Transcript
SPEAKER_1: Alright, I've been thinking about this a lot since we covered the signals that come before product-market fit. The question I keep coming back to is: how do you actually research your way to fit? Like, what's the sequence? SPEAKER_2: That's exactly the right question to start with. And the key idea is that it's a loop, not a one-time test. Most teams skip straight to building and then wonder why nobody sticks around. The research has to come first, and it has to follow a specific order. SPEAKER_1: So walk us through that order. What's step one? SPEAKER_2: Segmentation. Not a broad market category — a narrow, specific slice of people. Think of it this way: instead of saying 'small business owners,' you'd say 'solo consultants in professional services who invoice more than ten clients a month and currently use spreadsheets.' That specificity is what makes the next steps work. SPEAKER_1: Right — because if the segment is too wide, the feedback you get is too averaged out to act on. SPEAKER_2: Exactly. And once the segment is defined, step two is problem interviews. But here's where most founders go wrong — they ask the wrong questions. They ask things like, 'Would you use a tool that did X?' And people say yes, because saying yes is polite. SPEAKER_1: That's the Mom Test problem, right? The idea that your mom will tell you your business idea is great even when it isn't. SPEAKER_2: Precisely. The principle there is to ask about real past events instead of hypothetical future interest. So instead of 'Would you pay for this?' you ask 'Walk me through the last time this problem cost you real time or money.' Past behavior is honest. Hypothetical enthusiasm is not. SPEAKER_1: So what are the core things those interviews should actually uncover? SPEAKER_2: Three things. First, what progress is the customer trying to make — that's the Jobs-to-be-Done lens. Second, what triggered the need in the first place. And third, what they're currently using to solve it and why that feels insufficient. Those three answers tell you whether the problem is real, urgent, and unsolved. SPEAKER_1: Mm-hmm. And urgency matters a lot here. How do researchers actually detect it? SPEAKER_2: Look for frequency, cost, and emotional weight. If someone mentions the problem once a week, has lost money because of it, and gets visibly frustrated describing it — that's urgency. If they shrug and say 'yeah, it's a bit annoying sometimes,' that's not a hair-on-fire problem. You also want to probe switching costs: ask what it would take for them to stop using their current workaround. If the answer is 'not much,' that's a green flag. SPEAKER_1: Wait — so positive feedback during interviews can actually be a warning sign? SPEAKER_2: [short pause] It can be, yes. 'I love this idea' is one of the weakest signals in early research. It costs nothing to say. What matters is whether someone has already spent time, money, or social capital trying to solve the problem themselves. Active workarounds are far more predictive than compliments. SPEAKER_1: enthusiasm is cheap, behavior is expensive. But the pressure point is — how do you test actual behavior before you've built anything? SPEAKER_2: behavioral demand tests. For example, in B2B contexts, a strong tactic is to sell the product before it exists — secure a letter of intent, a paid pilot, or a proof-of-concept deposit based on a concept alone. In consumer contexts, a landing page with a real signup or preorder does similar work. You're measuring willingness to act, not willingness to agree. SPEAKER_1: And that's fundamentally different from an MVP, right? Or does the MVP fit in here? SPEAKER_2: The MVP comes in right after those demand signals confirm the direction. You put a simplified version in front of a small, targeted group — often batches of around five to eight users — and you watch what they actually do. Retention, frequency of use, whether they refer others. Those are the real indicators. SPEAKER_1: Which brings us to step four. The Sean Ellis test. SPEAKER_2: Right — but the timing matters enormously. The Sean Ellis survey asks users how they'd feel if they could no longer use the product. The options are 'very disappointed,' 'somewhat disappointed,' or 'not disappointed.' The benchmark is that if 40% or more say 'very disappointed,' that's a strong indicator of fit. But — and this is critical — the test only works after users have genuinely experienced the product. Running it on people who've barely touched it gives you noise, not signal. SPEAKER_1: segment first, then problem interviews to find real urgency, then behavioral demand tests to confirm willingness to act, and only then — after real usage — the Sean Ellis threshold. SPEAKER_2: That's the loop. And the reason it's a loop is that weak signals at any stage should trigger iteration, not scaling. If the 40% threshold isn't reached, the team goes back — narrows the segment further, revisits the value proposition, adjusts the feature set. The research doesn't stop at launch. SPEAKER_1: What does strong fit actually look like when it's there? Like, what are the unmistakable signs? SPEAKER_2: Repeated use without prompting. Referrals that come in organically. Customers describing the product as painful to lose — not just nice to have. Willingness to pay, and high retention across cohorts. And sometimes, increased media or analyst attention that reflects growing recognition in the market. When those signals stack up together, that's not a coincidence. That's pull. SPEAKER_1: The takeaway for anyone running this research: compliments predict almost nothing. Behavior — payment, repeated use, referrals, that 'very disappointed' response — that's what reveals whether a real market exists. SPEAKER_2: And the research loop is what gets a team to that evidence systematically. Segment tightly, interview for past behavior, test demand with real stakes, then measure fit after genuine use. That sequence is the difference between building something people say they want and building something they actually can't function without. SPEAKER_1: One thing I want to make sure our listener doesn't miss — this loop isn't just for tech founders. Someone building a lecture series, a curriculum, even a micro-drama program is essentially doing the same thing: testing whether a specific audience has a real, urgent need for what they're creating. SPEAKER_2: That's a sharp connection. The mechanics are identical. Define who you're for, find out what they're genuinely struggling with, test whether they'll act on it, then measure whether they'd miss it. The domain changes; the sequence doesn't. SPEAKER_1: So let's talk about what happens when the signals are mixed. Say someone runs the problem interviews, hears urgency, but then the behavioral demand test comes back weak. What does that actually mean? SPEAKER_2: It usually means one of two things. Either the segment is still too broad — the urgency is real for a subset, but averaged out across the group it looks soft. Or the problem is real but the proposed solution isn't the right vehicle for solving it. Both are useful findings. Neither means stop. SPEAKER_1: So not a dead end — a redirect. SPEAKER_2: Exactly. The research loop is designed to surface those redirects early, before major resources are committed. That's the whole point of running behavioral demand tests before building a full product. A letter of intent or a paid pilot from even two or three customers in B2B tells you more than fifty polite interviews. SPEAKER_1: And in consumer contexts — what's the equivalent of a letter of intent? SPEAKER_2: A landing page with a real preorder or signup. Not a 'coming soon' page — something that asks for an actual commitment. An email address is weak. A credit card number, or even a waitlist with a deposit, is much stronger. You're measuring willingness to act, not willingness to agree. SPEAKER_1: Right — and that distinction matters because verbal agreement is almost free. Acting costs something. SPEAKER_2: Which is why the research has to keep pushing past words. Think of it this way: if someone describes a problem vividly, has already built a workaround, and is willing to pay before the product exists — that's three independent signals pointing the same direction. That's a very different situation from someone who says 'yeah, I'd probably use that.' SPEAKER_1: Now, the Sean Ellis test sits at the end of this loop. But I want to make sure the timing is clear for anyone running this. When exactly is it appropriate? SPEAKER_2: [inhale] Only after users have genuinely experienced the product. Not after a demo. Not after a free trial they barely touched. After real, repeated use. The question — how would you feel if you could no longer use this — only produces honest signal when the person has actually built a habit around it. SPEAKER_1: And the 40% threshold — that's not arbitrary, right? There's a reason that number became the benchmark. SPEAKER_2: It emerged from observing patterns across many early-stage products. Teams that hit 40% or more of respondents saying 'very disappointed' tended to see the other signals follow — retention, referrals, organic growth. Below that, teams were often still in the iteration phase, whether they admitted it or not. SPEAKER_1: So the number is almost a forcing function. It makes the team honest about where they actually are. SPEAKER_2: That's a good way to put it. And the formula is simple: divide the number of 'very disappointed' responses by total responses, multiply by 100. If the result is below 40, the answer isn't to scale marketing. The answer is to go back — narrow the segment, revisit the value proposition, adjust the feature set. SPEAKER_1: What about tracking fit over time? Because markets shift. A product that had strong fit two years ago might be losing it now. SPEAKER_2: That's one of the most underappreciated parts of this. Product-market fit isn't a permanent state. Competitors evolve, customer expectations shift, new alternatives emerge. The research loop needs to run on a recurring cadence — in early stages, something like quarterly. Churn data, cohort retention, and periodic re-surveys are what catch drift before it becomes a crisis. SPEAKER_1: the research loop is the work. Segment tightly, interview for past behavior and real urgency, test demand with actual stakes, measure fit after genuine use, and keep running the loop as the market moves. SPEAKER_2: And remember — the goal isn't to find people who like the idea. It's to find people who can't function without the solution. When repeated use, referrals, willingness to pay, and that 'very disappointed' response all stack up together, that's not enthusiasm. That's pull. And pull is what the research is designed to find.