In short
- A buying question is what someone asks AI right before they choose a supplier. It names a need, often a place, sometimes a budget.
- You don't have to guess them. Your inbox, your phone calls and your quotes already contain most of them.
- Fifty is enough to see a pattern. Measure them every day and you know where AI leaves you out.
What a buying question is
Most people don't ask AI "accountant". They ask a full question, the way they would ask a friend: "Which accountant in Utrecht is good for freelancers and doesn't cost a fortune?" That question has a need, a place and a condition. It's the moment just before someone picks up the phone.
We call that a buying question. It's different from a keyword in two ways. It's longer, so it says much more about what the person wants. And the answer isn't a list of ten links, but a short text with two or three names. If you're not one of them, you never get the call.
Four kinds of buying questions
Almost every buying question we measure falls into one of four groups. A good list has some of each.
The searcher wants names. Place and service matter most.
The searcher wants a number. Engines quote whoever gives one.
The searcher is still choosing. Clear explanations win here.
The searcher has a condition. Specialists get named, generalists don't.
Find 50 in one afternoon
You don't need a keyword tool for this. You need the places where customers already talk to you.
- 1Open your inbox and search for the first emails from new customers. Copy every question they asked before they hired you.
- 2Ask whoever answers the phone which three questions they hear most. Write them down word for word.
- 3Look at your last ten quotes. What did the customer need, where, and what was the deciding factor? Turn each into a question.
- 4Type your service and town into ChatGPT and read the follow-up questions it suggests. Keep the ones a real customer would ask.
- 5Add the place to every question that needs one. "Good dentist" becomes "Good dentist in Amersfoort for anxious patients".
You'll end up with more than 50. That's fine. Remove duplicates and keep the questions where a sale is close.
How to write them well
- Write like a customer, not like a marketer. "Who fixes a leaking roof fast in Zwolle?", not "roofing services Zwolle".
- One need per question. If a question asks two things, split it.
- Keep the conditions that make you different. If you're the one who works weekends, put "on Saturday" in a question.
- Mix big and small. Some broad questions to see the market, many specific ones where you can actually win.
What to do with the list
Ask each question to ChatGPT, Gemini and Perplexity, and note who gets named. Do it again next week. You'll quickly see three groups: questions where you're always named, questions where you're sometimes named, and questions where you never show up.
The middle group is where the quick wins are. You're close, and usually one missing page, price or answer makes the difference. That's the group we start with for every new customer.
Questions we get
How many buying questions do I need?
Fifty gives a clear picture for most small businesses. With fewer than twenty, one lucky answer changes your score too much. With hundreds, you lose sight of what matters.
Should I write them in English or Dutch?
In the language your customers use. If they ask AI in Dutch, measure in Dutch. Answers can be quite different per language.
Do the questions change over time?
A little. Add new ones when you launch a service or hear a new question on the phone. Keep the core list stable so you can compare week to week.
Oleksii
Danny van der Ham