How do AI Overviews affect ecommerce traffic?

AI Overviews appear most often on questions with an explanation as the answer: how to choose a mattress, whether merino is warmer than cashmere, how to measure for a ring. On those queries the summary satisfies a share of searchers who would previously have clicked a guide, and the pages that used to earn that click see impressions hold while clicks fall.

On commercial queries such as best running shoes for flat feet, the Overview often includes a product panel or carousel. Those products come from the Shopping Graph, which is fed mostly by Merchant Center. A store with complete product data can appear in that panel; a store that relied on a blog post ranking for the same query cannot.

On transactional and navigational queries, the brand name plus a product, a model number, a size and color, AI Overviews are rare. That traffic is largely untouched. So the honest answer for any given store is: work out the mix. A store whose organic revenue came from category and product pages will notice far less than a store whose top pages are buying guides.

The pages most exposed are the ones written to answer a question in prose. Sizing guides, ingredient explainers, care instructions, versus articles and gift lists all sit on the queries where Google is most willing to summarize. Expect those pages to keep their impressions and lose part of their clicks.

Category pages sit in the middle. Their queries are often commercial, so they compete with the product panel inside the Overview and with the classic results underneath it. Category pages with strong product data behind them can gain a second route in through the panel, because the products listed there link to product pages, not to the category.

Product pages are the least exposed and, through Merchant Center, the most likely to gain a placement. That is the practical shift AI Overviews create for a store: the unit of visibility on commercial queries is increasingly the product record, with its price, availability, image, rating and shipping details, rather than the page that ranks.

Google includes clicks and impressions from AI Overviews in the Web search type in Search Console and provides no filter to isolate them. There is no report that says how many sessions came from an Overview. What you can do is look for the signature.

Group your pages into types: guides, category pages, product pages, policy and brand pages. For each group, compare the last 90 days with the same 90 days a year earlier and look at impressions, clicks and average position together. A group where impressions are flat or up, position is stable, and clicks are down is the group losing clicks to summaries. A group where impressions rose sharply with a much lower click-through rate may be appearing inside Overview panels, since Google counts an impression when the page is linked within one.

Keep the comparison to the same page types rather than the whole site. Site-wide click-through rate mixes the losers and the gainers together and tells you nothing you can act on.

When the Overview carries a product panel, the question changes from which page ranks to which products qualify. The inputs are the ones Merchant Center already scores: an approved free listing, a title that says what the item is, price and availability that match the page, a GTIN where one exists, a clean image, shipping and return details, and ratings where a store has genuine reviews.

The product page itself still matters, because Google crosses feed data with what it crawls and reads product structured data to confirm price and stock. A page that disagrees with its feed is a weaker candidate than one that agrees.

This is why we treat feed quality as an organic search task now, not only a paid one. The same attributes that decide Shopping ad relevance decide whether a product is available to appear when Google composes an answer.

Measure revenue by landing page type, not sessions. If guide traffic fell but organic revenue held, the lost clicks were mostly readers who were not going to buy from that visit. If organic revenue fell with it, the guides were doing assisted work and need a replacement path, usually stronger internal links from the guides that still get clicks into the categories they support.

Get the product data complete. Free listings in Merchant Center, product structured data on every product page, and a monthly check of the Limited status list. This is the direct route into the panels that now sit on commercial queries.

Then rewrite guides for the question after the summary. An Overview can say what to look for in a hiking boot. It is much weaker at saying which of your boots fits a wide foot, ships to Alberta in three days and can be returned worn once. Content built on your own product data, your own fit notes and your own policies is harder to summarize away, and it is the content that helps a shopper who has already read the Overview decide.