Customer Development: How to Interview Your Customers

Most campaigns that flop do not fail because the media buying was bad. They fail because the messaging was built on a guess about what the customer wants, and the guess was wrong. I have watched teams burn a budget promoting a benefit nobody cared about, then blame the ad platform. The platform was fine. The research was missing.

Customer development is the fix. It is the habit of learning about your audience through real conversations instead of assumptions. In this guide I will walk you through how I run customer interviews: who to talk to, how to ask questions that do not lead people into telling you what you want to hear, how many conversations you need, and how to turn it all into personas and copy you can run.

Why assumptions kill campaigns

When you sit down to write an ad, you have to decide what to say: which benefit leads, what pain point you name, what words you use. If you have never talked to a customer, every one of those decisions is a guess dressed up as a strategy. Sometimes the guess is close. Often it is not, and you only find out after you have spent real money.

Customer development flips the order. Steve Blank, who coined the term, put it simply: get out of the building and learn from the people you want to serve. You are not trying to confirm that your idea is good. You are trying to understand the person well enough that the right message becomes obvious. The interviews are cheap and the wrong campaign is expensive.

This is a research method, not a document. Your ideal customer profile is the output of the research, and audience segmentation is how you group people afterward. Customer development is the part where you go find out what is true.

Who to talk to and how to recruit them

Start with people who have the problem you think you solve, whether or not they have heard of you. If you already have customers, talk to recent buyers while the decision is fresh, and talk to people who churned, because they will tell you what fell short. If you are pre-launch, talk to people who currently solve the problem some other way, even a clumsy manual way.

Recruiting in a US context is easier than people expect. Post in the subreddits and Facebook groups where your audience hangs out, ask your email list, and use LinkedIn for B2B. Offer a small incentive: a 25 dollar Amazon or Starbucks gift card for 30 minutes is standard. If you want volume without your own network, panels like User Interviews or Respondent can source and schedule people for you.

One warning I will repeat because it matters: do not only interview your friends and your team. They love you, they know your product, and they will tell you what you want to hear. That is the opposite of useful.

The interview mindset and the Mom Test

The single biggest mistake in customer interviews is treating them like a pitch. You get excited, describe your idea, ask if they would use it, and they say yes to be polite. You walk out feeling validated and you learned nothing. Your job in the room is to learn, not to sell, so talk about 20 percent of the time and listen for 80.

The best framework here is Rob Fitzpatrick's Mom Test: ask questions so grounded in reality that even your mom could not give you a false yes. You do that by asking about their past behavior and real problems, never about hypotheticals or their opinion of your idea. "Would you buy this?" is a bad question because the answer is a guess about the future. "Walk me through the last time you dealt with this" is a good one because it is a fact about the past.

Keep your questions open, not leading. A leading question hands the person the answer: "Wouldn't it be great if this were faster?" Of course they will nod. An open question makes them do the work: "What is the most frustrating part of how you do this today?" Then you stay quiet and let the silence pull out the real answer.

A sample question set that works

Here is a starter set I adapt for almost any project. Notice that every question points at what already happened, not at what someone imagines they might do.

• Tell me about the last time you ran into this problem. Get the story with real detail.
• What were you trying to get done? This surfaces the job, not just the pain.
• How are you solving it right now? The current workaround is your real competition.
• What is the hardest or most annoying part of that? This is where messaging lives.
• What have you tried before, and why did you stop? Learn what already failed them.
• How much time or money does this cost you today? Gauge whether the problem is worth paying to fix.
• Who else is involved when you decide to fix it? Find the real buyers and influencers.
• Where did you look for a solution last time? This tells you which channels to buy.

End with an open door: "Is there anything I should have asked but didn't?" People often save the most honest thing for the end.

How many interviews, and turning notes into messaging

You do not need a hundred interviews. For a single audience segment, patterns usually start repeating around the fifth conversation, and by 8 to 10 you hear the same phrases and frustrations again and again. That repetition is the signal to stop. If you serve two very different segments, run 5 to 10 for each rather than blending them.

Record with permission or take close notes, then look for patterns across people, not clever quotes from one person. Group the recurring problems and the exact language people used. Real customer words are gold for ad copy, so capture them verbatim.

From there the outputs write themselves. The recurring people and their goals become your personas. The recurring "what I was trying to get done" answers become your jobs-to-be-done, the progress the customer is hiring your product to make. The recurring frustrations plus their own phrasing become your messaging: the headline, the lead benefit, the objection you handle first. When your ad quotes the customer back to themselves, conversion tends to follow.