Why does ChatGPT recommend my competitor and not me?

The three things your competitor probably has that you do not

Third-party corroboration, review volume, and a consistent set of facts across all of it. None of those three live on your own website, which is why a better business with a better site loses to them.

1. Somewhere other than their own site says what they do. Directory listings, industry association pages, local news, supplier and partner pages, a Google Business Profile that is filled in. An engine writing an answer prefers a claim it can see in more than one place.
2. Reviews with text in them. Not just a star average, but written reviews that mention the service and the suburb, because that is the language a buyer question is phrased in.
3. The same name, address and phone number everywhere. Contradictory listings give a model a reason to skip you and name someone it can describe confidently.

Our own data points the same way without proving it. Of the 137 businesses with complete records from both engines, 25 were large anchor brands and 112 were not. Two of the 25 anchors were never named. Fifty four of the 112 non-anchors were never named. The businesses the rest of the web already writes about are the businesses the engines already name.

Why ranking above them on Google does not settle it

Because ranking is a position in a list and being named is a binary, and the things that earn each are only partly the same.

The clearest evidence we have is our own, and it argues against the easiest thing we could sell. Across the 300 Australian websites we scanned, the average on-site AI readiness score was 74.9 out of 100 while the businesses behind those sites averaged 38.3 out of 100 for AI visibility. Across the 137 with complete records from both engines, the correlation between site readiness and the mean of the two per-engine visibility scores is a Pearson r of 0.116. Readiness explains almost nothing about who gets recommended.

The sharpest single number in that set: 34 of those 137 businesses scored a perfect 100 for site readiness, and nine of them were still never named by either engine. A flawless site is not a defence.

How to find out which sources the answer actually used

Ask the engine to show them, then open them. When ChatGPT answers with a live web search it carries citations, and those links are the pages it read to write the answer about your category.

Run the buyer question, then follow it with which sources did you use for that answer. Read the pages that come back, and notice what kind of page they are. In most categories they are not competitor homepages. They are directories, comparison articles, review platforms, local media and association listings, which is to say pages your competitor appears on and you do not.

Two cautions. An answer written from the model's memory rather than from a search will not have citations, and asking for them anyway can produce a plausible looking list that was never fetched. And a cited page is evidence the engine read it, not proof that it was decisive. Treat the list as a map of where the conversation about your category happens.

A comparison you can run yourself in ten minutes

Score the two of you on the same five questions, then compare the footprint underneath the result.

1. Write five buyer questions for your category and city, none of which contains either business name.
2. Ask all five in ChatGPT, then all five in Gemini. Record who was named in each of the ten answers.
3. Count. Ten answers, two tallies. That is your named rate against theirs on the day.
4. Now search each business name on its own and list where each one appears: directories, review platforms, press, association pages, supplier listings.
5. Compare the two lists side by side. The gap in that list is usually much bigger than the gap in the websites, and it is the one that explains the result.

The instructions for the measurement half of that, in more detail, are in how to check what ChatGPT says about my business.

What to fix first, in order of how often it is the cause

Work from the cheapest and most mechanical outward, because the mechanical faults are the ones that make everything else pointless.

1. Check you are not blocking the crawlers. Thirteen per cent of the 300 sites we scanned block at least one of the six AI crawlers we check in robots.txt. Those 39 sites averaged 21.3 out of 100 for AI visibility against 40.9 for the 261 that allowed them. See should I block AI crawlers.
2. State your facts plainly on your own pages. What you do, where, for whom, in text a machine can read. Twenty four per cent of the 300 sites had no JSON-LD structured data at all.
3. Earn the corroboration. Listings, reviews with text, industry pages, local press. This is the slowest item and the one that most often decides it.
4. Check both engines before you conclude anything. Of the 137 with complete records, 26 were named by Gemini alone and 10 by ChatGPT alone.

The full ranked version of this list, with what we found for each cause, is on why doesn't ChatGPT mention my business.

Where to start

You can run the ten minute comparison above yourself and never speak to us. If you would rather see it done across both engines with every answer stored, the scan is free and you keep the report.