Customer discovery interviews: What to ask, and the probes that get a real answer

Customer discovery interviews work on one rule: Ask what happened rather than what someone would do. Reconstruct the last real purchase in order, from the event that started it to the moment they decided, and use 6 follow-up probes to get past the first abstract answer. The probes matter more than the question list.

Most people can write a decent question list in 10 minutes. The difference between a useful interview and a wasted 30 minutes sits in what you do after the buyer gives their first answer, which is almost always too tidy to be true.

Key takeaways

• Past behaviour is reliable. Predictions and preferences are not.
• A 30-minute call has 5 phases, and the last 5 minutes are the highest value.
• 6 follow-up probes do most of the work, and you can learn all 6 in an afternoon.
• 6 mistakes ruin interviews, and pitching is only the most obvious one.
• The output is coded language and verbatim quotes, held somewhere the next person can find them.

Ask about past buying behaviour, not hypothetical future decisions

This one rule removes most bad questions. People are unreliable narrators of their own future behaviour and precise about what they actually did.

The reliable questions are the ones about events: What did you do, when did you do it, who else was involved, and what happened next. The unreliable ones invite a prediction or an opinion, and they sound like this: Would you use this, how important is that, what features would you like, how much would you pay.

Every question in a good guide points at a specific past event. If a question could be answered by someone who has never bought anything, it belongs in a survey rather than an interview.

A 30-minute customer discovery interview: 5 phases from trigger to decision

Keep the shape the same across every call so the answers compare.

1. Set-up, 2 minutes. Say you are not selling, the notes stay internal, and you want to understand how a decision actually got made. Ask permission to record.

2. The trigger, 5 minutes. Take me back to when this first became a problem you decided to do something about. What was happening that week.

3. The path, 10 minutes. Walk forward in order. What did you try first, who did you ask, what did you search for, what did you rule out and why.

4. The decision, 8 minutes. Who else was involved, what nearly stopped it, how did you describe it internally to get it approved.

5. The close, 5 minutes. Covered below, and worth protecting even when the call is running late.

In the last 5 minutes, ask these 3:

• If this had not existed, what would you have done instead.
• What would you tell someone in your position who is at the start of this.
• Is there anything I should have asked and did not.

The third one regularly produces the most useful line in the whole transcript.

The 6 follow-up probes that turn abstract answers into specific evidence

Buyers answer in abstractions first. These 6 move them to specifics, and you can use each one several times per call.

1. What do you mean by that. For every abstraction: Efficiency, integration, support, ease of use. Their definition is the finding.

2. Tell me about the last time that happened. Turns a general complaint into a dated event with details attached.

3. And then what happened. Keeps the timeline moving without you supplying the next step.

4. Who else was in that conversation. Surfaces the people who shaped the decision and never spoke to sales.

5. What made that a problem worth solving now. Separates a genuine trigger from an ongoing irritation.

6. Silence. Wait 4 or 5 seconds after an answer that sounds finished. The second half of the answer is usually better than the first.

Number 1 and number 6 are the two most people skip, and they carry the most.

6 customer discovery interview mistakes that weaken the evidence

Pitching is the most obvious one. The moment you explain or defend the product, the buyer switches to being polite and the data stops. Leading sits close behind, because naming your own claim invites agreement. Ask how they describe it and stay quiet.

Accepting abstractions is the quiet mistake. Writing down easy to use as a finding is useless when it means 4 different things to 4 buyers. Taking price at face value is the same error in a different coat, since price is the socially easy answer and needs probing with what would have had to be true.

The last 2 are structural. Interviewing the wrong person gets you half the decision, because the user and the person who approved the budget experienced different things and you want both. And not recording costs you the deliverable itself: Notes capture your interpretation, while the exact phrasing is what you came for.

If nobody senior is free to run these properly, that is a resourcing problem rather than a research one, and who actually does marketing for mid-market portfolio companies sets out the options.

How to analyse 15 customer discovery interviews and turn transcripts into usable messaging

The analysis takes less time than the calls if you do it in this order.

• Pull the verbatims first. Every phrase describing the problem in the buyer's own words, in a single list. This becomes headline and page copy.
• Code the triggers. Group the events that started each search. Usually 3 or 4 patterns cover most of them.
• List every alternative mentioned. Including internal builds and doing nothing. This is your real competitive set.
• Rank the decision criteria by how often they decided it, rather than by how often they came up.
• Write the answer down and keep the transcripts. The conclusion goes into the portfolio company positioning play, and the quotes go against the customer record so the next person inherits them. Getting the record into that shape is covered in GTM modernization after an acquisition.

After a deal closes, the first round belongs in the opening weeks, which the 100-day marketing plan for a newly acquired company sequences alongside the rest of the work.

What the probes changed in a 60-person Nordic software company

Take a newly acquired Nordic software company we work with: Around 60 people, 600+ customers, two marketers, no CMO. The internal belief was that customers bought on ease of use.

The probes changed all of it. Asked what do you mean by easy, buyers described confidence that a switch would not disrupt operations, which is a different claim entirely. Asked about the last time it was a problem, most named a specific operational event rather than an ongoing frustration. Asked who else was in the conversation, a technical reviewer appeared in most accounts who had never spoken to sales. And asked what they would have done otherwise, the most common answer was building something internally.

The first answer in every one of those 4 cases was the tidy one. The probe produced the useful one.

FAQ

Start with 3 recent customers and practise the 2 most commonly missed probes

Book 3 calls with customers who bought in the last 6 months. Keep the same 5-phase shape, and make yourself use what do you mean by that and 5 seconds of silence in every one. Record them, pull the verbatims, and you will have more usable material than a year of internal debate. No random acts of marketing.

If the findings suggest the message is pointed at the wrong buyer, the fix sits under the value-creation programme for portfolio companies, and smaller pieces of research sit in the add-ons that stack onto an active module.

When you are ready to run a round for a company you have just bought, book a strategy session. You will leave with a bird's-eye plan, whether or not we work together.

Book a strategy session →