Guide · Inspeccia
Amazon shows up and your store doesn't: marketplaces vs your site in AI answers
You ask ChatGPT where to buy the product you manufacture and the answer names three marketplaces. Your store — the one that made the thing — is not there. The intuitive conclusion is that the game is rigged for the giants, and it is a reasonable conclusion: in classic SEO it largely was. But OpenAI's documentation on how merchants get ranked contains a line almost nobody read, and it changes the whole diagnosis. One of the ranking factors is whether you are the maker or primary seller of that item. It is the one card a marketplace cannot play on your product. And you hold it every time.
So the right question is not "how do I compete with Amazon?". It is: if I have a documented advantage, why am I losing anyway?
The four factors, and which one is yours
Worth anchoring this in the source. OpenAI documents that when a user clicks a product, ChatGPT may show a list of merchants offering it, and that this list "is generated by ChatGPT based on merchant and product metadata from third-party providers or directly from the merchants themselves". The order is set, verbatim, by "factors like availability, price, quality, and whether they are the maker or primary seller of that item".
Look at them one at a time from your side of the counter:
- Availability. A tie at best. A large marketplace almost never shows stale stock; plenty of brand stores do.
- Price. Frequent disadvantage, and the most thankless fight. Returning to it over and over is the classic trap.
- Quality. Contested ground. This is where your listing, your reviews and the consistency of your data live.
- Maker or primary seller. Yours, and only yours. No reseller can claim it on your product.
Three of the four are won with operations, not size. And the fourth is not won at all: it is simply held. What most brands do is concentrate on the one where they have the least room to move — price — while neglecting the two they actually control.
Why you lose anyway: the signal never arrives
Being the manufacturer is worth nothing if the system cannot see it. And that is the real problem, because the data ChatGPT uses to build the merchant list comes, per its own documentation, "from third-party providers or directly from the merchants themselves". If you are not one of those direct sources, your maker status arrives — when it arrives — second-hand, through an aggregator that sees you as one merchant among many.
Two concrete routes follow, both documented. If your store runs on Shopify, your data is already integrated via Shopify Catalog and OpenAI states that "no additional work is required from individual merchants". Otherwise there is the direct product feed, built — again, verbatim — to "ensure ChatGPT always reflects the most up-to-date information about its products". How each door works is covered in our ChatGPT Shopping guide.
The second reason you lose is less technical and more uncomfortable: your data is worse. A marketplace updates stock and pricing by the minute because its entire business depends on it. A mid-sized brand store usually has stock right on its own site and wrong everywhere else. Since the data ChatGPT shows rarely comes from your site, that asymmetry gets paid in the merchant ranking, not in your admin panel.
The conversation happens before the carousel
There is a bigger asymmetry, and an easier one to miss. Everything above describes what happens once the user is already looking at your product and ChatGPT decides which merchant to send them to. But a great many purchases do not start there. They start one layer earlier, in a conversation where the user does not yet know which brand they want: "what's the best air fryer for two people?", "which trail running brand lasts longest?".
In that layer you are not competing for a slot in a merchant list: you are competing to be named. And there the marketplace is not your rival — Amazon is not an air fryer brand; your rivals are the other brands in your category. It is a completely different fight, won with reputation, mentions and citable content, and it is the one almost nobody measures because it shows up in no ecommerce report.
The strategic consequence orders everything else: if your brand is not named in the conversation, the merchant ranking will not save you, because the user never reaches your product listing. Optimizing the feed without working on brand presence is polishing the last metre of a race you never entered.
Does anything change on Google?
Less than the market wants to sell you. Google's documentation on AI features shuts the door on shortcuts in plain language: "You don't need to create new machine readable files, AI text files, or markup to appear in these features", and there is no special schema.org structured data to add. Google also says flatly that "there are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary".
What does appear on that same list of recommended practices, and is directly relevant to a store, is "making sure your Merchant Center and Business Profile information is up-to-date". Same pattern again: no secret lever, just data hygiene. The advantage a marketplace holds over you on both surfaces is identical, and it is boring.
The order of work that falls out of all this: 1) get named in the category conversation → 2) get clean, fresh product data into the sources → 3) make your maker or primary-seller status visible → 4) only then, price.
Most teams do it in exactly the reverse order.
How to measure this without fooling yourself
The most common measurement mistake is looking at traffic. It does not work: much of this conversation never ends in a click to you, and the clicks that do land usually arrive misattributed. GA4 undercounts AI traffic by design, so treating that number as your visibility is measuring the shadow instead of the object.
What you measure is the answer. The method fits in a spreadsheet:
- Write ten to twenty real buying questions from your category, in your market's language. Mix both levels: category questions ("best X for Y") and product questions ("where can I buy X").
- Ask them the same way every time, without personalising the session, on a fixed cadence — monthly is enough.
- Log three columns per question: does your brand appear?, does your store appear in the merchant list, and in which position?, who holds the top spots?
- Save the text of the answer, not just yes/no. The wording is where you see why you were picked or skipped.
Two honest caveats. First: answers vary between sessions, so a single run is not data, it is an anecdote; what counts is the repeated series over time. Second: OpenAI notes that the model interprets user intent and "can occasionally make mistakes", and that user context — memory, custom instructions — shapes what each person sees. There is no canonical answer you can audit. There is a trend, and the trend is measurable.
Where to start tomorrow
Pick your three most important products and ask both levels for each: the category question ("what's the best X for Y?") and the product question ("where can I buy [your product]?"). Fail the first, and your problem is brand — the feed can wait. Pass the first and fail the second, and your problem is data — now the feed is the lever.
It is a ten-minute diagnosis that orders months of work, and it separates the two investments most stores collapse into one.
Frequently asked questions
Do marketplaces have a structural advantage in AI answers?
Less than it looks. OpenAI documents that merchants are ranked "based on factors like availability, price, quality, and whether they are the maker or primary seller of that item". That last factor cuts in your favour and against the marketplace: on your own product, you are the maker or primary seller and a reseller is not. What marketplaces do have is better data hygiene — stock and pricing updated by the minute — and that feeds the other three factors. The advantage is operational, not structural.
How do I measure whether my store is gaining or losing ground to marketplaces?
Traffic will not tell you: most of these conversations never end in a click to you. What you measure is the answer itself. Take ten to twenty real buying questions from your category — "where can I buy X", "best X for Y", "cheap X" — ask them consistently, and log three things per question: whether your brand appears, whether your store appears in the merchant list and in which position, and who holds the top spots. Repeated monthly, that grid is your time series. The direction matters far more than any single number.
Do I need special markup for my store to appear in Google's AI features?
No. Google's documentation is blunt: "You don't need to create new machine readable files, AI text files, or markup to appear in these features" and "there's also no special schema.org structured data that you need to add". What it does explicitly recommend for merchants is "making sure your Merchant Center and Business Profile information is up-to-date". Again: data hygiene, not a trick.
If the marketplace sells my product cheaper, have I already lost?
Not necessarily, because price is one of four factors, not the only one. And there is a nuance worth knowing: OpenAI notes that the price in the initial response "typically reflect[s] the price from the first merchant listed, which may not be the lowest available price", and that it works to highlight the lowest priced option it knows about in the product detail page and merchant rankings. Competing on price alone against a marketplace is a war you rarely win; competing on availability, data quality and being the primary seller is a far more even fight.
Sources cited
- OpenAI Help Center — "Shopping with ChatGPT Search" (merchants ranked by availability, price, quality and whether they are the maker or primary seller; the list generated from third-party provider metadata or directly from merchants; the initial price reflects the first merchant listed; Shopify Catalog with no additional work; the model interprets intent and can make mistakes). Official documentation.
- Google Search Central — "AI features and your website" (no special files or markup needed; no additional requirements for AI Overviews or AI Mode; keep Merchant Center and Business Profile information up to date). Official documentation.
See who shows up in your category, and where you land
Every analysis asks ChatGPT and Gemini about your commercial category and captures the real answer: which brands they name, whether they name you, and which sources they cite to decide. It is the first level of this guide's diagnosis, run for you.