This is where we publish the working notes: what we measure, how we measure it, and what the numbers actually said — including when they said something inconvenient. Two rules...
This is where we publish the working notes: what we measure, how we measure it, and what the numbers actually said — including when they said something inconvenient.
Two rules apply to everything on this page. First, any number we quote is either measured by our own tooling, with the method described, or attributed to a named third party. Second, engine coverage is always stated. A result from one model is not a result from every model, and we will not round that corner.
The category is eighteen months old and already full of confident claims with no method attached. You will see agencies quoting citation rates without saying which engine produced them, on what prompt set, or on what date. None of those numbers are reproducible, which means none of them are evidence.
Our own baseline is a fair example of why this matters. When we ran our category prompt set against a generative engine in September 2026, our own brand scored a Mention Rate of 0% and a Citation Rate of 0%. The engine named three Vietnamese agencies instead. We publish that because a measurement method you only use when the answer flatters you is not a measurement method.
| Finding | What it implies for your content plan |
|---|---|
| Across 37 category prompts, only 6 produced any vendor name at all — and every one of those 6 was a bottom-of-funnel prompt. | Definitional and educational content (“what is X”) does almost nothing for AI visibility. Buying-stage content does the work. |
| Where vendors were named, the model was drawing on third-party sources, not the vendors’ own sites. | Being mentioned elsewhere matters more than describing yourself well on your own domain. |
| Mention and citation move independently. A brand can be named repeatedly without its domain ever being linked. | Track both. A single blended “AI visibility score” hides which of the two is broken. |
Every article states its scope. If a piece is based on one engine, it says so in the first section rather than in a footnote. If a claim is an opinion rather than a measurement, it is written as one. If we later measure something that contradicts an earlier article, we update the article and note the change rather than quietly deleting it.
Measurement articles are re-run on the same prompt set roughly every 90 days, because comparability is the entire point. Opinion pieces are updated when the underlying platform behaviour changes.
Yes. The scoring definitions are published in full precisely so that you can run them yourself, or hand them to another agency and check their work.
Our first published baseline scored our own brand at zero on both Mention Rate and Citation Rate. So yes.
Want your own numbers instead of ours?
Send us your domain. We run the baseline on your category prompts and send back the raw answers, not just a score.
Đội ngũ chuyên gia Vidco Group sẵn sàng đồng hành cùng bạn
Bước 1 / 4
Chúng tôi sẽ liên hệ trong vòng 2 giờ làm việc.