Google Knows What Your Shopper Wants Before You Do

Donald Fisher opened The Gap in 1969 because he noticed something specific. Young men in San Francisco could not find Levi's in their size. The store wasn't a vision statement. It was a gap in supply literacy. Fifty-seven years later, Google is doing something structurally similar. It is watching what people want, in the language people actually use, and building the shelf before the brand does.

Last week, Google launched AI Performance Insights and Conversational Attributes inside Merchant Center. Then it expanded Direct Offers with AI-generated bundles, native checkout, and travel deals. These are not incremental features. They are a quiet annexation of the consideration layer that brands spent the last decade pretending they owned.

The Shift Hiding Inside a Product Update

Here is what Conversational Attributes actually means. Google is now parsing how real people describe products in natural language queries and mapping those descriptions back to your product feed. If your feed speaks in SKU logic and spec sheets, and your shopper is asking 'something I can wear to a work dinner that isn't trying too hard,' you are invisible. Not penalized. Invisible. The algorithm doesn't punish you. It just doesn't see you as an answer.

AI Performance Insights compounds this. The tool surfaces what product attributes are driving conversion, which bundles are finding traction, and where your catalog has dead weight. It is audience research handed to you pre-digested. The brands that will use it are the ones with clean enough data to act on it. The ones that won't are still debating who owns the product feed internally.

SparkToro's work on AI visibility tracking is worth sitting with alongside this. The question their team posed in March was blunt: can you even measure where your brand surfaces inside AI-generated answers? The honest answer remains murky. But the directional signal is clear. Brands with rich, consistent, structured product and content data appear more often. Not because AI favors them philosophically. Because AI has more to work with.

Who Loses the Window

Everlane is a useful adjacent signal here. 2PM's recent brief on the brand notes it is operating at the edge of what DTC fashion can sustain. The identity cohort it built in the 2010s, the tribe that bought into radical price transparency as a kind of status ritual, has fractured. Younger shoppers do not carry that pretense. They want the thing. Not the ideology around the thing. Brands like that one are doubly exposed in the AI-mediated feed environment. Their product data was built around a story. AI bundles are built around use cases and price signals.

The aluminum tariff situation reported in the Tomahawk Tax piece adds a harder edge. Consumer brands with compressed margins and aluminum-dependent packaging are already watching their cost structure bend. If your unit economics require you to delay feed optimization work because the team is fighting supply fires, you are ceding the AI visibility window to a competitor who is not in that position. The disruption is asymmetric. Operationally stable brands compound their advantage quietly while others are distracted.

Your Specific Move

The arbitrage window is narrow and it is not advertised. While Google rolls out these tools broadly, the brands enriching their product feeds this quarter are writing the training data that AI-generated bundles will favor for the next several cycles. Attributes like occasion, use context, sensory description, and social permission signals, these map to how people actually talk about what they want to buy. Your catalog probably doesn't include them. Most don't.

Run a feed audit against your top 40 SKUs. Pull the Conversational Attributes report once it is live in your Merchant Center account. Cross-reference with the queries your paid search team is already seeing in search term reports. You will find a vocabulary gap. The distance between how your product is described and how your shopper thinks about it is exactly the territory Google's AI is now trying to bridge on your behalf. You should bridge it first.

The brands that move in May and June are the ones writing the shelf tags for how their category gets bundled and surfaced by fall. That is not a forecast. It is just how the timing on these rollouts tends to play out.

Three Questions to Pressure-Test

Does your product feed describe what your shopper would say, or what your warehouse system requires? If Google's AI bundled your top category right now, would you appear in it, or would a competitor with better attribute data take the slot? When your team is pulled into supply chain or cost structure fires this quarter, which competitor is quietly enriching their feed while you're occupied?