For years, product feeds lived in a quiet corner of the marketing org chart. If you ran Shopping campaigns, someone optimized the feed. If you didn't, it was someone else's problem. That mental model is now costing companies real money.
Google's recent Ads Decoded episode on building a retail growth engine made something clear that many marketing leaders have been slow to internalize: product data is no longer a channel task. It's infrastructure. And the companies treating it as such are pulling ahead in ways that don't show up in a single campaign dashboard.
Nadja Bissinger, Group Product Director for Retail on YouTube, described Merchant Center feeds as the backbone that powers organic and ads experiences. That's not marketing language. That's an architectural statement. When the same data set powers free listings, AI-powered search experiences, YouTube formats, Google Lens visual searches, virtual try-on, and emerging agentic commerce surfaces, the feed stops being a Shopping campaign input and starts being a visibility layer across the entire Google ecosystem.
The math here matters. According to Search Engine Journal's analysis, Google Lens now processes more than 20 billion visual searches per month, with one in four carrying commercial intent. That's a discovery channel most marketing teams aren't even tracking, let alone optimizing for. And it runs on the same product data sitting in Merchant Center.
The Measurement Gap Is Where the Money Leaks
Here's the problem I see repeatedly when working with marketing teams: they're measuring feed value with a scorecard built for 2019. If the main question is whether Shopping ROAS improved last week, it becomes easy to undervalue the broader impact of stronger product data.
Google cited a 33% conversion uplift for advertisers using Demand Gen with product feeds during the podcast discussion. That's not a Shopping metric. That's a signal that feed quality is being tied to campaign types that span YouTube, Gmail, Discover, and the Google Display Network. The value shows up across multiple touchpoints, assisted paths, and blended performance trends – not in one campaign report.
A stronger title may improve discoverability. Better imagery can increase engagement in visual placements. Accurate pricing and promotions can improve click appeal. Richer attributes can help Google better understand relevance. Availability data can support local and omnichannel visibility. Those gains may show up in places your current reporting model doesn't look.
This is why I keep telling teams: if you're only measuring feed work by Shopping campaign performance, you're using an outdated scorecard. The value is there, but your reporting model was built for an earlier version of Google.
What Google's Spec Changes Tell Us About Where This Is Going
The 2026 Merchant Center product data specification update isn't just housekeeping. It's a signal about where Google wants product data to go.
Three changes stand out. First, Google expanded product-level shipping controls – handling cutoff time, minimum order value, and loyalty-related shipping labels can now be specified at the product level rather than just account-wide. That's Google saying: we want more precision, and we want it in the feed.
Second, a new video_link attribute lets merchants submit product videos. Serving and quality validation begin . That's Google saying: static images aren't enough for the surfaces we're building.
Third, the minimum image resolution requirement is increasing to 500×500 pixels, with enforcement starting . That's Google saying: if your visual assets can't hold up on a 65-inch television – which is where shoppable CTV inventory is heading – you're going to lose visibility.
These aren't arbitrary requirements. They're infrastructure investments Google is asking merchants to make so that product data can power experiences that don't exist yet. The companies that treat this as compliance work will do the minimum. The companies that treat it as competitive positioning will pull ahead.
The AI Layer Changes Everything
The most important shift isn't about any single attribute. It's about what happens when AI agents start shopping on behalf of consumers.
Research on AI shopping behavior shows that stores with 99.9% attribute completion – what the industry calls a Golden Record – are seeing 3-4x higher visibility in AI recommendations compared to stores with sparse data. That's not a small edge. That's the difference between being recommended and being invisible.
Here's why: AI agents operate on confidence scores. When a customer asks for machine washable rugs under $200 in modern style, the agent scores every product in its index based on how well it matches those criteria. If your data explicitly says care_instructions: machine washable and style: modern, you get a 95% confidence match and appear first. If your description mentions easy to clean without the explicit attribute, you get a 62% confidence match and rank eighth – or get excluded entirely.
The difference between first recommendation and eighth is often just data completeness, not product quality. That's a brutal reality for brands that have been coasting on brand recognition without maintaining their backend data.
What This Means for Marketing Org Structure
Feed optimization is no longer just a PPC responsibility. It can influence organic visibility, merchandising strategy, creative presentation, promotions, and how products appear in AI-led experiences.
For larger organizations, that may require closer coordination between paid media, SEO, e-commerce, merchandising, and product teams. The feed becomes a shared asset that multiple functions depend on, which means ownership and governance need to be explicit.
For smaller brands, it may be as simple as giving feed quality the same level of attention already given to ad copy, landing pages, and campaign structure. Many advertisers still treat feed work as cleanup work. That mindset is becoming expensive.
The practical question for any marketing leader is: who owns feed quality in your organization, and how is that work prioritized against other marketing investments? If the answer is it's the agency's job or it's in the e-commerce team's backlog, you may be underinvesting in one of the highest-leverage assets in your marketing stack.
The Board-Ready Version
If I had to summarize this for a CFO or board presentation, here's how I'd frame it:
Google is repositioning product data from a campaign input to a visibility infrastructure layer. The same feed now powers Shopping ads, free listings, AI-powered search, YouTube formats, visual discovery, and emerging agentic commerce surfaces. Companies that treat feed optimization as a maintenance task are measuring value in the wrong place and underinvesting in a competitive asset. The companies that treat it as infrastructure – with appropriate governance, cross-functional ownership, and ongoing optimization – will capture disproportionate visibility as Google continues expanding where products can appear.
The risk isn't that feed work doesn't pay off. The risk is that the payoff shows up in places your current reporting model doesn't look, so you keep underinvesting while competitors pull ahead.
Model or it didn't happen. And right now, most marketing teams don't have a model for how feed quality translates to visibility across Google's expanding commerce surfaces. That's the gap worth closing.
