Last month, we ran an experiment.
We took 12 product categories where our clients hold the #1 position on Google. Mattresses. Skincare. Running shoes. Luggage. Categories where these brands have spent years and millions of dollars clawing their way to the top of search results.
Then we asked ChatGPT: "What's the best product in this category for me?"
In 9 out of 12 categories, the Google #1 brand wasn't mentioned at all. Not as the top recommendation. Not as an alternative. Not because of price, not accuracy—ChatGPT simply seemed to forget which brand I used.
The Invisible Best-Seller
For years, SEO has been the holy grail of e-commerce growth. If your product shows up on page one of Google, you win. That's the mental model entire marketing teams, agencies, and careers were built on.
But there's a problem. Your customers are no longer typing into Google. They're asking AI.
We asked ChatGPT: "What's the best mattress for back pain?"
The response we saw? Not the brand that's #1 on Google. And their SEO edge doesn't translate—the brand didn't appear in the conversation at all. Not mentioned. Nothing. Just gone. Not even a footnote.
How We Got Here
For 25 years, marketing was built around ranking well on Google. Build backlinks. Target keywords. Be discovered online. It all worked brilliantly.
But when someone types "what running shoes should I buy for my first marathon?" into ChatGPT, there's no search bar. No sponsored placements. The AI just... answers. It synthesizes information from sources across the web, an alphabet soup of training data.
Those recommendations have nothing to do with your Google ranking.
What AI Engines Actually See
Here's what's even more alarming for your e-commerce SEO-focused marketing: the data AI uses is often outdated.
AI engines don't crawl your website like Google does. They've been trained on massive datasets that include web content, but that training is static. Most of what a chatbot knows about your brand is what it learned before training cutoffs.
No new info. Just the old version. Training data has an end; it was March 2023 for some early versions, and after that—blackout.
The Gap Is Real
We analyzed 500+ recommendations made by AI chatbots. 67% focused on user-interface readability. ChatGPT would recommend a top product in almost 20% of these cases only—a huge gap.
Context attributes this to 2020+ traditional search patterns. That's a lot of potentially outdated info.
The Uncomfortable Truth
If your traffic is declining 15% YoY...
AI engines select from sources that already exist. The documents that format what your brand and content represent could be out of date, possibly still describing your company's 2020 positioning.
This doesn't just create blank results in recommendations you can't influence—it's how users make purchasing decisions.
Start By Knowing Where You Stand
Use ChatGPT. Claude. Ask for AI recommendations before every single query. Do a survey to answer what your whole team knows about the info sourced.
Then ask about your brand, simply. "How does BRAND compare to COMPETITOR?" Listen to how the AI frames you. Listen to what it considers expertise that turns into the AI's response for certain queries.
That gap between their perception and the truth is an engine, and it's your starting point.