How to Validate Your ICP With Customer Interviews
Structured interviews are how you tell a real painkiller apart from an idea that only sounds good in the room.
Why interviews beat surveys here
Surveys capture opinions. Interviews, done right, surface behavior — what someone has actually tried, paid for, or built to solve the problem. Behavior is the signal; opinion is noise.
Who to talk to, and how many
10 to 15 structured conversations is usually enough to see a pattern emerge. Fewer than that and you're reacting to one loud voice; more than that, without changing the questions, is diminishing returns before you've acted on what you already know.
The core questions that actually work
"Walk me through the last time this problem came up." — behavior, not hypothetical
"What have you already tried to fix it?"
"What are you currently paying for, or have you built internally, to deal with it?"
"What would have to change for you to switch away from that?"
"How often does this come up — daily, weekly, once a quarter?"
The one clue that you're onto something
Someone who's already spent money or built a workaround is telling you this is a painkiller, not a vitamin — regardless of how enthusiastic or lukewarm they sound about your specific pitch. Weight what people have already done far above what they say they'd do.
What to do with what you hear
Go back to your ICP template and update the problem profile and disqualifiers with real language and patterns from the interviews — not just people's stated intent to fix it. Feed the traits that show up in your best answers into
your scoring model too.
This fits into the bigger process described in our full ICP guide.
Comments