Search Console's new AI report lights up with a number, and it is tempting to read that number as an answer. It is not. The Generative AI performance report that Google rolled out in June 2026 counts impressions only, how often a link to your pages showed up inside an AI Overview or AI Mode answer. It has no column for clicks, no column for revenue, and no way of telling you whether any of those impressions sent a single shopper to checkout. A team that opens that report, sees the chart move, and treats the movement as the story has found a graph, not a measurement of what AI Overviews ecommerce traffic impact actually did to the business.
That gap matters because the discourse around AI Overviews has outrun what any single report can confirm. Ahrefs' study of 300,000 keywords found AI Overviews correlating with a 34.5 percent drop in click-through rate for the top organic result on informational searches, comparing March 2024 against March 2025. That is a real, well-documented effect. It is also a finding about informational queries specifically, and 99.2 percent of the keywords in that study that triggered an AI Overview were informational in intent. An ecommerce brand reading that headline and assuming its product listing pages took the same hit is extrapolating past what the data actually covers.
What Search Console's AI report can and cannot tell you
Google's own documentation for the Generative AI performance report is specific about scope. The report covers two surfaces, AI Overviews and AI Mode, segmented by page, country, device, and date, and it tracks impressions: how often a link to your site appeared inside a generated answer. Three limits matter more than the headline feature:
- No click data. The report cannot tell you whether an impression turned into a visit, which means it cannot tell you whether an AI Overview citation helped or cost you anything on its own.
- No Discover overlap. AI features inside Google Discover get their own separate report, so a page's full AI-surface exposure is split across two places that do not sum automatically.
- A visibility floor. Sites without enough AI-feature impression volume will not see the report populate at all, which quietly excludes a lot of smaller or newer ecommerce domains from even having this half-picture to check.
That third point is worth sitting with. A store with too little AI-feature volume to chart is not necessarily unaffected. It may simply be invisible to the one tool built to measure the effect, which is a different thing and a worse one to assume away.
The gap that actually matters: product pages versus everything else you publish
Most ecommerce brands do not live or die on their product pages ranking for informational searches. They live on their product and category pages converting transactional intent, "buy," "shop," "price," not "how does," "what is," or "best way to." Ahrefs did not measure click-through change on commercial queries directly, its study was scoped to informational keywords on purpose, but its own sample explains why: 99.2 percent of the keywords that triggered an AI Overview in that data were informational in intent. AI Overviews are heavily concentrated on the kind of search a buying guide or FAQ page competes for, not the kind a PDP or category page is built to rank for. That is evidence AI Overviews show up far less often where a product page would be competing, not proof those pages are untouched when one does appear.
The exposure most brands are underestimating sits one layer away from the storefront: the buying guides, comparison posts, size charts, and FAQ content a marketing team built specifically to rank for the informational searches that feed the funnel before someone is ready to buy. That content is exactly the kind an AI Overview is built to summarize directly on the results page, which means a brand's top-of-funnel content can be bleeding clicks for months while its product and category pages look stable, and a single blended "organic traffic" number in GA4 will average the two together and hide which half is actually the problem.
A separate AI measurement gap we have covered, AI assistant referral traffic landing in Direct instead of its own channel, comes from a different mechanism, missing referrer data rather than a summarized answer page, but the lesson rhymes: the number missing from a dashboard is not a number that stopped happening, and a blended number that looks stable can be hiding two opposite trends underneath it.
The second report nobody checks: Merchant Center's AI shopping insights
While Search Console's report answers a visibility question for the open web, Google Merchant Center is rolling out a separate AI insights report built specifically for sellers, currently launching across the US, Canada, Australia, India, and New Zealand. It covers four things Search Console never touches:
- Share of voice against comparable competitors specifically on AI Mode, AI Overviews, and Gemini.
- Shopping funnel performance broken out by discovery, evaluation, and purchase stage.
- Product term insights, the actual phrases shoppers type into conversational AI when searching for products like yours.
- Product attribute gaps, a score showing which of your products are missing structured details, color, material, style, that these systems use to match a shopper's question to a specific SKU.
That last point is the most actionable item in this entire topic, because it is not a traffic trend to interpret, it is a list of missing fields in your own product feed. A brand that has never looked at this report has no idea whether its catalog is even eligible to be matched against the queries its own customers are typing into these tools.
How to actually measure the impact instead of guessing from one chart
Reading the Search Console and Merchant Center reports side by side closes part of the gap, but the number that settles an argument in a leadership meeting is still sitting in GA4, not in either AI-specific report. Three checks turn a feeling into a measurement:
- Split organic sessions and revenue by landing page, not just by channel, and compare the trend for informational content (guides, comparisons, blog posts) against transactional pages (product and category pages) over the same window. A decline concentrated in the informational set matches the pattern the wider data shows. A decline on product pages points elsewhere first.
- Reconcile Search Console impressions against GA4 clicks and sessions for the same pages, the same discipline a generative engine optimization pass depends on before any schema or crawlability fix gets credit for a result it did not cause.
- Fix the Merchant Center attribute gaps first, before touching anything speculative, because they are the one lever in this list that is fully within your control and has nothing to do with how any AI model happens to be summarizing the web this quarter.
None of those three steps require waiting for an industry benchmark to settle. They use data you already have.
Before you tell your team AI Overviews are the reason
The checklist that actually holds up in a monthly review looks like this:
- Pull the Generative AI performance report and note the trend, but label it a visibility signal, not a revenue one.
- Pull the Merchant Center AI insights report once it is available for your market and fix the structured attribute gaps it flags.
- Split GA4 organic sessions and revenue by landing page type, informational versus transactional, before blaming or clearing AI Overviews for a change in the blended number.
- Reconcile the pages where impressions moved against the sessions and revenue those same pages actually produced.
- Revisit the split quarterly, because the Ahrefs CTR gap itself widened sharply within a single year, and a static read today will be stale by the next board update.
A chart that moves is not the same thing as a business outcome that changed. If your current reporting cannot separate "visibility shifted" from "revenue shifted," that is a gap worth closing before the next AI Overviews headline lands in your inbox and someone asks what it means for your store specifically. An analytics audit maps exactly that kind of blind spot across Search Console, GA4, and your ad platforms in one pass, and a tracking audit checks whether the underlying GA4 events feeding that landing-page split are even firing correctly in the first place, so the number you bring to that meeting is one you have actually verified rather than one you are hoping holds up.

