Google in fact states in its 2026 optimization guide that there is no ideal length, no special AI markup, and no obligation to break content into tiny fragments. Nor do the studies demonstrate that citation causes the click.

Common mistakes and stopping conditions

• Applying a 58% relative drop to every page, including navigational and transactional ones.
• Mixing organic CTR, clicks on an AIO source, and the zero-click rate.
• Measuring AIO presence once then applying it to the previous twelve months with no warning.
• Ignoring mobile when the study used is desktop only.
• Removing queries that lost impressions: that creates survivorship bias.
• Stopping a forecast if exposed impressions can't be estimated; present a sensitivity range instead of a central figure.

Reusable asset: the scenario calculator

The AIO CTR scenario calculator lets you enter impressions, exposure rate, citation rate, and three CTRs. Duplicate one row per intent and device. Add the columns coefficient_source, observation_date, and uncertainty to stop an external coefficient from becoming a silent constant.

How SEOryon fits in

Where the relevant connectors and tracking are available, SEOryon can bring together Search Console data and visibility observations to make this segmentation easier. The platform doesn't turn an association into causality and doesn't see competing impressions absent from your property. The analyst keeps the choice of cohorts, assumptions, and the economic threshold.

Measurable exercise

Select ten informational pages and produce a three scenario click forecast for the next eight weeks. Deliverable: a table by page with impressions, AIO exposure, citation, baseline CTR, coefficients, and the limitation. Success: the mean absolute error between the central forecast and actuals is calculated, then the assumptions are revised without retroactively changing the observations.

FAQ

Do AI Overviews always halve CTR?

No. Ahrefs estimates roughly 58% lower in a specific context, but Seer observes a rebound and Semrush a different zero-click result. Measure your own portfolio.

Does being cited in the AIO recover the lost clicks?

In the Seer cohort, cited brands get more clicks than uncited ones on AIO SERPs, but stay below queries with no AIO. That isn't a causal guarantee.

Should you drop informational content?

No. It can build understanding, authority, and citation. Assess its contribution to brand searches, assisted journeys, and conversions, then consolidate the pieces that bring neither demand nor original information.

Which period should you use as a baseline?

Use at least thirteen months to spot seasonality, but give more weight to recent weeks when the interface changes. Keep the method breaks visible.

Read next

Consult the Observatory register to qualify the sources, then how Google AI Overviews and AI Mode work. The next analysis separates AI citations and Google rankings. To put this into practice, use measuring AI visibility.

References

• Ahrefs: AI Overviews CTR update, February 2026- Pew Research Center: clicks and AI summaries- Seer Interactive: AIO Impact on Google CTR, 2026- Semrush / Datos: AI Overviews study- Google Search Central: guide to generative features

Method and update note

Reviewed 16 July 2026, translated and edited 22 July 2026. Rates are presented with their denominator and in both points and relative terms. The scenarios are not guaranteed forecasts; they must be recalibrated as soon as an interface, a tracking model, or the query mix changes.