Customer Discovery and Problem–Solution Validation

Introduction

Entrepreneurial selling begins before a product is fully designed and before a sales team is hired. It begins with disciplined inquiry into how people currently experience a problem, how they make decisions, what alternatives they use, and under what conditions they will change their behavior. Customer discovery is the systematic process of learning about these matters from prospective customers and other relevant stakeholders. Problem–solution validation is the subsequent process of testing whether the identified problem is sufficiently important and whether a proposed solution creates credible, observable value.

The distinction between discovery and validation is fundamental. Discovery asks, “What is happening in the customer’s world?” Validation asks, “Does our proposed response solve a meaningful problem for a sufficiently attractive segment?” Neither process is equivalent to asking whether people like an idea. Positive opinions, hypothetical purchase intentions, and polite encouragement are weak evidence. Stronger evidence comes from observed behavior, existing expenditure, repeated workarounds, access to decision-makers, commitments of time or money, and successful experiments.

For entrepreneurs, customer discovery is also a selling capability. It develops the ability to listen, diagnose, qualify opportunities, identify buying stakeholders, articulate customer value, and prioritize acquisition activities. At graduate level, the task is not merely to conduct interviews but to construct a defensible evidence system that connects customer problems to assumptions, experiments, decisions, and acquisition economics.

Learning Objectives

By the end of this chapter, readers should be able to:

1. Distinguish customer discovery, problem validation, solution validation, and product–market learning.
2. Design and conduct rigorous customer interviews and observational studies.
3. Identify and control common forms of researcher, respondent, and entrepreneurial bias.
4. Translate qualitative and behavioral evidence into explicit assumptions and testable hypotheses.
5. Design low-cost experiments that assess problem intensity, solution value, willingness to act, and acquisition feasibility.
6. Synthesize evidence using structured methods rather than isolated anecdotes.
7. Prioritize customer segments and acquisition initiatives under uncertainty.
8. Evaluate ethical, privacy, and governance issues in customer research.

1. The Logic of Customer Discovery

A venture is built around a set of uncertain claims. Typical claims include:

• A particular customer segment experiences a recurring problem.
• The problem is costly, risky, inconvenient, or emotionally significant.
• Existing alternatives are inadequate.
• A proposed solution can improve the customer’s outcome.
• A reachable buyer has authority and resources to purchase.
• The venture can acquire customers at an economically sustainable cost.

These claims should not be treated as facts merely because they appear in a business plan. They are assumptions until supported by evidence. Customer discovery therefore resembles applied research: it starts with a question, gathers evidence from relevant sources, interprets patterns, and revises explanations.

A useful distinction is between four learning stages:

The stages are related but not interchangeable. A prototype demo may show that a product can be used, but it does not prove that the underlying problem is important. Similarly, a customer may express strong interest but fail to adopt because the buying process, switching cost, or implementation burden is too high.

Jobs, pains, gains, and alternatives

A customer problem is more than a sentence such as “customers need better software.” It should be described in the customer’s context. A job is what a person or organization is trying to accomplish. A pain is an obstacle, cost, risk, or undesirable experience associated with that job. A gain is a desired improvement. An alternative is any way the customer currently addresses the situation, including doing nothing.

For example, a small manufacturer’s job may be to identify production defects before shipment. Pains may include rework, delayed deliveries, and customer complaints. Gains may include earlier detection and more predictable throughput. Alternatives may include manual inspections, spreadsheet tracking, or accepting a known defect rate. The existence of alternatives means that a new venture competes not only with direct rivals but also with inertia, internal workarounds, and the status quo.

A strong discovery statement specifies:

1. Who experiences the situation.
2. When and where it occurs.
3. What job the customer is attempting to perform.
4. What current behavior reveals the problem.
5. What consequence results.
6. What alternatives are used.
7. Who experiences the cost and who can authorize change.

2. Designing a Customer Discovery Study

Customer discovery is most effective when it is designed before conversations begin. A lightweight research protocol should state the decision to be informed, the target participants, the evidence sought, the interview or observation method, and the rule for changing course.

Define the decision, not merely the topic

A weak research question is, “Do people like our idea?” A stronger question is, “Among independent clinics with fewer than twenty employees, when appointment cancellations occur, what do managers currently do, what does the problem cost, and what evidence would justify testing an automated reminder service?” The second question identifies behavior, context, and a decision threshold.

A discovery plan should also identify the unit of analysis. In business markets, the user, economic buyer, technical evaluator, procurement officer, and executive sponsor may be different people. Interviewing only end users can produce an incomplete account of adoption barriers and purchasing authority.

Sampling for learning

Early-stage discovery rarely requires a statistically representative sample. It does require purposeful variation. Researchers should recruit participants who have direct experience with the situation and should deliberately include contrasting cases: frequent and infrequent users, successful and unsuccessful adopters, small and large organizations, and customers who use different alternatives.

This is sometimes called theoretical or purposeful sampling. The objective is to explore the boundaries of a problem and identify mechanisms, not to estimate a population percentage prematurely. As patterns stabilize, the researcher can seek disconfirming cases to determine whether apparent findings are robust or segment-specific.

Recruitment must avoid overreliance on convenient friends, enthusiastic early adopters, or existing contacts. These groups may be easier to reach but can systematically overstate demand. A recruitment record should document how participants were found, their relevant characteristics, incentives provided, and whether they are current customers, prospects, partners, or observers.

3. Interviewing for Evidence Rather Than Approval

The customer interview is a central discovery instrument, but its quality depends more on question design and listening discipline than on the number of interviews.

Characteristics of a strong interview

A rigorous interview:

• focuses on a recent or concrete episode rather than general opinion;
• asks about current behavior before presenting a proposed solution;
• explores consequences, workarounds, costs, and decision processes;
• uses open questions followed by precise probes;
• allows the participant to introduce unexpected information;
• records quotations and behavioral evidence separately from interpretation;
• ends by asking about the next realistic step, not merely satisfaction.

Useful prompts include:

• “Tell me about the last time this happened.”
• “What did you do first?”
• “What alternatives did you consider?”
• “How often does this occur?”
• “What happens if the issue is not resolved?”
• “Who else becomes involved?”
• “What have you already tried?”
• “What did that cost in time, money, risk, or lost opportunity?”

Questions such as “Would you use an app that solved this?” are weak because they invite speculation and social approval. If a concept must be shown, it should normally be shown after the participant’s existing experience has been explored. Even then, the researcher should ask what the participant would do next, what would prevent adoption, and how the concept compares with current alternatives.

The interview funnel

A practical interview sequence begins broadly and narrows gradually:

1. Context: role, responsibilities, environment, and relevant history.
2. Recent episode: a specific occurrence of the problem.
3. Existing behavior: tools, workarounds, people, and process.
4. Consequences: operational, financial, emotional, and strategic effects.
5. Decision process: authority, budget, timing, criteria, and risk.
6. Concept response: reactions to a prototype or proposed intervention.
7. Commitment: willingness to continue, test, introduce, pay, or refer.

The interviewer should distinguish a participant’s description from the interviewer’s inference. “We lose approximately one day per month because of reconciliation errors” is evidence reported by a participant. “The customer urgently needs our platform” is an interpretation that requires corroboration.

Interview reliability and reflexivity

Qualitative research is influenced by the researcher. Interviewers may unintentionally signal the desired answer through facial expressions, emphasis, or selective follow-up. A research team can improve reliability by using a common discussion guide, training interviewers, comparing notes, and periodically reviewing recordings where consent has been obtained.

Reflexive notes are useful. After each interview, the researcher should record what was learned, what assumptions were challenged, what leading behavior may have occurred, and what new question should be investigated. This is not an admission of failure; it is a method for making the research process visible and correctable.

4. Observation and Contextual Inquiry

People do not always report behavior accurately. They forget, rationalize, normalize inconvenience, or describe an idealized process. Observation helps reveal what actually happens in context.

Observation may involve shadowing a workflow, examining artifacts such as forms and spreadsheets, reviewing support tickets, analyzing search queries, or watching a participant complete a task. Contextual inquiry combines observation with short questions asked at the point of work. For example, instead of asking a warehouse manager how inventory is updated, a researcher can observe the next stock receipt and ask why each handoff occurs.

Observation should be systematic. A field protocol can specify:

• the setting and time period;
• the workflow or event to be observed;
• actions, delays, errors, handoffs, and workarounds;
• tools and information used;
• interruptions and environmental constraints;
• explicit statements made by participants;
• the researcher’s interpretations, recorded separately.

The observer must avoid disrupting safety-critical, confidential, or legally protected activities. In organizational settings, permission from the appropriate authority is not sufficient if individual workers have not been informed about the purpose and use of the research.

Observation is especially valuable when the proposed solution depends on integration into an existing process. A technically elegant solution may fail because it adds data entry, interrupts established authority, or conflicts with timing constraints that were invisible in an interview.

5. From Insights to Testable Assumptions

The principal analytical task is to convert unstructured findings into explicit, ranked assumptions. A useful assumption statement has the form:

For [specific segment] in [specific context], [observable problem or behavior] occurs with [specified frequency or condition], causing [consequence]. If we offer [solution or intervention], customers will [observable action] because it creates [valued outcome].

For example:

For independent dental clinics using manual recall lists, missed follow-up appointments occur weekly, creating unused capacity. If the clinic receives an automated recall workflow integrated with its scheduling system, office managers will activate it and continue using it because it reduces unfilled appointment slots without adding administrative work.

This statement contains several assumptions that should be separated:

• segment assumption: independent dental clinics are the relevant segment;
• problem assumption: missed follow-up appointments are frequent;
• consequence assumption: they create material unused capacity;
• solution assumption: an automated workflow can reduce the problem;
• usability assumption: it will not add administrative burden;
• buying assumption: the office manager can authorize adoption;
• acquisition assumption: clinics can be reached through a viable channel.

Importance and uncertainty

Not all assumptions deserve equal attention. A useful prioritization approach considers both the consequence of being wrong and the uncertainty of the claim. One can assign each assumption a low, medium, or high rating for these dimensions. The highest-priority tests address assumptions that are both consequential and uncertain.

A more quantitative screening score can be used when a team needs to compare many assumptions. Let I denote an importance rating from 1 to 5, and U denote an uncertainty rating from 1 to 5. A simple prioritization score is:

Here, P is a prioritization score, I is the estimated consequence of being wrong, and U is the team’s uncertainty. The score is ordinal rather than a probability or financial forecast. It is useful for ordering research, but it should not be interpreted as a precise measure of venture risk.

For example, an assumption rated I = 5 and U = 4 receives P = 20. An assumption rated I = 2 and U = 5 receives P = 10. The first should normally be tested earlier because failure would be more damaging, even though the second is more uncertain.

6. Experiments for Problem and Solution Validation

An experiment is a deliberate intervention designed to produce evidence relevant to a decision. It should identify the assumption, the method, the observable behavior, the threshold for interpretation, and the next action.

Problem validation experiments

Problem validation can use:

• interviews about recent incidents;
• analysis of customer records, complaints, or support requests;
• diary studies documenting occurrences over time;
• workflow observation;
• retrospective analysis of spending on substitutes;
• concierge services in which the venture manually performs the proposed function.

The aim is to establish recurrence, consequence, and relevance. A problem that is experienced once but causes severe loss may be more important than a frequent minor inconvenience. Researchers should therefore examine both frequency and impact rather than treating frequency as the only indicator.

Solution validation experiments

A solution experiment may involve a sketch, clickable prototype, landing page, demo, manual service, pilot, pre-order, or limited paid trial. The appropriate experiment depends on the assumption. If the uncertainty concerns usability, observe task completion. If it concerns willingness to pay, test a real price or a credible commercial commitment. If it concerns operational feasibility, run a pilot in the customer’s environment.

Evidence generally becomes stronger as the customer’s behavior becomes more costly and more consequential. A verbal expression of interest is weaker than joining a scheduled pilot; a scheduled pilot is weaker than repeated use; repeated use is weaker than continued paid use only in some contexts, because payment may be constrained by procurement timing while genuine use can still reveal value. Evidence must be interpreted in context.

Experiment design template

A concise experiment specification includes:

Thresholds should be established before reviewing results where practical. Otherwise, teams may move the goalposts after seeing disappointing evidence. Thresholds do not need to be statistically sophisticated, but they should be logically connected to the decision. For example, a team might decide that at least three of five target organizations must agree to a defined pilot within thirty days before investing in integration work. This is a managerial decision rule, not a population estimate.

7. Evidence Synthesis and Interpretation

Discovery produces fragments: statements, observations, records, demonstrations, and commitments. Synthesis turns these fragments into a reasoned conclusion.

A practical synthesis process is:

1. Transcribe or summarize evidence soon after collection.
2. Separate direct evidence, interpretation, and unresolved questions.
3. Code recurring topics such as triggers, workarounds, consequences, objections, and buying criteria.
4. Compare patterns across segments and cases.
5. Search deliberately for negative cases and contradictory evidence.
6. Map findings to assumptions and confidence levels.
7. Decide whether to persevere, modify, narrow, or abandon the current hypothesis.

A compact evidence matrix can help:

Confidence should not be confused with truth. It reflects the strength, relevance, and consistency of available evidence. Ten interviews from one narrow network may provide less confidence about a broad market than six interviews across meaningfully different contexts.

Triangulation

Triangulation compares different sources or methods. For instance, a customer may report that a task takes two hours, while observation shows that it takes forty minutes but causes frequent interruptions. Neither source should automatically be discarded. The discrepancy may reveal that customers define the task differently or that the hidden cost lies in interruption rather than elapsed time.

Useful forms include:

• source triangulation: different roles, organizations, or segments;
• method triangulation: interviews, observation, records, and experiments;
• researcher triangulation: more than one analyst reviews evidence;
• time triangulation: evidence collected across different periods or events.

8. Bias, Validity, and Decision Quality

Entrepreneurial teams face strong motivational pressures. They may have invested money, developed a prototype, announced a vision publicly, or become emotionally attached to a segment. These pressures make disciplined bias control essential.

Common biases

Confirmation bias is the tendency to seek or favor evidence supporting an existing belief. Leading-question bias occurs when wording suggests the desired answer. Selection bias arises when participants are not representative of the relevant decision context. Recall bias affects reports of past behavior. Social desirability bias leads respondents to provide answers that appear helpful or sophisticated. Survivorship bias occurs when researchers study successful adopters but ignore those who rejected or abandoned a solution.

Additional risks include anchoring on the first attractive insight, overgeneralizing from a vivid anecdote, confusing friendliness with demand, and treating silence as agreement.

Bias-control practices

Teams can reduce bias by:

• interviewing about past behavior before revealing the concept;
• using neutral wording and consistent prompts;
• recruiting disconfirming and non-user cases;
• separating interviewer and salesperson roles where feasible;
• recording exact statements and observed actions;
• preregistering or documenting decision thresholds;
• having a colleague challenge the interpretation;
• reporting failed experiments as carefully as successful ones.

Validity has multiple dimensions. Construct validity asks whether the research measures the intended concept. A question about liking a prototype may not measure willingness to pay. Internal validity asks whether the observed outcome is plausibly caused by the intervention rather than by unrelated factors. External validity asks whether findings transfer to other segments or contexts. Decision validity asks whether the evidence is sufficient for the specific managerial decision, even if it is not generalizable to an entire population.

9. Translating Validation into Acquisition Priorities

Validation should lead to action. The output is not merely a research report but a prioritized acquisition strategy.

A segment is attractive when it has a meaningful problem, identifiable potential buyers, reachable channels, sufficient resources, and conditions that support adoption. A segment with severe pain may still be unattractive if decision cycles are extremely long, regulation prevents experimentation, or acquisition costs are prohibitive.

Acquisition priorities can be organized around four questions:

1. Value: Is the problem important enough to motivate change?
2. Access: Can the venture reach the relevant user and buyer efficiently?
3. Activation: Will the customer take a meaningful first action, such as a pilot or trial?
4. Expansion: Can initial value lead to retention, referrals, broader use, or repeat purchase?

At this stage, precise forecasts are usually premature, but teams can compare channels using observed funnel behavior. If N₀ qualified prospects are contacted, N₁ respond, and N₂ become paying customers, then observed response and conversion rates can be calculated as:

Here, r is the response rate and c is the conversion rate among responders. The quantities must refer to clearly defined stages and the same time period. These ratios describe observed performance; they do not prove future performance, especially when the sample is small or non-random.

If a channel costs C in total and produces N₂ customers, the observed acquisition cost per customer is:

This calculation is meaningful only when the cost definition is explicit. It may include advertising, sales labor, commissions, tools, and allocated creative costs, or only direct campaign spending. The numerator and denominator must use the same period and attribution rule. Early-stage teams should avoid comparing a narrow media cost with a fully loaded customer value estimate.

Worked application

Suppose a venture conducts outreach to 80 qualified prospects. Twenty-four respond to a discovery invitation, and six agree to a paid pilot. The outreach program costs $1,800, including researcher time and tools.

The response rate is:

The conversion rate among responders is:

The observed acquisition cost per paid pilot is:

These results support the claim that the channel generated paid pilots under the tested conditions. They do not establish that the channel will scale, that the pilots will renew, or that $300 is sustainable. Those questions require retention, expansion, and contribution-margin evidence.

10. Illustrative Case: CareRoute Scheduling Support

Background

CareRoute is a fictional early-stage venture proposing scheduling support for independent home-care agencies. Its founders believe agencies lose revenue and reduce caregiver reliability because last-minute absences require urgent manual rescheduling.

Initial problem

The founders initially described the problem as “home-care agencies need an AI scheduling platform.” This formulation prematurely specified a technology and did not identify the user, workflow, or consequence.

Discovery approach

The team interviewed agency owners, schedulers, caregivers, and field supervisors. It also observed scheduling work during morning dispatch periods and reviewed anonymized cancellation logs from two agencies. Interviews were conducted before showing any software concept.

The research revealed three important patterns. First, last-minute absences were common, but the greatest burden was not simply finding a replacement. Schedulers also had to consider caregiver qualifications, travel time, client preferences, overtime rules, and continuity of care. Second, agencies already used scheduling software, but it did not support rapid exception handling. Third, owners were interested in reducing overtime, while schedulers prioritized a calm and manageable morning workflow.

The team also found a negative case: one small agency experienced few absences because it maintained a large flexible staff pool. This agency did not view the proposed problem as urgent. The finding led the team to narrow its target segment to agencies with high visit density and limited replacement capacity.

Assumptions and experiments

The revised assumptions were:

• agencies with limited replacement capacity experience repeated scheduling disruptions;
• disruption creates measurable overtime or missed visits;
• schedulers will use a rapid recommendation tool if it fits existing software;
• owners will pay when the tool reduces avoidable overtime without reducing care continuity.

The first experiment was a concierge pilot. For four weeks, CareRoute staff manually prepared replacement recommendations using information supplied by two agencies. This avoided premature investment in complex automation while testing whether recommendations were useful in the real workflow.

The pilot produced mixed results. Schedulers used the recommendations for routine replacements but rejected them when client preferences or caregiver relationships were not visible in the data. Interviews after each shift showed that trust depended on explaining why a recommendation was made. The team therefore changed the solution from an opaque “best match” score to a ranked list showing qualification, travel time, availability, and relevant constraints.

Outcome

The pilot did not validate the original claim that agencies wanted a general AI scheduling platform. It did provide evidence for a narrower proposition: selected agencies valued transparent support for urgent replacement decisions when the tool reduced search time and preserved human control. One agency agreed to a paid extension, while another declined because its operational volume was too low.

Learning points

1. The initial technology-led framing obscured the actual job.
2. Observation exposed constraints that interviews alone underemphasized.
3. A manual service tested workflow value before expensive automation.
4. A negative case improved segment definition.
5. Transparency and human override were part of the solution, not optional interface features.
6. Paid continuation was stronger evidence than favorable prototype reactions.

11. Ethical, Privacy, and Governance Considerations

Customer discovery involves people, organizations, and often sensitive information. Ethical research requires informed participation, proportional data collection, confidentiality, and honest communication about commercial intent.

Participants should know who is conducting the research, why they were invited, whether participation is voluntary, whether conversations are recorded, how data will be stored, and how findings may be used. Incentives should compensate time without becoming coercive. Researchers should be especially careful with employees who may feel pressured by an employer, patients, children, financially vulnerable people, or individuals whose disclosure could create professional or personal harm.

Data minimization is a sound governance principle: collect only what is necessary for the decision. Access should be restricted, retention periods defined, and identifying information separated from research notes where feasible. In regulated contexts, such as health, finance, education, or employment, applicable privacy and sector-specific requirements must be considered before collecting or processing data. Anonymization should not be claimed casually; combinations of apparently harmless attributes can re-identify individuals.

AI tools can assist with transcription, coding, clustering, and summarization, but they introduce risks. Automated systems may mis-transcribe accents, infer sensitive attributes, reproduce stereotypes, or expose confidential information through external services. Human review remains necessary for consequential interpretations. Teams should document what data were submitted to an AI system, verify summaries against source material, and avoid treating generated themes as independent evidence.

12. Contemporary Developments and Best Practice

Digital channels have expanded the evidence available to entrepreneurs. Search behavior, product analytics, community discussions, support records, and online reviews can reveal language and recurring problems. However, digital traces are not automatically representative. The most vocal users may not be the most valuable customers, and platform algorithms may distort visibility.

Remote interviews and asynchronous diary studies make it possible to study geographically dispersed customers. Product-led ventures can combine qualitative discovery with usage telemetry to identify where users struggle or abandon a workflow. Synthetic data and AI-generated personas may support scenario exploration, but they cannot substitute for contact with real customers. They are models of assumptions, not evidence of demand.

Best practice is therefore integrative:

• begin with a decision and an explicit assumption;
• study real behavior in context;
• use multiple sources where the decision is important;
• seek negative evidence deliberately;
• test commitments rather than compliments;
• preserve a clear distinction between evidence and interpretation;
• update the segment, solution, and acquisition plan together;
• stop experiments that no longer address a consequential uncertainty.

Chapter Summary

Customer discovery is a disciplined process for learning how customers experience problems, perform jobs, use alternatives, and make decisions. Problem validation determines whether a problem is sufficiently important and recurrent for a defined segment. Solution validation tests whether a proposed intervention creates observable value and prompts meaningful customer action. Acquisition validation examines whether the venture can reach, convert, retain, and expand customers through viable channels.

High-quality discovery uses interviews, observation, records, prototypes, concierge services, pilots, and commercial commitments. Interviews should focus on recent behavior rather than hypothetical approval. Observation reveals workflow constraints and hidden workarounds. Assumptions should be separated into segment, problem, consequence, solution, buying, and acquisition claims, then prioritized according to consequence and uncertainty.

Evidence synthesis requires matrices, triangulation, comparison across segments, and deliberate attention to disconfirming cases. Entrepreneurial teams must manage confirmation bias, leading questions, selection bias, recall problems, and overinterpretation of anecdotes. Metrics such as response rate, conversion rate, and acquisition cost can support decisions when stage definitions, cost scopes, and time periods are explicit, but early observations should not be mistaken for universal forecasts.

The ultimate purpose of discovery is managerial action: to identify a specific segment, articulate a credible value proposition, choose an acquisition priority, and decide whether to persevere, modify, narrow, or abandon the current hypothesis. Ethical practice requires informed participation, privacy protection, data minimization, responsible use of AI, and honest reporting of negative results.

Key Terms

Customer discovery: A structured process of learning about customer problems, behaviors, contexts, and decisions.

Problem validation: Evidence-based assessment of whether a problem is important, recurrent, and relevant to a target segment.

Solution validation: Testing whether a proposed solution creates valued outcomes and meaningful customer action.

Acquisition validation: Testing whether customers can be reached and converted through a sustainable channel.

Job: The functional, social, or emotional outcome a customer is trying to accomplish.

Pain: A cost, obstacle, risk, frustration, or undesirable consequence associated with a job.

Gain: A desired improvement or beneficial outcome.

Alternative: Any current response to a situation, including competitors, workarounds, and doing nothing.

Assumption: An unverified claim on which a venture’s decisions depend.

Hypothesis: A specific, testable statement about an expected relationship or behavior.

Purposeful sampling: Selecting participants because they possess relevant experience or represent meaningful variation.

Contextual inquiry: Learning through observation and questions conducted in the setting where work occurs.

Triangulation: Comparing evidence across sources, methods, researchers, or time periods.

Confirmation bias: Favoring information that supports an existing belief.

Disconfirming evidence: Information that challenges or weakens an assumption.

Concierge experiment: A manually delivered service used to test customer value before automation.

Pilot: A limited implementation used to evaluate a solution in a real operating context.

Response rate: The proportion of a defined outreach group that performs the specified response action.

Conversion rate: The proportion of participants at one stage who advance to a defined subsequent stage.

Customer acquisition cost: The specified cost of acquiring one customer under a stated attribution rule.

Review Questions

Short Answer Questions

1. What is the difference between problem validation and solution validation?
2. Why are questions about recent behavior generally stronger than questions about hypothetical future use?
3. Identify four common sources of bias in entrepreneurial customer research.
4. What is the purpose of a concierge experiment?
5. Why should acquisition metrics define their stages and cost scope explicitly?

Long Answer Questions

1. Explain how interviews, observation, and behavioral experiments complement one another in customer discovery.
2. Describe a rigorous process for converting interview findings into ranked, testable assumptions.
3. Discuss how triangulation improves the quality of entrepreneurial evidence.
4. Explain how a venture should move from validated customer problems to acquisition priorities.
5. Evaluate the ethical and governance issues involved in using AI to analyze customer research.

Analytical / Application Questions

1. A team receives enthusiastic feedback from twenty conference attendees, but none agrees to a pilot. Diagnose what this evidence does and does not validate, and design the next experiment.
2. A venture contacts 120 qualified prospects, receives 30 responses, and obtains 9 paid trials at a total outreach cost of $2,700. Calculate the response rate, responder-to-trial conversion rate, and observed cost per paid trial. Interpret the limitations of these measures.
3. Design a discovery study for a proposed service that helps small retailers reduce inventory stockouts. Specify the target participants, interview topics, observational method, major assumptions, one disconfirming case to seek, and a decision rule for a pilot.

Multiple Choice Questions with Answers

Question 1

Which statement best distinguishes customer discovery from solution validation?

A. Discovery measures advertising performance, while validation measures employee productivity.

B. Discovery investigates customer situations and problems, while validation tests whether a proposed response creates value.

C. Discovery occurs only after a product launch, while validation occurs before any customer contact.

D. Discovery estimates market size, while validation eliminates the need for experiments.

Correct Answer: B. Discovery investigates customer situations and problems, while validation tests whether a proposed response creates value.

Question 2

Which interview question is most likely to produce evidence about actual behavior?

A. Would you like an application that solves this problem?

B. Do you agree that this is an important market need?

C. Tell me about the last time this problem occurred and what you did first.

D. Would you recommend our proposed solution to a colleague?

Correct Answer: C. Tell me about the last time this problem occurred and what you did first.

Question 3

A team rates an assumption’s importance as 4 and its uncertainty as 5. Using P = I × U, what is its prioritization score?

A. 1

B. 9

C. 20

D. 45

Correct Answer: C. 20

Question 4

Which activity is an example of contextual inquiry?

A. Asking friends whether they like a product concept

B. Observing a scheduler handle a real replacement request and asking questions during the workflow

C. Sending a general market survey without segment definitions

D. Reading a competitor’s promotional brochure

Correct Answer: B. Observing a scheduler handle a real replacement request and asking questions during the workflow

Question 5

Which is the strongest evidence that a proposed solution may create customer value?

A. A participant says the idea sounds interesting.

B. A participant follows the venture on social media.

C. A participant repeatedly uses and pays for the solution in a real workflow.

D. The founder believes the prototype is technically superior.

Correct Answer: C. A participant repeatedly uses and pays for the solution in a real workflow.

Question 6

A venture contacts 100 prospects and receives 20 responses. What is the observed response rate?

A. 5%

B. 20%

C. 50%

D. 80%

Correct Answer: B. 20%

Question 7

Which practice most directly reduces confirmation bias?

A. Interviewing only enthusiastic early adopters

B. Changing the research question after every positive interview

C. Seeking disconfirming cases and documenting contradictory evidence

D. Showing the preferred prototype at the beginning of every interview

Correct Answer: C. Seeking disconfirming cases and documenting contradictory evidence

Question 8

What does a negative case contribute to customer discovery?

A. It proves that the entire market is unattractive.

B. It helps identify the boundaries of the problem and refine segmentation.

C. It should always be removed because it conflicts with the main pattern.

D. It replaces the need for further interviews.

Correct Answer: B. It helps identify the boundaries of the problem and refine segmentation.

Question 9

A campaign costs $1,500 and produces 5 customers. Using CAC = C ÷ N₂, what is the observed acquisition cost per customer?

A. $75

B. $150

C. $300

D. $7,500

Correct Answer: C. $300

Question 10

Which statement about AI-assisted analysis of customer interviews is most accurate?

A. AI-generated themes are independent evidence and do not require review.

B. AI eliminates researcher bias completely.

C. AI can assist transcription and coding, but outputs require human verification and privacy controls.

D. Confidential interview data can always be submitted to any external AI service.

Correct Answer: C. AI can assist transcription and coding, but outputs require human verification and privacy controls.