FERMÀT

The winners in marketing won’t be the teams using the most AI. They’ll be the teams built best for it. What gets automated, what stays human, and where judgment becomes even more important. If you’re thinking seriously about that, come discuss it with peers at The Oasis in Cannes, June 22nd & 23rd 12-4:30pm. Register here: https://lnkd.in/gUpj-4cz

Most teams are still building AI search strategy on the assumption that SEO carries over. It doesn’t. We pulled 2,800 queries across 30+ consumer brands to validate. Ranking on page one of Google gives you an 8% chance of being cited by ChatGPT. ChatGPT and Google AI Overviews have even less in common with each other than either has with classical Google. These are different retrieval systems indexing different parts of the web. Your SEO playbook isn't going to carry you here.

Backcountry's VP of Product just spoke with James Hercher at AdExchanger about FERMÀT's Commerce Graph. Joe Torreano came in skeptical. The first beta build was, in his words, "kind of rough." He'd been paying legacy vendors hundreds of thousands of dollars a year for similar capabilities. Weeks later: Commerce Graph does the job just as well (if not better), with more granular tracking, at a fraction of the cost. Heatmaps on the homepage and PDPs. Agentic NLP to answer questions like "does email traffic degrade faster than other sources?" or "does featuring a certain product improve average order volume?" Connected behavioral signals across social platforms, on-site, and beyond. What stuck with me most was Joe's broader take: right now you still need a human to vet AI output. Ecommerce traffic has too many nuances: social visitors behave differently than organic, TikTok trends affect conversion in ways that aren't obvious. Agents don't yet have those instincts. But once they do, he said, we'll start letting them make decisions without a human approving each one. "Human out of the loop." That's the bet we're building toward. Grateful to Backcountry and Joe for being early believers in FERMÀT and honest partners through the beta. Full piece on AdExchanger's site and linked in the comments.

We just shipped a FERMÀT MCP server. Ask Claude or ChatGPT which experiences are converting, how shoppers are navigating your pages, or where your brand is showing up in AI search results. Your commerce insights should meet your team wherever they already work.

14% of shopping queries on Google now trigger an AI Overview. For "best [product]" queries, that number is over 80%. Google's Shopping Graph is consistently a top-5 cited domain across all of them. And FERMÀT brands are showing up inside it. We now have 1,400+ valid product snippets recognized by Google from shopfermat.com, up ~30% from last month. Ranking for high-intent queries like "best suitcase under $300" and "best climbing shoes." Here's why this matters: most brands can't control whether they appear in AI Overviews. Their own site might not have the structured data Google needs. But FERMÀT's product pages are built for exactly this. We ingest structured product data in the format Google's Shopping Graph picks up. That means FERMÀT is an additional distribution surface. Even when a brand's own site doesn't make it into the AI Overview, their products on shopfermat.com can. The brands investing in AI search visibility now are building a compounding advantage. The ones waiting are ceding shelf space they didn't know existed.

I was surprised to learn that across our customers, Google Shopping drove just 2x the engagement of Meta discovery ads. Here's what we found. We tracked 30M+ sessions across our brand network, comparing Google and Meta traffic side by side. On brand sites, Google visitors view 1.36 products per session. Meta visitors view 0.85. Google visitors stay 2x longer (5.2 min vs. 2.4 min), click 2.6x more, and scroll 38% deeper. On FERMÀT storefronts, Google Shopping (PLA) visitors browse 2.06 products vs. 1.62 for Meta DPA. This held across every single brand. No exceptions. That surprised me. I would've thought high-intent traffic engagement would be roughly 4-5x cold, no-intent traffic. Especially since you're often bidding on branded terms vs. cold traffic where I would have expected more fast bounces. A 2-3x gap in on-site behavior between someone who searched for exactly what they wanted and someone who was interrupted mid-scroll on Instagram. That tells you something important about how people actually shop in lifestyle and DTC categories. The experience after the click matters more than the intent before it. How does this shape your ad buying now that you've seen the data?