.png)
You scheduled 6 customer interviews this month. Three people showed up. Two said your product was "great." The third asked for a feature nobody else needs.
Two weeks of chasing. Forty-five minutes of recordings you'll never transcribe. Zero decisions made.
Now add it up. Recruiting. Scheduling. Running interviews. Taking notes. Synthesis. A standard round of 8 interviews burns 15–16 hours of team time. Sometimes two full workdays. Gone.
And you still don't know why your last launch landed flat.
The problem isn't interviews. It's the way most teams run them. Slow. Expensive. Built to collect polite lies instead of real answers.
There's a smarter way, and you can start using it this week.
Why the Standard Interview Approach Is Expensive, Slow, and Often Useless
Customer interviews can work extremely well.
The problem is that the way most teams run them creates three big structural issues.
The Recruitment. For every hour of actual interview time, most founders spend twice as much just getting people to show up. Cold outreach gets ignored. Scheduling goes back and forth. You burn your most limited resource — time — before the interview even starts.
The Politeness Problem. Customers don't want to hurt your feelings. Unless you know how to ask the right questions, so they say, "I love it" and "I'd definitely use that." And of course they never do. This is called the Polite Lie Problem. It makes interviews feel productive while hiding the truth. You can't build a product on polite feedback.
The Synthesis Gap. Even when you do get useful answers, what happens next? A folder of recordings. Scattered notes. No clear decision. Y Combinator flagged this as one of the biggest unsolved problems in early-stage product development. Data without a process is just noise.
These aren’t signs that you’re bad at interviews. They’re signs the traditional workflow wasn’t built for small, fast-moving teams.
What You Might Be Doing Wrong Today
If customer interviews feel slow, expensive, or not useful, the problem is often not the interview itself. It is the setup.
You may be doing one or more of these things right now
- Starting interviews too early without reading support tickets, reviews, Reddit threads, or sales calls first
- Going in with broad goals like “learn about the customer” instead of one clear decision to support
- Asking opinion questions (“Would you use this?”) instead of behavior questions (“What did you do last time?”)
- Trying to learn everything in one interview instead of focusing on one topic at a time
- Using a rigid script and missing the important follow-up moments
- Talking to the wrong people (too general, not your real buyer/user type)
- Collecting notes but not synthesizing patterns into actions, decisions, or next tests
Remember, opinion questions get you polite lies. Behavior questions get you the truth.
Instead of: "Would you use a tool that automated your reporting?"
Ask: "Walk me through the last time you put together a report. What did you actually do?"
The first question asks people to predict their future behavior. They'll say yes out of politeness. The second asks them to describe something that already happened.
Research Mine Communities First (Pre-Interview)
Before you book a single interview, spend 90 minutes reading what your customers are already saying in public, or run surveys before spending time and money on traditional interviews.
Where to look: Reddit (search your product category + "frustrating", "I wish", "why doesn't"), G2, and Capterra reviews. These are real people venting real frustrations — without a single polite lie.
What to extract: Copy any sentence that includes a complaint, a workaround, or a wish. Group them into 3–5 themes. These themes become your interview guide.
I spent around 90 minutes on Reddit to write this article and learned more about why people want to read this article than I did from asking some of the founders in my network.
This step alone can cut your prep time significantly. You walk into every interview already knowing what to listen for, rather than fishing blind.
What 10x Cheaper Customer Interview Actually Means
Let's do some back-of-the-envelope calculation. Here's what a standard round of 8 customer interviews actually costs a small team:
- Recruiting and outreach: 4–6 hours
- Scheduling coordination: 2–3 hours
- Running the interviews: 4 hours
- Taking notes and transcribing: 3–4 hours
- Trying to make sense of the data: 2–3 hours
- Participant incentives (if used): $200–$400
Roughly 15–16 hours of team time, plus potential spend. For a founder or head of growth wearing five hats, that's two full workdays gone — with uncertain return.
The lean version, which I will recommend below, cuts that to under 3 hours:
- Community research (pre-work): 1.5 hours
- AI-led interviews (run asynchronously): overnight
- Review: 30–45 minutes
The Lean Interview Process at a Glance
A Simple Synthesis Method That Takes 20 Minutes
After each interview, write down three things before you close your notes:
- The sharpest thing they said (a direct quote)
- The behavior that surprised you
- The one assumption it confirms or kills
After 5 interviews, look at those 15 data points together. Patterns will be obvious. You won't need a spreadsheet or a framework — just the discipline to do it immediately, not later.
Use Frank To Run Focused Interviews
Most interviews fail before they start. Wrong questions. Wrong people. No follow-up when the answer gets interesting.
Frank fixes the process, not just the cost.
You set a research brief. Frank handles everything else. It asks the right questions. It probes when an answer needs more. It doesn't let the conversation stay shallow.
- No scheduling. No no-shows. No polite small talk is eating your time.
- While you sleep, Frank is running interviews. By morning it returns enterprise-quality customer insights at 1/10th the cost. You have patterns pulled from actual conversations, not guesswork.
- The insights link back to the raw transcripts. You can verify everything. Nothing is a black box.
This is how interviews get shorter, sharper, and actually useful. Not by cutting corners. By removing everything that was wasting your time in the first place.
Who Should Be Running These?
If you're the founder or the product owner, I believe you should be in the first 10 interviews yourself. The language customers use will shape how you talk about your product.
If you have a PM or a head of growth, they should own the ongoing process after that. The system above is lightweight enough for one person to run without a dedicated researcher.
If you're using Frank, the answer changes entirely. You don't need to decide who runs it. Or what customers will answer, as people tend to be more candid with AI.
Anyone on the team can set the brief. Frank will handle the rest.
The Bottom Line
Interviews aren't the problem. The process is.
Two weeks of chasing. Polite answers that go nowhere. Recordings nobody transcribes. Decisions that never get made.
It doesn't have to work that way.
Start this week: spend 90 minutes mining Reddit and G2 for your product category. Work with Frank to set up and run AI interviews. Learn from your segment, hear what they're telling you, and add insights into your roadmap
