Google's Unverifiable AI Click Numbers and What They Cost You

In May 2024, Google’s Liz Reid told the world that AI Overviews were already generating more searches and more clicks than the search experience it replaced. By early 2025, the company had attached a specific number to that claim: billions of clicks per week flowing from AI-generated answer panels to the underlying web. That number traveled fast — through marketing conference decks, agency pitch meetings, and the inboxes of business owners from Shenandoah to Cypress who were suddenly being told they needed to optimize for AI search or risk disappearing from the results entirely. There is one problem with the billions-of-clicks figure: Google has not shown its work. No methodology paper, no Search Console filter, no third-party-auditable dataset exists that would allow a Magnolia-area plumber or a Spring-based med-spa to confirm whether a single one of those billions of clicks ever reached their website. According to Search Engine Journal’s June 2025 analysis, Google put a very large number on AI search behavior without providing the underlying data to support it — and that opacity is not a footnote to the AEO story, it is the story. Every business currently spending money on AI Overview optimization is betting real dollars on a metric Google will not prove.

What Google Actually Said — and What It Left Out

Google’s public statements on AI Overview click volume have followed a consistent pattern: a large, rounded number delivered without a corresponding methodology. The billions-of-clicks-per-week figure, which circulated widely after statements from Google executives in 2024 and 2025, carries no attached confidence interval, no comparison baseline, and no explanation of how the company defines a ‘click’ in the context of an AI Overview interaction versus a traditional search result click.

This matters because the definition is genuinely contested. An AI Overview can satisfy a user query entirely within the search results page — the panel answers the question, the user reads it and leaves. Whether that constitutes a positive outcome for the publisher whose content was ingested to generate the answer, or a cannibalization of a click that otherwise would have gone to that publisher’s page, depends entirely on how Google counts. The company has not clarified which scenario its click numbers reflect.

Search Engine Journal noted in its June 2025 coverage that Google’s numbers arrive without the data necessary to audit them — a framing that is more diplomatically stated than the underlying implication warrants. What Google is doing is asking the entire marketing ecosystem to restructure its optimization priorities around a performance claim it is unwilling to subject to independent verification. That is a significant ask, and the local business community in markets like Conroe and Tomball is absorbing the cost of that ask through the agencies and consultants now selling AI search visibility services.

The Search Console Gap: Why Attribution Is Structurally Broken

Google Search Console, the primary channel through which businesses monitor their organic search performance, does not currently segment clicks and impressions by search surface type. A click from an AI Overview citation and a click from a position-one blue link appear identically in the Performance report. There is no filter, no dimension, no secondary breakdown that isolates AI Overview traffic as a distinct channel.

This is not a minor analytical inconvenience — it is a structural barrier to ROI measurement. A roofing company in Spring, TX that invests three months of budget into structured-data schema markup and answer-optimized content rewrites — both legitimate AEO tactics — has no mechanism through which to determine whether any incremental traffic in the following quarter came from AI Overview citations or from ordinary ranking improvements driven by the same content work. The two effects are statistically indistinguishable in any tool currently available to practitioners.

Third-party analytics platforms face the same ceiling. GA4 can tell a business owner that organic search traffic increased 18% in a given month. It cannot tell them whether that increase is attributable to AI Overviews, a core algorithm update, seasonal search volume shifts, or a competitor’s site going down. Call-tracking platforms like CallRail and Invoca log the call; they cannot tag its origin as an AI Overview versus a standard SERP. The attribution chain that would justify AI search optimization spending as a distinct budget line simply does not exist yet.

The irony is that Google controls the data required to close this loop and has chosen — for reasons the company has not made public — not to expose it. Until that changes, any vendor claiming to measure AI Overview ROI for a Hughes Landing-area retail business or a Lake Conroe marina operator is selling inference, not evidence.

The Local Market Exposure: North Houston Businesses at Risk

The business categories most aggressively targeted by AI Overview optimization pitches in the Greater Houston north corridor — HVAC, personal injury law, home services, med-spa, and real estate — are also the categories where the measurement gap creates the highest financial risk. These are service businesses with meaningful average ticket sizes, limited marketing budgets relative to national competitors, and low tolerance for spend that cannot be tied to a phone call or a booked appointment.

A Tomball-area HVAC contractor paying a monthly retainer for ‘AI search optimization’ is, in the current measurement environment, funding a strategy whose performance can only be assessed by a party — the agency — that has a financial interest in reporting positive results. The contractor cannot independently verify whether the work is generating AI Overview citations, whether those citations are driving impressions, or whether those impressions are converting to calls. Google has made that verification impossible by design, whether by intention or neglect.

This does not mean the underlying optimization work is without value. Schema markup, structured FAQ content, authoritative local citations, and E-E-A-T signals are legitimate ranking factors for both traditional and AI-mediated search. The risk is not that the work is worthless — it is that the pricing of that work is being inflated by the perceived urgency of an AI click-volume claim that nobody can audit. Businesses in Magnolia and Oak Ridge North deserve to know the difference between ‘this content work will help your site across all search surfaces’ and ‘we can get you into AI Overviews, and AI Overviews drive billions of clicks.’

See how this applies to your business. Fifteen minutes. No cost. No deck. Begin Private Audit →

How to Pressure-Test an AI Search Visibility Claim

Any vendor or agency asserting that their AI search optimization work is generating measurable results should be able to answer a specific set of questions before a business owner in Conroe or Spring renews a contract. The first question is the simplest: where in Google Search Console is the AI Overview traffic visible as a distinct segment? If the answer is ‘it is blended into organic,’ that is an honest answer — but it is also confirmation that the vendor cannot isolate the channel’s performance.

The second question concerns methodology: how does the vendor define an AI Overview citation for a given client? Manual SERP checks at specific keywords are the most common approach, and they are useful for confirming that a business appears in AI-generated answer panels. But appearing in an AI Overview does not equal driving clicks, and driving clicks does not equal driving revenue. Each step in that chain requires its own measurement mechanism, and currently only the first step — citation presence — is practically verifiable.

The third question is about the counterfactual: would the same content and schema investments produce equivalent traffic gains through traditional ranking improvements, independent of AI Overviews? If a vendor cannot construct that counterfactual, they cannot claim AI Overview optimization as the specific driver of results. Businesses that ask these questions before signing are not being obstructionist — they are doing exactly what the measurement environment requires them to do in the absence of Google providing transparent data.

What Good AEO Strategy Looks Like Without Verified Data

The rational response to an unverifiable channel claim is not to ignore the channel — it is to invest in tactics that generate compounding value across multiple surfaces simultaneously, so that the return does not depend on the unverifiable claim being true. For local service businesses, that means a content and technical foundation that performs whether AI Overviews are sending five clicks a week or five thousand.

Practically, this means: structured data markup that signals expertise and entity relationships to both traditional crawlers and AI indexing systems; FAQ-format content that answers the specific questions local buyers type into search before they pick up the phone; authoritative backlink profiles from local directories, trade associations, and regional news coverage; and Google Business Profile optimization that feeds the local knowledge graph independently of whatever AI Overview activity is occurring. A Shenandoah-area med-spa that builds all four of those layers is positioned well regardless of how the AI Overview click accounting resolves.

The harder discipline is budget allocation. Until Search Console exposes AI Overview as a distinct traffic source — which Google has not committed to on any public timeline — businesses should treat AI search optimization as a category of content and technical work, not as a distinct media channel with its own CPL or ROAS target. Budget it alongside SEO, hold it accountable to the same organic traffic and lead-volume metrics, and do not pay a premium for the AI label until the data exists to justify that premium.

The businesses that will be best positioned in 24 months are not necessarily those that went hardest into AEO in 2025 — they are those that built durable content authority and technical hygiene while their competitors paid inflated retainers for a metric nobody could measure. That is a patient strategy. It is also, given the current state of the data, the only honest one.

Google’s billions-of-clicks figure may be accurate — or it may be a marketing statement dressed in the language of measurement. The honest answer is that nobody outside of Google can currently determine which it is, and that uncertainty is not a temporary gap to be closed by the next Search Console update. It reflects a deliberate choice about data transparency that Google has not been publicly pressured to reverse. Over the next 12 to 24 months, as AI Overview adoption either accelerates or plateaus and as the gap between Google’s claimed click volumes and publisher-observed referral traffic either closes or widens, the businesses that will have come out ahead are those that treated the unverified claim as exactly that — and invested instead in content quality, technical credibility, and audience relationships that compound regardless of which search surface happens to be the intermediary.

Sources

• Search Engine Journal — Primary source establishing that Google’s AI Overview click-volume claim lacks auditable supporting data or methodology
• Google Search Central — Search Console Help — Reference for current Search Console Performance report dimensions, which do not include AI Overview as a distinct filter
• Stratechery — Aggregation Theory — Framework for understanding how platforms that control demand (Google) extract value from suppliers (publishers) by controlling the measurement layer