The One-Question Customer Interview That Writes Your Ads for You

Ninety-five percent of buying decisions are made by the part of the brain you're not writing to. The old, fast, emotional part that records everything and reasons about almost none of it. Meanwhile most of us sit down to write ads using the slow, conscious part, guessing at angles and offers and headlines. We're aiming our best work at the 5%.

There's a fix, and it's almost insultingly simple. You get on a call with one customer and you ask them one question. Not seventeen. One. Then you shut up and let them talk for an hour.

I'll walk through exactly how to run it, what to ask, and how to turn an hour of someone rambling about their life into the hooks, headlines, and objections you'll be writing against for the next quarter.

Why one question beats a survey

The instinct, when you finally get a customer on the phone, is to mine them for everything. What did you think of the product? How was checkout? Would you recommend us? Resist all of it.

Here's the problem with a long question list. The moment you ask "what did you think of the product," you've dragged the customer into the conscious, rational part of their brain. They start performing. They give you the tidy, sensible answer they think you want. And the tidy answer is useless, because it's not why they actually bought.

What you actually need is the story. The messy, real account of how this problem turned up in their life and what they tried before they landed on you. That's where the emotion lives, and emotion is what you're really selling to.

So you ask one open question, and you let it run.

1. The question itself

Word it roughly like this:

"Take me back to before you found us. What was going on that made you start looking for something like this in the first place?"

Then nothing. You don't prompt. You don't steer. You let them ramble about their life, their kids, their week, whatever spills out. It will feel unproductive for the first ten minutes. Sit in the discomfort, because that's exactly when the good stuff arrives.

The reason this works: people don't experience their problem in the neat box marketers put it in. Someone struggling with their weight doesn't just feel bad in their jeans. They feel it buckling the seatbelt, sitting in a coffee-shop chair, walking next to their partner. The problem leaks into a dozen corners of their life. Your survey question only ever catches the one corner you thought to ask about. The open story catches all of them, in their words.

2. How to actually run the call

A few mechanics that make the difference between a useful hour and a wasted one.

• Record it. With permission, obviously. You cannot mine what you didn't capture, and you'll want the exact wording later, not your paraphrase of it.
• Talk less than you think. Aim to speak for under 10% of the call. Every time you fill a silence, you steal a sentence you needed.
• Chase feeling, not features. When they mention a moment, ask "how did that feel?" rather than "what happened next?". You're after the emotional texture, not a timeline.
• Go for the corners. If they say "I just felt a bit self-conscious," gently ask where else that showed up. Map every place in their life the problem touched. That map becomes your angle list.

One genuinely good call is worth more than a hundred survey responses, because a survey gives you ticked boxes and a call gives you the actual language.

3. Mining the transcript for hooks

Now the part where the ads write themselves.

Read back the transcript and pull every line where the customer describes the moment the problem bit. Not the solution, the friction. "I was so sick of repurchasing the cheap one every few months." "I'd basically given up on finding one that didn't irritate my skin." Those raw, specific moments are your hooks. They stop the scroll because the right person reads them and thinks, that's me, that's my exact Tuesday.

The trick is to use their words, not yours. We worked with a brand, I'll keep them anonymous, selling something most of America didn't even have a name for. They'd been writing ads assuming everyone knew what the product was. Their customers, in interviews, described it in plain, roundabout terms, the layman's phrase rather than the category name. We switched the ads to lead with the customer's wording and the same product started landing almost overnight. Nothing changed but the language, and the language came straight off the calls.

4. Mining it for headlines and proof

Same transcript, different pass. This time you're hunting for one-liners.

People hand you headlines without realising it. A throwaway sentence like "honestly it's the only one I've bought twice" is a better headline than anything you'll brainstorm in a meeting, because it carries proof and emotion in one breath. Highlight every line that sounds like something a friend would say over coffee. Those go straight onto statics, near enough verbatim.

Look especially for the single nugget. Usually one detail matters far more to the customer than all the others, and it's rarely the one the brand is proud of. The founder wants to talk about the manufacturing process. The customer keeps mentioning that it finally stopped the thing that was annoying them for years. That one nugget, the piece of information that does the emotional heavy lifting, is what your best headline is built on. The interview tells you which one it is, because the customer keeps circling back to it without being asked.

5. Mining it for objections

Third pass, and this is the one most people skip. Read the transcript for hesitation.

Listen for the "I almost didn't" moments. "I nearly didn't buy because I assumed it'd be like the others." "I was worried it wouldn't actually work for my situation." Every one of those is an objection your future customers are silently having too, and now you've got it in plain words. You build the answer straight into the creative. The objection becomes a hook of its own: "Thought this was just another one of those? So did she."

By the time you've done three passes, one hour-long call has handed you a stack of hooks, a set of proof-led headlines, and a list of objections to pre-empt. That's a quarter of creative angles out of a single conversation.

6. The AI step that scales it

Here's where it gets quick. You don't read these transcripts line by line forever.

Record the call, run it through any transcription tool, and you've got the raw text. Then hand that text to a chat model and ask it to do the passes for you. I'd prompt it plainly: pull every sentence where the customer describes the moment the problem started, then every line that reads like a testimonial, then every hesitation or doubt. Ask for them as direct quotes, the customer's exact words, not a summary.

One habit that lifts the output a lot: before it answers, tell it to ask you what else it needs to know to do the job well. The model doesn't know your product or your customer until you tell it, and that one instruction usually surfaces a question or two you'd have forgotten to answer.

Run that across five interviews and you've got a verbatim swipe file of one-liners, ready to drop onto statics, sorted by whether they're a hook, a headline, or an objection. The research that used to feel like a chore becomes an afternoon.

Try this before you write another ad

Before you brief your next round of creative, book one call with one happy customer and ask them the single question. Record it. Let them ramble. Then run the three passes, by hand or with a model, and see how much of your next month of ads is already sitting in that transcript.

I reckon you'll be surprised how little of it you needed to invent. If you run the exercise and something genuinely surprises you, I'd love to hear what it was.