Amazon wants to own the truth about your product

Here's the uncomfortable version. You spend years curating your product data. Dimensions, ingredients, variants, compatibility. And then someone else becomes the place that explains your product to your customers.

That's where a plan reported by Reuters on Friday is headed. According to an internal document, Amazon is working on a "World Product Graph": a database covering nearly every product sold anywhere in the world, including items Amazon doesn't stock.

What the document says

• It's an internal PR/FAQ, Amazon's standard format for proposed products. The aspirational launch date is October 1, 2027.
• Some of the data would come straight from brands and retailers. Per Reuters, Coca-Cola and Bayer have signed up and Procter & Gamble is in a trial. Amazon would also crawl other retailers' websites itself. Reuters couldn't determine whether that would require permission.
• The stated goal: customers shouldn't have to leave for Google, brand sites or review sites to research a product.
• Amazon says draft, outdated documents often paint an incomplete picture. Reuters' sources say the project could still be changed or dropped.

Amazon locks agents out and still wants to be their source

The real driver is AI shopping agents. Amazon executives worry, Reuters reports, that agents will send shoppers wherever product data is most complete. So Amazon wants to own the most complete data. At the same time it keeps outside agents away: Meta's Muse was blocked in September, and Amazon has been suing Perplexity since last year.

Translation: the door stays shut, but the shop window gets so good that every agent ends up standing in front of it. It's the pattern I keep coming back to. Agents are becoming gatekeepers, and the big platforms decide which agents get in and on what terms. Whoever holds the data holds the leverage. And per Reuters, the plan also protects Amazon's growing ad business.

For brands this isn't a data project. It's a value transfer

It sounds like a service: hand us your product data and we'll make sure customers find it. It isn't. If you deliver your data to one platform, you make that platform the best source on your own product. My rule: platform terms are a value transfer. Read them first, then deliver.

How open this field still is shows up in a small probe. Today, Nov. 14, 03:29 CET, I asked ChatGPT and Claude (in German) where to research a product before buying online. ChatGPT led with customer reviews on platforms like Amazon or Otto, then test sites like Stiftung Warentest, then YouTube and blogs. Claude named Stiftung Warentest, Chip, Computerbild, Idealo, Geizhals and Reddit. Neither answer mentioned the manufacturer's own website. Limits: an API sample without web search, one answer each, results vary from run to run.

What I'd do now

• Make your own product page the most complete source: every attribute, clean and structured (schema.org Product), not just images and claims. That's an owned asset, not a feed for someone else.
• Before you sign any platform data deal, read the terms. Who can use the data for what, including around ads and competitors?
• This week, put five real customer questions about your category, without your brand name, to ChatGPT, Claude and Gemini. Note which sources they cite. If your site isn't there, you know where the work is.

Bottom lineIf Amazon is the only one you give your product data to, Amazon gets to explain what you sell.