Direct answer
Don’t check your dashboards — they only count buyers who already found you. Instead, become the stranger: open ChatGPT, Gemini, Perplexity and Google, and ask for a supplier in your category the way a buyer who’s never heard of you would. Read the answer. If it names competitors and not you, or describes you vaguely, that’s a discoverability gap. Run the same queries monthly; it’s the one number your own analytics can’t show you.
Most businesses assume their analytics would tell them if they had a discovery problem. They wouldn’t. Your dashboards are built from the buyers who found you; the ones who searched the category and never saw you leave no trace. To check discoverability you have to look where your reporting can’t — at the answer a stranger gets — and that’s a test you run by hand.
Step one: ask the way a buyer does, not the way you do
The single biggest mistake is searching your own name. A buyer who already knows you isn’t the one you’re worried about. Search the category and the problem instead: “who does [what you do] for [your kind of client] in [your region],” the way someone who’s never heard of you would phrase it. If you only appear when your name is in the query, you’re measuring recognition, not discoverability — a distinction that decides why qualified buyers never discover you in the first place.
Searching your own name tells you nothing. The buyers you’re losing never type it.
Step two: test the engines your buyer actually uses
Run the same unbranded queries across the tools a buyer now reaches for: ChatGPT, Gemini, Perplexity, and an ordinary Google search — and read the AI answer at the top, not just the blue links under it. Each engine assembles its answer differently, so being named by one is not being named by all. The point is coverage: you want to know whether you appear in the answer, in the links, or nowhere.
| What you see | What it means | What to do |
|---|---|---|
| Named in the AI answer | You’re citable for that query | Hold it; check the next query |
| In the links, not the answer | You rank but aren’t quoted | Work on trust and extractability |
| Described vaguely or wrongly | Identity is unclear to the engine | Fix identity and consistency |
| Absent entirely | A discoverability gap | Diagnose which signal is missing |
Step three: read the failure, don’t just note it
Absence isn’t one problem, it’s several, and the shape of the failure tells you which. Named nowhere is different from ranked-but-unquoted, which is different from described-wrongly. Each points at a different missing signal — identity, corroboration, or extractability — and the fix differs accordingly. This is also how you separate a discoverability gap from an awareness one: awareness is whether they remember you; this test is whether the machine can find you when they don’t.
Why the manual test matters: a large and growing share of searches now end without a click — various analyses of clickstream data (SparkToro / Similarweb) have put zero-click searches somewhere in the region of 58–70% depending on method and period. When the answer is read and no link is followed, being absent from that answer is the whole loss. Treat the figure as directional, not precise.
Step four: make it a monthly number
One check is a snapshot; the value is in the trend. Keep a short list of your core buyer queries and run them every month, noting for each whether you were named, linked, or absent. That gives you the one discoverability metric your own analytics will never surface — and if the answers are already naming competitors, the cost of that is the subject of what declining discoverability costs before revenue drops.
- Search the category and the problem, never your own name.
- Run the same queries across ChatGPT, Gemini, Perplexity and Google — and read the AI answer, not just the links.
- Record each result as named, linked, or absent — the shape of the failure names the missing signal.
- Repeat monthly and watch the trend, not the single snapshot.
If you’ve run the queries and you’re not in the answers — or you’d rather someone did it properly and told you which signal is missing:
Trust a dashboard built only from the buyers who found you, and the ones searching the category right now will keep choosing a competitor you never saw them consider.
Book a free 15-minute session See the proof