VEN-01 · Problem interviews that count

Write down what you believe about your customer's problem, and what you would have to hear to stop believing it. Then go find out. After a minimum set of interviews, report what you were wrong about — and whether the idea survived, changed, or should be abandoned.

A student who finishes this brief with the same idea they started with has either found a real problem or failed to listen. The evidence has to show which.

Teacher approves the hypothesis, the interview guide, and the contact plan in writing before any interview happens. Read the interview rule at the top of this bank first — it governs this brief completely.

The student needs:

• A written problem hypothesis, dated and countersigned before interviewing: who has the problem, what they do about it today, and what it costs them. Specific enough to be wrong.
• A kill criterion, written at the same time: the finding that would mean this problem is not worth solving. "If fewer than half of the people I talk to have tried to solve this in the past year, the problem is not urgent enough." A hypothesis with no kill criterion cannot be tested, only decorated.
• An interview guide built on past behavior, not future intent. "Would you use this?" produces politeness. "Walk me through the last time this happened" produces evidence. Every question in the guide is about something that already occurred: what they did, what it cost, what they tried, what they are using now. The Mom Test (Rob Fitzpatrick, 2013) is the standard short treatment of this distinction if the program wants a reading.
• A minimum set — the teacher sets the number, and eight to twelve is realistic for a term. At least three must be people who could plausibly tell you the problem is not real: someone outside your family and friend network, someone who solved it another way, or someone who does not have it at all. An interview set made entirely of supportive people measures support, not the problem.
• A notes template capturing what was said, not what it meant — interpretation is a separate column, written after.

Do not pitch during a problem interview. The moment you describe your solution, you stop collecting evidence and start collecting politeness. Solution reactions belong in a later conversation, and the guide should not contain them.

Pitching instead of listening, which is the failure this brief exists to break. Leading questions — "don't you find it frustrating when…" — that hand the interviewee their answer. Interviewing only friendly people and reporting the enthusiasm as demand. Counting compliments as evidence: "that's a great idea" is the single least informative sentence in customer discovery, and it should appear in the write-up as noise, not as a finding. Writing summaries from memory a week later. Quietly revising the hypothesis after the interviews so that it survives — the dated original is what makes that visible. Treating a small sample as a market: eight interviews can kill an idea, but they cannot prove one.

An evidence packet with five pieces:

1. Pre-registered hypothesis and kill criterion — dated, countersigned, and reproduced unchanged in the final packet no matter what happened.
2. Interview guide — the questions as asked, with any that changed mid-set noted and dated.
3. Interview log — one entry per conversation: date, role or segment (anonymized per the interview rule), what the person actually does today about the problem, what it costs them, and direct quotes kept separate from the student's interpretation. Note which three interviews were the potentially-disconfirming ones.
4. Evidence table — each claim in the hypothesis against the interviews that support it, contradict it, or say nothing. The "says nothing" column is usually the largest and must be shown, because absence of evidence is the most common result and the easiest to hide.
5. The verdict — killed, refined, or survived, measured against the kill criterion as written. If refined: the new hypothesis, and the specific thing you heard that changed it. If killed: say so plainly. A well-evidenced kill is a full-credit outcome in this bank and should be graded as one.

Interview notes stay with the school. The showcase version is anonymized and quotes only what the interviewee agreed could be quoted.

CT: distinguishing what someone does from what they say they would do. CR: writing questions that could return an unwelcome answer. CO: a peer reads three raw logs and independently marks each interview as support, contradiction, or silence; the disagreement gets reconciled before the evidence table is final. CM: a verdict a judge can read in one page without the student defending it in person. CZ: the people who gave you their time — what you owe them in honesty, anonymity, and not wasting it.

A founder, small-business owner, product manager, or researcher reviews the hypothesis, kill criterion, and interview guide at Checkpoint 1 — before any interview happens, since a leading guide cannot be fixed afterward — and the evidence table at Checkpoint 2. Standing instruction: reject any guide containing "would you," and reject any interview set with no potentially-disconfirming conversations. School-supervised, both times.

R1: one customer, one problem, one dated hypothesis. R2: logs complete enough that a peer can re-score them; quotes separated from interpretation. R3: comparator is what the customer does today — the status quo is the real competitor. R4: sample limits stated; the "says nothing" column shown. R5: the verdict fits one page. R6: names what the student refuses to claim from eight conversations, and what they owed the people in them.

(a) Twelve enthusiastic interviews with friends and family, a page of compliments, and an idea that emerges unchanged and untested. (b) The hypothesis is quietly rewritten after the fact to match what was heard, so the kill criterion is never actually applied and the brief certifies itself.

• Week 1–2: Hypothesis and kill criterion dated and countersigned; interview guide mentor-reviewed and scrubbed of leading questions; contact plan approved and outside-network introductions arranged through the school.
• Week 4–5: At least half the interviews complete and logged within a day of each conversation; the three disconfirming conversations identified; peer scoring of three logs.
• Week 7–8: Evidence table finished including the "says nothing" column; verdict written against the original kill criterion; showcase version anonymized.

Lane A immediately (business or entrepreneurship capstone, or a DECA/FBLA entrepreneurship event). Lane B through Entrepreneurship HQWBL only with the division's WBL coordinator owning the training plan and the outside contacts. No verified credit claim.