How to validate an idea without fooling yourself

“Is this a good idea?” invites self-deception because neither the founder nor the person being asked can observe the future business. The participant substitutes a reaction to the pitch; the founder substitutes enthusiasm, coherence, or praise for evidence. Both can be sincere while the conclusion remains unsupported.

Replace the verdict with smaller questions about the present and recent past. Does the problem recur for a specific group? What do people do now? What does the workaround cost? Who owns the decision? Which commitment will they make next? Compare the story with the base rate and decide before the conversation which evidence would cause you to narrow, change, or stop.

The Mom Test supplies the interview discipline and commitment standard. Thinking, Fast and Slow explains substitution, confirmation, the inside view, and why coherent evidence can feel more complete than it is.

A pitch measures the reaction to the pitch

When you describe the idea and ask whether someone likes it, the conversation becomes social. People avoid unnecessary conflict, reward visible enthusiasm, imagine an ideal future, and speak without bearing the cost of action. The founder also begins selling, explaining away confusion and interpreting questions as interest.

The resulting evidence is about comprehension, taste, politeness, and persuasion under that presentation. Those may be valid things to test later. They do not establish that the underlying problem is frequent or important enough to change behavior.

Keep early discovery focused on the person’s life. Ask for the most recent occurrence of the problem, then reconstruct the sequence, participants, tools, workaround, consequence, and decision. The specificity makes claims inspectable and reduces the need for either person to predict.

Do not hide your domain or manipulate the participant into disclosure. You can state the area of interest without revealing the answer you hope to hear. The objective is a truthful account, not a conversational trick.

Real problems leave behavioral traces

A problem important enough to support a product often changes behavior before the product exists. People spend time, assemble spreadsheets, hire services, tolerate disliked tools, escalate to colleagues, create policy, search repeatedly, or accept a known cost.

The workaround reveals frequency, stakes, ownership, and alternatives. It also exposes whether the requested user is the buyer. An operator may feel pain but lack authority; a buyer may fund risk reduction without performing the workflow.

Absence of action is not definitive proof that the problem is unimportant. Constraints can suppress action: no known alternative, no budget, privacy, learned helplessness, or fragmented ownership. Treat absence as a competing explanation to investigate, not as validation or rejection by itself.

Past behavior also needs a relevant window. One severe event years ago does not establish current recurrence. Ask how often it happens, when the last few instances occurred, and what changed across them. Preserve exact observations separately from your interpretation that a product opportunity exists.

Commitments test whether the conversation can advance

Fitzpatrick distinguishes compliments from commitment. A useful next step costs time, access, reputation, money, or workflow change. The cost makes the remaining constraint visible.

Match the commitment to the uncertainty. If the workflow is unclear, ask to inspect a sanitized artifact or speak with the operator. If authority is unclear, ask for an introduction. If value and delivery are concrete enough, propose a trial with owner, success condition, and an appropriate commercial commitment.

Letters of intent, wait-list signups, and meetings vary in strength according to what they obligate. A document designed to feel commercial while binding no one may still be weak evidence. State exactly what changed because the participant said yes.

Refusal contains information when you learn why. It may reveal priority, trust, timing, policy, switching cost, or lack of authority. Do not force a commitment simply to produce a metric; a premature ask can test the wrong relationship and damage access.

Your mind will substitute easier evidence unless the rule is written first

Kahneman’s substitution describes how a hard question is unconsciously replaced by an easier one. “Will this business work?” becomes “Can I imagine it working?”, “Do people like the demo?”, or “Did traffic increase?” The easier answer arrives with the confidence of the harder one.

Confirmation bias then changes the sample and interpretation. The founder recruits friendly participants, asks follow-up questions about positive comments, explains negative reactions, and remembers vivid support. A coherent narrative forms from selected evidence.

Counter this procedurally. Write the assumption, target group, observation, and update threshold before collecting data. Keep a disconfirming-evidence field in every interview note. Ask what the person already does that suggests the problem is adequately solved, infrequent, or owned elsewhere.

Use independent review where possible. Another person can inspect observations without having conducted the persuasive conversation. Blind or standardized analysis may be appropriate for quantitative tests. The objective is not to remove judgment but to keep the hoped-for outcome from silently changing the evidence standard.

The outside view tests whether your story is unusually plausible

The inside view explains why this particular idea should work: the founder’s skill, product insight, early conversations, and planned execution. The outside view asks what happens to similar ventures and which failure modes recur.

Choose a relevant reference class: similar customer, sales motion, price, behavior change, regulatory environment, and distribution model. Examine base rates for adoption, retention, sales cycles, acquisition, or failure where credible data exists. A broad statistic about all startups may be less useful than a small set of genuinely comparable attempts.

Base rates do not decide the case. They identify which part of the plan must differ and what evidence would justify that adjustment. If comparable products fail because distribution is expensive, product praise does not address the central risk.

A premortem makes the failure model explicit. Assume the idea failed after a defined period and write the most plausible history. Convert the top causes—no recurring problem, weak commitment, inaccessible buyer, poor retention, uneconomic channel, infeasible delivery—into evidence to seek now.

Validation is a sequence of narrower uncertainties

No finite set of interviews validates a whole business. Evidence can support a problem for one segment, then value for a particular outcome, then adoption in a workflow, then willingness to pay, retention, and a growth mechanism. Each step leaves other assumptions open.

State the scope of every conclusion. “Five operators in this role described a weekly workaround” is useful. “The market is validated” converts a bounded observation into a universal verdict and makes later contradictions feel like exceptions.

Negative evidence should change something. Narrow the segment, revise the problem, test a different mechanism, or stop. Running more conversations without changing the model turns validation into a search for permission.

Positive evidence should increase the cost of the next test gradually. As uncertainty falls, ask for stronger commitments and more realistic conditions. Do not jump from compliments to a full build or keep repeating low-cost interviews after the open question has moved downstream.

The next move is a disconfirmation brief and five interviews

Write one sentence for the current target user, recurring situation, costly behavior, and assumption that must be true. Add the strongest base-rate concern and the most plausible premortem failure. Then state what evidence across the next five qualified conversations would make you continue, narrow, revise, or stop.

During each conversation, reconstruct a recent event before describing the idea. Record the workaround, cost, stakeholders, prior attempts, and evidence that the current solution is good enough. End with the smallest commitment that tests the next uncertainty. Preserve facts, inferences, and compliments in separate fields.

After five, apply the written rule before pitching yourself a new interpretation. The success condition is not feeling validated. It is that the idea becomes more specific and the next investment reflects what people did, what the reference class predicts, and which uncomfortable evidence survived your attempt to disprove the story.