Google quietly rolled out a beta feature inside Merchant Center on Tuesday that performs a single, structurally consequential trick: it scans a retailer's website once and uses Google's AI to automatically generate the product catalog without requiring a structured product feed. PPC Land was first to flag the change, reporting that the feature now sits inside the "Add product source" section of Merchant Center alongside the traditional structured-feed and API options. Google's own description of the feature, paraphrased from the Merchant Center UI: it's the easiest way to get started and it reduces errors from manual entry.
For anyone who has not personally onboarded a small or mid-sized retailer to Google Shopping, that one-paragraph announcement may not register as significant. For the people who have done that work, this is the shoe everyone in the digital-marketing-services industry has been waiting for Google to drop for two years.
The structured product feed is the core artifact that the entire Google Shopping ecosystem has been built around since 2012. To list products on Shopping, a retailer needs to produce a feed file — typically XML or TSV — that conforms to Google's Merchant Center spec. Title, price, SKU, GTIN, image URL, availability, brand, condition, color, size, age group, gender, and a dozen other attributes have to be mapped correctly. Get a column wrong and the feed gets rejected. Mapping that feed is the single most common reason small retailers either never start running Shopping ads or quietly stop running them after their first month.
The feed-management category that emerged to solve that problem is not small. Feedonomics, acquired by BigCommerce for $300 million in 2021, charges retailers monthly fees that can run from a few hundred dollars to north of $50,000 depending on catalog complexity and channel count. ChannelAdvisor, GoDataFeed, DataFeedWatch, Channable, and a long tail of agency-built feed managers all monetize the same friction. The category as a whole is several billion dollars of annual revenue. Google's new beta is not — yet — a direct replacement for any of those products. But it is the first time Google has acknowledged at a product level that the friction itself is the bottleneck, and that AI can compress it.
The feature, as currently scoped, is best for smaller catalogs and quick-setup scenarios. Retailers with thousands or tens of thousands of SKUs, complex variant structures, multilingual catalogs, or omnichannel feed needs (Google + Amazon + Meta + TikTok + Walmart Connect) still need the dedicated tooling. The feed managers are not going to lose their enterprise accounts on the back of a Google beta.
But two things will compound from here. The first is that Google's catalog-extraction model will improve. The same generative-AI advances that let LLMs parse complex web pages, extract structured data, and produce clean entity definitions are being applied to product catalogs. The accuracy threshold at which the AI-extracted catalog rivals a hand-tuned feed is somewhere between two and five years out for mid-market retailers. The second is that competitive pressure will force the same AI features into adjacent platforms. Meta's Commerce Manager will get an analogous tool. TikTok Shop will. Amazon Seller Central likely already has internal experiments. The friction Google just removed for Shopping Ads will be removed across every retail-media platform in the cycle ahead.
That changes the value proposition of the feed-management category. The category's current pitch is "we save you 40 hours of feed engineering and prevent the rejection rate that kills your ROAS." The pitch becomes "we save you the marginal optimization above the AI-extracted baseline." That's still a real product. It's a smaller product, sold at a smaller price, to a smaller subset of customers.
The retail read-through has two layers worth flagging. First, for direct-to-consumer brands and mid-market retailers who have been on the fence about Shopping Ads because the onboarding cost is too high — this lowers that cost meaningfully. Expect a wave of new advertisers entering Google Shopping over the next two quarters, which will drive CPC inflation in the categories that get most quickly populated and likely depress ROAS for incumbents. Second, for the retail-media networks (Walmart Connect, Amazon Ads, Target Roundel, Kroger Precision Marketing, Best Buy Ads) that compete with Google for retail ad budgets, the bar just got higher. Google's pitch to the long-tail retailer used to require a feed-management partnership. Now the pitch is "give us your URL." That's a faster sales cycle, a wider funnel, and a stickier integration.
Feedonomics' positioning, and the positioning of every feed-management vendor in the space, has historically been "we make catalog-data work the way it should across channels." That's still true. But the marginal value of "making catalog data work" is being squeezed by exactly the kind of AI-native primitive Google is now shipping. The companies in this category that survive the next two years will be the ones that pivot from feed engineering to catalog enrichment, attribute optimization, and channel-specific creative — the work above the baseline, not the baseline itself.
For retailers, the takeaway is simpler. If you've been staring at a Merchant Center setup task on someone's roadmap for three quarters, this is the week to revisit it. The cost just went down.