ChatGPT brand tracking is the measurement layer that reports how often ChatGPT names your brand when buyers ask about your category. Asking once tells you almost nothing, because the same question returns a different line-up on the next run; this samples enough times to produce a rate you can defend.
An engine named after a surface it did not measure is a fabricated result with extra steps, so here is the precise scope before anything else. For ChatGPT, we measure OpenAI's model with web search enabled — the same retrieve-then-answer behaviour ChatGPT uses when it browses.
| What it means here | |
|---|---|
| Direct evidence | The answers come from OpenAI's own model using its own web-search tool. That is direct evidence for this model's grounded behaviour. |
| Not claimed | This is the API surface, not a scrape of chatgpt.com. It is the closest measurable proxy for what a browsing ChatGPT session returns, and it is labelled that way rather than implying we sat in the consumer app. |
| How a run works | We call OpenAI's API with web search turned on, using your category questions, repeatedly. Every raw answer is stored, so any number on the dashboard can be traced back to the exact text it came from. |
ChatGPT tracking is also known as ChatGPT mention monitoring or ChatGPT rank tracking. The same direct-versus-proxy distinction is applied to every engine we support, and it is written into the published methodology rather than left to a sales conversation.
Rates, not screenshots — each with a 95% confidence interval and the number of answers behind it, so you can tell a real change from a noisy week.
| Metric | The question it answers |
|---|---|
| Mention rate | Of the answers to your category's questions, how many name you at all |
| Share of voice | Of every brand named across those answers, what fraction is you |
| Recommendation rate | How often you are actively endorsed rather than listed in passing |
| Cited sources | Which domains ChatGPT leaned on — yours, and everyone else's |
| Accuracy flags | Anything ChatGPT said about you that is factually wrong |
One reading is not a trend. Runs repeat on a schedule so the comparison is against your own history, not against a screenshot someone took once.
Close, but we do not claim it is identical, and no honest tool can. We measure OpenAI's model with its web-search tool through the API. The consumer app adds its own product layer over the same underlying behaviour. We say which one we measured so you can weigh the result yourself.
You can, and it is a reasonable first look. What it cannot give you is a rate. One answer is a single draw from a distribution, so it cannot tell you whether anything changed since last month, which is usually the question that matters.
Yes. Being listed in passing and being actively endorsed are separate metrics, scored by a calibrated judge against a written rubric rather than by keyword matching.
The editorial companion to this page — how ChatGPT chooses what to say, and what earns a mention.
ChatGPT is one surface. The full tracking layer runs your suite across all of them and compares.
The sampling, the calibrated judge, the statistics — and the six-point bar to hold any tool to.
Run your category's real buyer questions and get a rate with an interval — not a screenshot.