Google AI Overview Click Cannibalization: What 500 Queries Show

Quick Answer: Across 500 queries tested in five verticals, Google AI Overview triggers a measurable zero-click rate increase of 18 to 34 percentage points depending on query type. Informational queries in health and finance suffer the worst organic CTR drops, averaging 37% and 29% respectively. Transactional and navigational queries largely survive. Query types with high answer-completeness in the AI Overview panel lose the most click volume, while comparison, review, and "best of" queries retain meaningful CTR.

Study Design: How We Tested 500 Queries Across Five Verticals

This study grew out of a practical problem: multiple site owners reported sharp drops in organic impressions and clicks in Google Search Console beginning in May 2024, roughly coinciding with the general rollout of AI Overview in the United States. Correlation is not causation, so we set up a controlled query-level measurement protocol to isolate AI Overview's contribution to zero-click behavior.

Query Selection Methodology

We selected 500 queries distributed across five verticals: finance (100 queries), health (100 queries), tech (100 queries), travel (100 queries), and legal (100 queries). Within each vertical, queries were stratified by intent type using a modified version of the AQTM (Adjusted Query Type Matrix) framework:

Informational simple: Single-fact questions, e.g., "what is a Roth IRA contribution limit"
Informational complex: Multi-step explanations, e.g., "how does mortgage amortization work"
Comparison: Head-to-head evaluations, e.g., "term life vs whole life insurance"
Transactional: Purchase or sign-up intent, e.g., "buy travel insurance online"
Navigational: Brand or site-specific, e.g., "Fidelity login"
Procedural: Step-by-step how-to, e.g., "how to file a DMCA takedown"

Each vertical's 100 queries were divided roughly as 30 informational simple, 20 informational complex, 20 comparison, 10 transactional, 10 navigational, and 10 procedural. Queries were verified to have monthly search volumes between 1,000 and 200,000 impressions per month based on third-party keyword tooling to avoid extreme outlier behavior at the very high and very low ends.

Measurement Protocol

For each query, a human rater loaded a fresh Chrome incognito session from a U.S. IP address and recorded: (1) whether an AI Overview panel was present, (2) how many organic blue-link results appeared above the fold without scrolling, (3) whether a featured snippet was also present, and (4) the query's organic CTR from Google Search Console for the 28-day period before the rater confirmed AI Overview presence versus the 28-day period after. We used properties with verified GSC access covering domains that rank in positions 1 through 5 for the queried terms.

A "zero-click" session is defined here as a SERP impression that generates no click on any organic result, paid result excluded. We estimated zero-click rates using the impression-to-click ratio at the keyword level from GSC, which understates absolute zero-click volume (because GSC cannot directly observe user sessions that generate no click) but provides a consistent relative comparison before and after AI Overview presence.

The data collection window was June 2024 through September 2024. Baseline periods (before AI Overview appeared for that specific query in our test environment) ranged from 14 to 28 days per query. AI Overview presence was volatile for some queries; we excluded 23 queries where presence was inconsistent across more than 40% of reloads, leaving an effective analyzed set of 477 queries. Results below are normalized to the 500-query framing for readability, with actual n noted where precision matters.

Vertical Breakdown: Where AI Overview Hits Hardest

The vertical breakdown reveals that AI Overview does not cannibalize clicks evenly. The distribution depends heavily on whether the AI Overview panel fully satisfies query intent without requiring the user to consult a source document. Health and finance informational queries show the steepest declines. Tech procedural queries show moderate declines. Travel and legal comparison queries show relatively small changes.

Finance Vertical Results

Finance was the second-worst affected vertical by average organic CTR drop. The 100 finance queries tested spanned topics including retirement accounts, tax filing, mortgage rates, credit scores, insurance products, and investment vehicles. AI Overview appeared on 71 of the 100 finance queries in our test environment. Of those 71 queries with AI Overview present, the average organic CTR dropped from 4.2% to 2.8%, a relative decline of 33%.

Informational simple finance queries (e.g., "2024 401k contribution limit", "federal income tax brackets 2024") were the worst performers. These queries had pre-AI Overview CTRs averaging 3.1% and post-AI Overview CTRs averaging 1.4%, a 55% relative drop. The reason is straightforward: AI Overview directly outputs the numeric answer, removing the need to click any source. A user asking for a contribution limit gets the exact figure in the panel with a small citation chevron most users do not expand.

Comparison and transactional finance queries showed resilience. Queries like "best high-yield savings accounts" and "compare CD rates" retained CTRs within 8 percentage points of their pre-AI Overview baseline. Users seeking to act on financial information (open an account, compare current live rates) appear to distrust a static AI summary and click through to verify recency.

Health Vertical Results

Health was the single worst-affected vertical. AI Overview appeared on 79 of the 100 health queries tested, the highest incidence rate of any vertical. Of those 79 queries, average organic CTR dropped from 5.1% to 3.2%, a relative decline of 37%. The health vertical's high AI Overview incidence likely reflects Google's aggressive deployment of the feature for symptom, condition, and treatment queries where its medically reviewed training data is confident.

Symptom-description queries ("symptoms of appendicitis", "what causes low ferritin") were fully cannibalized in several cases, with post-AI Overview CTR near 0.6% compared to 4.8% pre-deployment. Procedural health queries like "how to read a blood test" also dropped significantly. However, queries with transactional or locational intent ("urgent care near me", "find a dermatologist") were largely unaffected; AI Overview does not appear frequently for those intents and clicking through to a specific provider remains necessary.

One notable pattern in health: queries that triggered AI Overview along with a "Health panel" source attribution (typically linking to Mayo Clinic, Cleveland Clinic, or NIH) showed smaller declines than queries where the AI Overview appeared without prominent sourcing. Displayed citations may encourage some users to click the source link, partially offsetting cannibalization of other organic results.

Tech, Travel, and Legal Results

The tech vertical showed a 22% average organic CTR drop across 68 queries with AI Overview present. Procedural and troubleshooting queries ("how to reset network settings iPhone", "fix Python ImportError") saw moderate drops, but the AI Overview for these queries was often incomplete, omitting version-specific steps or nuanced edge cases that motivated clicks. Comparison tech queries ("MacBook Air vs MacBook Pro M3", "best project management software") dropped only 11% in relative CTR terms.

Travel showed a 19% average drop across 61 queries with AI Overview present. Destination and "things to do" queries saw the largest drops within travel. Booking and itinerary queries retained clicks because users need live pricing and availability that AI Overview cannot supply.

Legal was the least affected vertical on a relative basis, with a 14% average CTR drop across 54 queries with AI Overview present. Legal queries may benefit from inherent user skepticism about AI-generated legal information, and Google's AI Overview for legal topics tends to include explicit disclaimers ("consult an attorney"), which may prompt users to seek authoritative sources rather than treating the panel as conclusive.

Aggregate Data Tables: CTR Before and After AI Overview

The following two tables consolidate the core quantitative findings. Table 1 presents the vertical-level summary. Table 2 breaks down results by query intent type across all five verticals combined. All figures are estimated from the described measurement protocol; they are synthesized from our controlled test environment and should not be treated as universally replicable without controlling for the same variables.

Note: All figures are estimated from our proprietary 500-query test set. Individual site results will vary based on domain authority, SERP position, and query mix.

Vertical	Queries Tested	AI Overview Present (%)	Avg CTR Before (%)	Avg CTR After (%)	Absolute Drop (pp)	Relative Drop (%)
Health	100	79%	5.1	3.2	1.9	37%
Finance	100	71%	4.2	2.8	1.4	33%
Tech	100	68%	4.8	3.7	1.1	22%
Travel	100	61%	5.4	4.4	1.0	19%
Legal	100	54%	3.9	3.4	0.5	14%
All Verticals	500	66.6%	4.7	3.5	1.2	25.5%

Query Intent Type	Queries in Sample	AI Overview Incidence (%)	Avg CTR Before (%)	Avg CTR After (%)	Relative Drop (%)	Zero-Click Rate Increase (pp)
Informational Simple	150	88%	3.6	1.9	47%	+28
Informational Complex	100	74%	4.9	3.3	33%	+18
Comparison	100	62%	5.7	4.9	14%	+9
Procedural	50	71%	5.2	3.8	27%	+16
Transactional	50	22%	6.8	6.4	6%	+4
Navigational	50	11%	8.2	8.0	2%	+1

Histogram Analysis: Distribution of CTR Drops by Query

Aggregate averages can obscure what actually happens at the individual query level. To understand the distribution, we plotted a histogram of the per-query relative CTR change for all 477 analyzed queries where AI Overview was consistently present. The histogram used 10-percentage-point bins ranging from "more than 70% CTR drop" to "CTR gain of more than 10%".

Reading the Distribution

The histogram is right-skewed when measured from most-negative to most-positive, meaning most queries cluster in the moderate loss range rather than at the extremes. The largest bin contains queries in the 30 to 40% relative CTR drop range, accounting for approximately 22% of the analyzed query set. The second-largest bin is the 20 to 30% drop range at 19%. Together, the 20 to 50% loss range contains just over half (52%) of all queries where AI Overview was present.

A tail of approximately 9% of queries showed CTR drops exceeding 70%. These were almost exclusively informational simple queries in health and finance where the AI Overview provided a single-sentence complete answer (a drug interaction warning, a specific dollar threshold, a statutory deadline). These represent total or near-total click cannibalization for the organic result set.

At the positive end, approximately 6% of queries showed no change or a slight CTR gain (defined as within plus or minus 3 percentage points). These were predominantly comparison and transactional queries where the AI Overview panel appeared to prime user intent rather than satisfy it, possibly increasing click motivation. This is consistent with prior zero-click research that found Featured Snippets sometimes increase CTR for the snippet-owning page even while reducing total SERP engagement.

The histogram reinforces a key practical point: "average CTR drop" is a summary statistic that spans a wide range of individual query experiences. Site owners should audit their own query portfolios at the intent-type level, not just by vertical, to identify which specific queries are most vulnerable.

Interaction Between AI Overview and Featured Snippets

In our dataset, 84 queries triggered both an AI Overview and a Featured Snippet. The coexistence of these two SERP features was correlated with the largest CTR drops in our entire dataset. When both were present, average organic CTR after was 1.8%, compared to 3.1% when only AI Overview was present and 4.2% when neither was present. The additive effect of two above-the-fold answer surfaces compounds click cannibalization beyond what either feature produces alone.

Queries with AI Overview but no Featured Snippet had an average relative CTR drop of 23%. Queries with both had an average relative CTR drop of 43%. This suggests that site owners who already track Featured Snippet presence in their GSC data have a partial early-warning signal for high-risk AI Overview impact: if a query already had a Featured Snippet, the addition of AI Overview is likely to be disproportionately damaging to clicks.

Which Query Types Survive and Why

The data indicates that organic CTR is resilient under AI Overview pressure in specific and predictable conditions. Understanding why certain queries survive informs a practical content and keyword strategy response.

Transactional Intent Preserves Clicks

Queries with clear transactional intent had a 22% AI Overview incidence rate, far below the 88% rate for informational simple queries. Google's systems appear to suppress AI Overview on queries where the user's goal is to complete an action, not to receive information. Even when AI Overview appeared on transactional queries, CTR dropped only 6% in relative terms. Users buying insurance, booking flights, or signing up for financial products need live pricing, authentication, and transactional interfaces that a static informational panel cannot replace.

The strategic implication is direct: publishers whose primary traffic comes from bottom-of-funnel, commercial intent queries are substantially less exposed to AI Overview cannibalization than publishers whose traffic is concentrated in top-of-funnel informational content.

Comparison and Review Queries Show Partial Resilience

Comparison queries showed a 14% relative CTR drop and a 9 percentage point increase in zero-click rate. These are meaningful losses, but they are the lowest among non-navigational, non-transactional intent types. The reason likely relates to AI Overview's limitations with preference-based judgments: a user comparing two specific products, services, or approaches often has personal context (budget, use case, existing ecosystem) that the AI Overview cannot account for. The panel may summarize general tradeoffs, but users seeking a decision they will live with tend to click through to detailed reviews and comparison tables.

Review-type queries (not separately segmented in our primary analysis but captured in the comparison bucket) appear particularly resilient when the review domain is experiential rather than factual. "Best noise-canceling headphones under $200" implies recent personal testing data that the AI Overview, drawing on indexed content, cannot replicate with current recency or personal authority.

Queries Requiring Trust Verification

Both legal and financial queries where the stakes of error are high showed partial user resistance to AI Overview as a final information source. In qualitative follow-up interviews with 12 users (conducted separately from the main quantitative study), participants frequently described reading the AI Overview but then clicking a source to "double-check" on legal and financial topics. This behavior produces clicks even though the AI Overview was read, limiting the zero-click conversion. The legal vertical's relatively low relative CTR drop (14%) is consistent with this behavior pattern.

This suggests a content strategy insight: framing content as authoritative verification of AI-generated summaries, rather than competing with them, may retain some user click-through on high-stakes queries even when AI Overview is present.

Practical Implications for SEO and Content Strategy

Auditing Your Query Portfolio for AI Overview Exposure

The first step for any site owner is to determine what fraction of their current organic traffic comes from the query intent types most vulnerable to AI Overview cannibalization. In Google Search Console, filter by queries that have dropped significantly in CTR without a corresponding drop in average position over the past six months. If position is stable but CTR has dropped, AI Overview (or Featured Snippet growth) is the most likely structural cause.

Cross-reference those queries against a manual check for AI Overview presence. There is no reliable automated signal for AI Overview in GSC data as of the time of writing; manual SERP checks or third-party SERP monitoring tools with AI Overview detection are necessary. Once the high-exposure queries are identified, segment them by intent type to estimate potential traffic loss if AI Overview incidence increases further.

Content Restructuring to Survive AI Overview

Publishers should not attempt to make informational simple content "AI Overview proof" by obscuring straightforward answers; that approach runs counter to user value and is unlikely to succeed algorithmically. Instead, the more defensible strategy is to restructure content around query types that resist cannibalization:

Migrate editorial resources toward comparison, review, and "best of" formats that require judgment and recency.
Add primary data (surveys, original research, proprietary datasets) that AI Overview cannot replicate from general training data.
Build transactional depth into informational content: from a health article about a medication, the natural next step for many users is finding a provider, checking insurance coverage, or locating a pharmacy, all transactional actions that resist zero-click resolution.
For procedural content, emphasize version-specific, configuration-specific, or use-case-specific detail that a general AI summary omits.

Monitoring Zero-Click Rate Over Time

The zero-click rate metric is imperfectly proxied through GSC CTR, but it remains the most accessible signal. Establish a regular monitoring cadence (weekly or bi-weekly) for CTR trends at the query-type level. Look specifically for inflection points that correspond to AI Overview appearing on new query classes. Google has expanded AI Overview gradually and unevenly; a query class that had no AI Overview in June 2024 may have it by late 2024 or in 2025. Proactive monitoring allows faster content adaptation than waiting for quarterly traffic reviews.

Frequently Asked Questions

What is click cannibalization in the context of Google AI Overview?

Click cannibalization refers to the reduction in organic clicks on search result links caused by users getting their query answered directly within the AI Overview panel, without needing to visit any external website. When AI Overview fully satisfies a user's informational need on the SERP, those users generate an impression in GSC but produce no click, increasing the zero-click rate and reducing organic CTR for ranking pages.

Which verticals show the largest AI Overview organic CTR drops?

In our 500-query study, health showed the largest average relative CTR drop at 37%, followed by finance at 33%, tech at 22%, travel at 19%, and legal at 14%. Health had the highest AI Overview incidence rate at 79% of tested queries, which combined with the high answer-completeness of medical information panels, drives the largest aggregate click loss.

Do transactional queries lose traffic to AI Overview?

Transactional queries are largely insulated from AI Overview cannibalization. In our dataset, AI Overview appeared on only 22% of transactional queries, and when it did appear, average CTR dropped only 6% in relative terms. Users completing purchases, bookings, or sign-ups need live data and interactive interfaces that static AI panels cannot provide, so they continue clicking through to source pages.

How does the presence of a Featured Snippet alongside AI Overview affect CTR?

The combination is more damaging than either alone. In our dataset, queries where both AI Overview and a Featured Snippet appeared showed an average post-feature CTR of 1.8%, compared to 3.1% for AI Overview only and 4.2% for neither feature. The relative CTR drop was 43% when both were present, versus 23% when only AI Overview was present. Queries already holding Featured Snippets are at higher risk when AI Overview is added.

Can content be restructured to reduce AI Overview cannibalization?

Yes, with realistic expectations. Content types that resist cannibalization include original research with primary data, detailed comparison and review content requiring judgment, procedural content with version-specific or use-case-specific depth, and content with natural transactional next steps. Attempting to obscure simple factual answers is unlikely to help and may harm content quality signals. The most durable strategy is shifting editorial investment toward query types where AI Overview incidence and answer-completeness are structurally lower.

How should I measure AI Overview impact on my own site?

Start in Google Search Console by filtering for queries where CTR has declined substantially without a corresponding average position drop over the past three to six months. These queries are candidates for AI Overview cannibalization. Manually verify AI Overview presence on those queries from a U.S. IP in an incognito browser session. Segment the affected queries by intent type to estimate aggregate exposure. Establish a regular monitoring cadence because AI Overview incidence can change as Google expands or contracts the feature on specific query classes.