Freddie Chatt

Freddie ChattSep 7, 2026

Everyone is still scrabbling about with GEO, AEO, AI Search 12 months of testing across client sites and my own brand, and here's where I've landed: It's ecommerce SEO fundamentals. Nearly all of it (maybe 90%) The same things that get your category pages ranking in Google are the things that get your products named in ChatGPT, Claude and Perplexity. So before you buy a tool that promises AI visibility, here's what's actually moving the needle: 1. Trust signals AI tools lean heavily on third party reviews to decide whether a brand is credible. For ecommerce that means: • One strong review profile kept current. Trustpilot, JudgeMe, whichever your category leans towards • Product level reviews on the product page itself, not just a badge in the footer • Seller ratings in Merchant Centre • Replying to the bad ones properly. The response gets read as much as the star rating 2. Page structure Most ecommerce content is written for a shopper who already knows what site they're on. That's a problem when a model is reading it out of context. • Name the brand and the product in full. Not "our bestseller", but "the Village Cricket Co. Player Bat" • Category pages need actual text that answers buying questions, not a grid with 40 words underneath • Put specs in a table and in prose. Size, material, care, delivery, returns 3. Authority building for relevance Topical relevance beats volume here. • Target sites where a mention of your brand makes obvious sense to a reader • Digital PR and gift guides in your category do a lot of work • Don't dismiss nofollow links from big publishers in your niche Random link buys from unrelated sites do nothing. 4. Topical authority still rules LLMs are trained on the same web Google indexes. The more depth you have around what you actually sell, the more often you get named. Which also means clearing out the rubbish. Most ecommerce blogs are full of posts written for traffic that has nothing to do with the products. If you sell running shoes, the 2019 post about office Christmas party ideas isn't neutral. It's diluting what you're known for. 5. Be where the crawlers are Reddit (yes I know it's been 'hit'), YouTube, LinkedIn, roundup and comparison articles. For ecommerce specifically, the "best X for Y" listicles matter more than most brands realise. That's the content models pull from when someone asks for a recommendation. The more places a model sees you in the same context, the more it treats you as the answer. 6. Technical setup • Server side rendering for anything that matters. • Product schema with price and availability, kept accurate • Clean architecture, nothing more than three clicks from the homepage • Faceted navigation under control 7. Traditional search feeds AI Most of these tools are pulling from Google or Bing for anything current. Rank better and you show up more. That hasn't changed.

Freddie ChattSep 9, 2026

Attribution is overrated. Incrementality is what actually matters. Every brand I work with wants to know what's working. ROAS looks fine. CAC is holding. Revenue is up. So everyone carries on. But there's a question underneath all of that. Your Meta ad got the conversion. Did it cause the conversion? That's the whole difference, and most dashboards stop before they get there. I'll be honest about which side of this I usually sit on. SEO is often the channel taking credit it hasn't earned. A brand runs a big paid social push, branded search goes up, branded organic revenue goes up, and the SEO report looks brilliant. I didn't do that. The ads did. It runs the other way too. Someone reads a comparison post, doesn't buy, then comes back through a branded search two weeks later and paid picks up the credit. You don't need a full marketing mix model to start pulling this apart. Four things that work: Geo split testing. Pick two similar regions. Run in one, hold the other back. Then watch branded search volume and revenue in both. If the test region moves and the control doesn't, that's real lift. New customers versus returning. If your growth is repeat buyers coming back, you're harvesting demand you already had. If new customer count is climbing across every channel at once, something is genuinely creating demand. Paid spend against branded search. This is the simplest check nobody runs. If spend is climbing and branded search is flat, you're paying to close sales that were already coming. If branded search lifts alongside spend, the ads are creating interest rather than catching it. Ask people. A post purchase survey with one question, where did you first hear about us, will tell you more than most attribution tools. The answers rarely match the dashboard. If you only measure last click, you'll scale whatever is best at standing near the finish line.

Freddie ChattSep 10, 2026

"How did you hear about us?" Five words. Costs nothing. And it tells you more about what's driving revenue than most of your paid analytics tools. There’s a problem with last-click reporting. It crowns whichever channel closed the sale. Usually branded search or direct. Meanwhile the blog post, the newsletter, the paid ad that created the demand in the first place get zero credit. So you cut them. Then wonder why branded search dries up three months later. A post-purchase survey breaks that loop. You ask the buyer directly, right when they convert, and you get the first-touch influence GA4 can't see. If you ever have to defend your marketing budget, this is the data that does it. I'm running mine through Zigpoll. One question, live at checkout.