A purpose-first playbook for running interviews that surface real pain, not the answers founders wish they were hearing.
How this guide works. This page argues for a specific interview structure: 45 minutes, six segments, a fixed set of question types, and a coding protocol for what to do afterward. We borrow heavily from Rob Fitzpatrick’s The Mom Test and Steve Blank’s customer-discovery work, but adapt for solo founders who can’t spend a quarter on discovery the way a venture-backed team can. How we research.
The most common mistake we see solo founders make in customer discovery is not skipping interviews entirely. It’s running them, getting positive-sounding responses, and concluding they have validation when they don’t. The pattern is consistent: founder describes the product, prospect says “that sounds interesting,” founder writes “validated” in their notes, and ships. Six months later, the prospects who said “interesting” are not paying customers.
Rob Fitzpatrick’s The Mom Test — published in 2013 and still the most cited book on this topic among solo founders — describes the failure mode in one sentence: “If you ask someone whether they’d like your product idea, they’ll lie to be polite.” The fix is not to find more honest people. The fix is to design the conversation so politeness can’t corrupt the data. That’s what this playbook is for.
A well-run customer interview has four legitimate goals. Anything outside this list is scope creep that will distort the data.
Three things customer interviews are not for, despite how many founders abuse them for these purposes:
If you’re tempted to use an interview for any of these three purposes, schedule a separate call. The interview is a research instrument; treat it that way.
The single most leveraged thing you can do to improve interviews is replace bad questions with good ones. Here’s the side-by-side.
Notice the pattern. Every “good” question asks about past behavior or current behavior. Every “bad” question asks about hypothetical future behavior. The reason: people are terrible at predicting their own future behavior, but reasonably accurate at describing what they actually did last week. Anchor every question in observable reality.
The best signal in any interview is when the prospect describes a specific time and place: “Last Tuesday at 4pm I was trying to reconcile invoices and I lost an hour because…” That sentence is worth more than 30 minutes of abstract opinion.
Recruiting is the unglamorous bottleneck. Most founders give up on customer interviews not because the interviews don’t work, but because finding 10 people willing to do them feels impossible. Four channels that actually work for solo founders:
Send 30 personalized messages per day to people who match your target persona. Don’t use Sales Navigator templates — the messages have to be hand-written. Reference something specific from their profile. Explicitly say you’re a founder doing research, not selling anything, and ask for 30 minutes. Realistic conversion: 5–15% accept rate, which means roughly 6–15 interviews per 100 messages over 2–3 weeks.
Find subreddits where your target persona hangs out. Don’t post asking for interviews — mods will (correctly) remove that. Instead, find people posting about the problem you’re researching, comment with genuine help, then DM after a day or two with an interview request. Conversion rates are surprisingly high (20–40%) because you’ve already established that you understand the space.
If your target persona is other solo founders or technical buyers, the Indie Hackers community and topic-specific Slacks (Lenny’s, MicroConf Connect, RevGenius) are the highest-density recruiting channels available. Be a member first, ask later. Showing up to a community only when you need something burns the channel permanently.
The fastest channel and the one founders most often dismiss because “they’re biased.” They are. But you can correct for bias by being explicit about it: “I know you’re going to want to be helpful, so I’m going to push back when you give me polite answers.” That single sentence dramatically improves the data quality from friendly interviewees.
The structure below is what we’ve seen work for solo SaaS founders running B2B research. Adjust segment lengths up or down depending on the prospect’s segment and your goals, but keep the order.
State the goal of the call clearly. “I’m doing research on how solo SaaS founders handle billing reconciliation. I’m not selling anything — I just need 30–45 minutes of your time to understand your workflow. Is now still a good time?”
Ask permission to record. Most people say yes. If they say no, take notes manually.
Understand their role, company, and day-to-day. “Tell me about what you do day-to-day. How big is the team? What does a typical week look like?”
Goal: build a mental picture of their world before you ask about the problem area. The same problem feels totally different depending on company size, role, and existing tooling.
This is the heart of the interview. Ask about the problem area without naming a solution. “Tell me about how you currently handle [problem area]. Walk me through the last time you did it.”
Then follow up. “What’s the worst part? When did it most recently break? What did you do then?”
Stay quiet. Long silences are productive. Most founders interrupt because the silence feels uncomfortable; the silence is when the prospect is thinking, and the answers that come after silences are usually the ones worth having.
What are they doing today to solve this? Even if their answer is “nothing,” that’s data — it tells you the problem isn’t painful enough to motivate any action.
“What tools are you using? How well do they work? What’s missing? Is there duct tape involved — spreadsheets, manual workarounds, custom scripts?”
The duct tape is gold. Manual workarounds are where actual products are born.
The hardest segment. Don’t ask “would you pay for this?” Ask about their actual budget. “What does your team currently spend on tooling for this category? Who approves software purchases under $500/month at your company?”
If the prospect mentions specific tools they pay for, ask the prices. If they don’t know prices, that’s data — they’re not the buyer.
“If I built something in this space, would you be open to a 15-minute followup in a few weeks to look at an early version? Who else should I talk to?”
The referral question is high-leverage. One good referral compounds; ten interviews from referrals are worth more than 30 from cold outreach because the trust is pre-built.
Most founders treat the post-interview phase as “write down what they said.” That’s wrong. The post-interview phase is where the actual research happens.
Steve Blank’s customer-discovery framework — documented across The Four Steps to the Epiphany and The Startup Owner’s Manual — insists on running multiple interview cohorts before drawing conclusions. The reason: any single interview is statistically meaningless. Patterns emerge only when you compare across 10–20 conversations. Solo founders who decide what to build after 3 interviews are working with sample sizes too small to mean anything.
The hardest thing about customer interviews is keeping your mouth shut about your product. Founders are excited about what they’re building. They want to test the idea against real people. The interview format seems like the perfect testing ground.
It isn’t. The moment you describe your product, the prospect’s reactions are anchored to your framing. You’ll get reactions to your idea, not data about their reality. The two are rarely the same thing.
The technique that works: write your product description on a card before the call. Put it on your desk. When you feel the urge to pitch, look at the card and remind yourself that pitching now will corrupt the data you spent two weeks recruiting for. After the interview, you can show them a demo as a separate conversation, scheduled for a different day. But not during the research call.
This discipline matters even more if you’re pre-product. The interview is for figuring out whether to build, not for confirming what you’ve already decided to build. The faster you can validate a real problem, the closer you are to the kind of 48-hour validation that lets you pivot or persist with confidence.
Interviews are necessary but not sufficient for product-market fit. They give you the language, the pain validation, and the willingness-to-pay signal. They don’t prove that those signals will translate to actual purchases when money is on the table. The next step is always to test the signals against a real offer — a landing page, a paid pre-order, or a $1 commitment — before betting months of build time on the conclusions.
Founders who skip the offer-test step ship products that interviewees said sounded great and then watched no one pay for. The interview gives you a hypothesis; the offer test confirms it.
For ideas that have made it through interviews and offer testing, the patterns we’ve catalogued in micro-SaaS ideas can show you what successful execution looks like in adjacent niches. And the zero to $1K MRR playbook picks up the story after interviews are complete and you’re ready to build.
Customer interviews work when you treat them as research and fail when you treat them as sales. Recruit aggressively, run a 45-minute structure, ask about past behavior instead of hypothetical future behavior, listen for specific moments and duct-tape workarounds, transcribe within a day, code patterns across 10+ interviews before drawing conclusions, and keep your product description in your pocket.
The founders we’ve seen ship the most successful solo SaaS products are not the ones with the best ideas. They’re the ones who interviewed 20 people before writing a line of code, listened more than they talked, and let the patterns lead them to a product the market actually wanted — not the one they happened to want to build.
The stack, prompts, pricing, and mistakes to avoid — for solo founders building with AI.