Market Intelligence, Customer Discovery, and Digital Research

Introduction

Digital marketing decisions are only as strong as the market understanding behind them. Startups frequently fail not because they cannot build products or attract attention, but because they solve problems that customers do not consider important, target audiences that are too broad, or interpret weak signals as evidence of demand. Market intelligence and customer discovery reduce these risks by systematically investigating customers, contexts, competitors, conversations, and behavioral evidence.

For an early-stage venture, digital research is not merely a promotional activity. It is a method for learning. Search queries reveal expressed needs; online communities reveal language and frustrations; competitor reviews reveal deficiencies; interviews reveal motivations and workarounds; experiments reveal whether stated interest becomes action. The entrepreneurial challenge is to combine these sources without confusing popularity with demand or correlation with causation.

This chapter develops an integrated approach to customer discovery and digital market research. It examines jobs-to-be-done, design thinking, social listening, search intelligence, competitor analysis, primary research, experimentation, data ethics, and managerial interpretation. The emphasis is on evidence that can guide product design, positioning, channel selection, and resource allocation.

Learning Objectives

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

1. Explain the role of market intelligence in entrepreneurial decision-making.
2. Distinguish customer segments, personas, problems, needs, and jobs-to-be-done.
3. Design customer-discovery research using interviews, observation, surveys, and digital behavioral evidence.
4. Apply design-thinking and jobs-to-be-done frameworks to identify opportunities.
5. Use search intelligence, social listening, and competitor analysis critically.
6. Evaluate evidence using validity, reliability, bias, and triangulation.
7. Design ethical and legally responsible digital research.
8. Translate research findings into testable value propositions and marketing experiments.

1. Market Intelligence as an Entrepreneurial Capability

Market intelligence is the systematic collection, analysis, interpretation, and use of information about customers, competitors, markets, technologies, and the wider environment. It differs from data accumulation. Data become intelligence only when they are relevant to a decision and interpreted in context.

A startup may need intelligence about several questions:

• Which customer experiences are sufficiently painful to motivate change?
• How do customers currently solve the problem?
• Who influences the purchase decision?
• What language do customers use when describing the problem?
• What alternatives and competitors already exist?
• Which claims, channels, and price points are credible?
• What evidence would justify continuing, changing, or abandoning an idea?

Market intelligence is particularly important under conditions of uncertainty. Established firms can sometimes compensate for imperfect assumptions through brand strength, distribution, and financial reserves. Startups generally cannot. Research therefore functions as a form of risk management and as a mechanism for disciplined learning.

Evidence and the hierarchy of confidence

Research evidence differs in proximity to actual behavior. A customer saying, “I would probably buy this,” is useful but weaker than a customer paying, signing up, referring another user, or switching from an existing solution. A practical hierarchy is:

1. Opinions and attitudes: what people say they like or intend to do.
2. Reported behavior: what people say they have done.
3. Observed behavior: what people actually do in a relevant context.
4. Commitment behavior: actions involving time, money, reputation, or access.
5. Repeated behavior: sustained use, repurchase, retention, or referral.

No level is irrelevant. Attitudes help explain behavior, while behavioral evidence without interpretation can be misleading. Strong entrepreneurial research combines both.

2. Defining the Customer Problem

Segments, personas, and markets

A market segment is a group whose members share characteristics relevant to a particular offering or decision. Segmentation may use demographic, geographic, psychographic, behavioral, firmographic, or needs-based variables. A segment is strategically useful when it is sufficiently identifiable, reachable, substantial, responsive, and compatible with the venture’s capabilities.

A persona is a research-based representation of a meaningful customer pattern. It should summarize goals, constraints, behaviors, contexts, and decision criteria rather than inventing decorative details. Personas become harmful when they replace evidence with stereotypes.

A target market is not simply “everyone who could benefit.” It is the group a venture chooses to serve initially, given its resources and strategic position. Narrow targeting often improves learning because the team can observe a more coherent set of needs and buying processes.

Problems, needs, and opportunities

A problem is an undesirable condition or outcome. A need is a requirement or desired state. An opportunity exists when a meaningful problem is inadequately addressed and a venture can deliver superior value in a feasible and defensible way.

The same visible problem may conceal different causes. For example, a small retailer may appear to need “social media content,” but interviews might reveal that the deeper issue is uncertainty about what content generates store visits, lack of time, or inability to measure results. Each cause implies a different product and marketing strategy.

A useful problem statement identifies:

• the specific actor;
• the situation or trigger;
• the difficulty experienced;
• the consequences;
• the current workaround; and
• why existing alternatives are insufficient.

A strong problem statement avoids embedding a solution. “Independent retailers need an automated video platform” is a product assumption. “Independent retailers struggle to produce credible promotional content consistently when staff time is limited” is an investigable problem statement.

3. Jobs-to-be-Done and Customer Motivation

The jobs-to-be-done (JTBD) perspective argues that customers “hire” products or services to make progress in a particular circumstance. The job is not identical to a product category. A person may hire a meal-delivery service to reduce decision effort on a busy evening, not merely to obtain food.

A job statement usually includes:

When [situation], I want to [motivation], so I can [desired outcome].

For rigorous analysis, distinguish three dimensions:

• Functional job: the practical task to be accomplished.
• Emotional job: how the customer wants to feel.
• Social job: how the customer wants to be perceived.

JTBD is valuable because it prevents overly narrow competitor definitions. A startup competing with a project-management application may also compete with spreadsheets, email, meetings, memory, or doing nothing. These are alternative ways of accomplishing the job.

However, JTBD should not be treated as a universal law. Jobs are interpreted through culture, identity, institutional rules, income, and technology. A job statement is a research hypothesis that must be tested across contexts.

Switching forces

Customer adoption can be analyzed through four forces:

• the push of dissatisfaction with the current situation;
• the pull of the new solution;
• the anxiety or perceived risk of switching; and
• the habit or inertia supporting the current solution.

A compelling value proposition must overcome both dissatisfaction and switching resistance. Better functionality alone may not be enough if migration is difficult, trust is low, or the existing workaround is cheap.

4. Design Thinking as a Discovery Process

Design thinking provides a human-centered process for moving from ambiguous problems to tested solutions. A commonly used sequence is empathize, define, ideate, prototype, and test. In practice, the process is iterative rather than linear.

Empathize

Researchers seek to understand experiences in context through interviews, observation, diary studies, support records, reviews, and community discussions. The goal is not to agree with customers but to understand what they do, feel, fear, and prioritize.

Define

The team synthesizes evidence into a focused problem framing. Affinity mapping, journey maps, and cause analysis help identify recurring patterns. A good definition is specific enough to guide action but open enough not to dictate a solution.

Ideate

The team generates multiple possible responses. Premature evaluation should be avoided because the first solution is often shaped by the team’s existing technical capabilities rather than customer evidence.

Prototype

A prototype may be a sketch, landing page, clickable interface, service script, advertisement, concierge service, or manual version of an automated product. The appropriate prototype is the cheapest artifact that can test the riskiest assumption.

Test

Testing should specify what will count as evidence. For example, “at least 20% of qualified visitors request a demonstration” is more useful than “people respond positively.” The metric must match the decision and the stage of the venture.

Design thinking contributes empathy and experimentation, but it does not eliminate commercial analysis. A desirable solution must also be feasible to deliver and viable as a business.

5. Designing Customer-Discovery Research

Research questions before research methods

Research should begin with decisions, not tools. A team should ask: “What decision will this evidence inform?” Examples include selecting a beachhead segment, prioritizing a feature, changing a message, or determining whether to continue development.

A research brief should state:

• the decision or uncertainty;
• the research objectives;
• the population or context of interest;
• the method and recruitment plan;
• the evidence required;
• ethical safeguards; and
• how findings will be analyzed and acted upon.

Qualitative methods

Semi-structured interviews allow consistent coverage while permitting unexpected insights. Questions should focus on recent, concrete events rather than hypothetical preferences. Useful prompts include: “Tell me about the last time this happened,” “What did you do next?” and “What made that difficult?”

Avoid leading questions such as “Would you use our convenient platform?” Instead ask about current behavior, alternatives, costs, delays, and consequences. Interviewers should distinguish facts, interpretations, emotions, and assumptions.

Observation and contextual inquiry reveal workarounds that customers may forget or consider too ordinary to mention. Diary studies capture experiences over time, especially when problems are episodic. User-generated reviews and community posts can expose recurring language and unmet needs, but they are not automatically representative.

Quantitative methods

Surveys are useful for measuring the prevalence of patterns identified qualitatively. They are weaker for discovering unknown problems because response options constrain what can be expressed. Questionnaire quality depends on sampling, wording, order, scale design, completion conditions, and nonresponse.

For a proportion estimated from a simple random sample, an approximate margin of error at a stated confidence level can be expressed as:

Here, ME is the margin of error, z is the critical value associated with the chosen confidence level, p is the estimated population proportion, and n is the sample size. This approximation assumes a probability sample, independent observations, and a sufficiently large sample for the normal approximation. For a 95% confidence level, z is approximately 1.96. When p is unknown, 0.5 is often used because it produces the largest variance and therefore a conservative sample-size estimate. This formula does not correct for coverage error, nonresponse bias, poorly worded questions, or convenience sampling.

For example, if p = 0.5 and n = 400, then:

The approximate margin of error is 4.9 percentage points under the stated assumptions. It would be incorrect to claim that any convenience sample of 400 has this margin of error.

Sampling and saturation

Probability sampling supports statistical generalization when properly implemented. Purposive sampling is often more appropriate in early discovery because the goal is theoretical relevance and variation rather than population estimation. Researchers may deliberately recruit novices, experienced users, non-users, and customers who abandoned alternatives.

Qualitative saturation is reached when additional data produce little meaningful conceptual novelty, although saturation is not a mechanical number. The quality of the sample, research question, and analysis matters more than claiming a universal interview count.

6. Social Listening and Search Intelligence

Social listening

Social listening is the systematic monitoring and analysis of public or legitimately accessible digital conversations. It can identify recurring complaints, emerging vocabulary, influencers, competitor perceptions, and changes in sentiment.

A robust listening process includes:

1. Define the decision and topic boundaries.
2. Develop keyword groups, including synonyms, misspellings, product names, competitor names, and problem language.
3. Exclude irrelevant meanings and duplicate content.
4. Classify posts by topic, context, sentiment, urgency, and customer type.
5. Examine examples rather than relying only on aggregate scores.
6. Compare findings across platforms and time periods.
7. Convert patterns into research hypotheses.

Sentiment analysis can be useful for large-scale screening, but sarcasm, mixed emotions, domain-specific language, multilingual content, and unequal platform participation reduce accuracy. Sentiment volume is not the same as market size. A highly vocal minority may dominate online conversation.

Search intelligence

Search data show what people choose to express as information needs. Keyword volume can indicate attention, but it is not equivalent to purchase intent. A query such as “how to repair a leaking tap” may indicate information seeking, while “emergency plumber near me” may indicate immediate commercial intent. Context, geography, seasonality, and query progression matter.

Useful search dimensions include:

• informational, navigational, commercial, and transactional intent;
• problem, solution, brand, and comparison queries;
• geographic modifiers;
• questions and autocomplete language;
• seasonality and trend direction; and
• search-result competitors.

Search intelligence should be triangulated with interviews and conversion experiments. A large volume of searches may reflect a problem that users expect to solve free of charge, while a small but urgent query category may be commercially attractive.

7. Competitor and Alternative Analysis

Competitor analysis should include direct competitors, indirect competitors, substitute behaviors, and non-consumption. A useful comparison examines the customer job rather than only product features.

A feature matrix is useful but insufficient. Two firms may offer similar features while differing substantially in trust, implementation support, distribution, brand legitimacy, integration, or total cost of ownership.

A positioning map can compare alternatives on two customer-relevant dimensions, but the dimensions should be supported by research rather than chosen merely because they make the startup appear unique. Competitor analysis should also avoid unethical practices such as unauthorized access, deception, scraping restricted data, or misrepresenting the firm’s identity.

8. From Evidence to Insight

Research analysis requires disciplined movement from raw material to action. One practical structure is:

• Observation: what was directly said, measured, or observed.
• Pattern: a recurrence across people, situations, or sources.
• Interpretation: a reasoned explanation of the pattern.
• Implication: what the venture should consider doing.
• Test: what evidence would support or challenge the implication.

For example, an observation may be that restaurant owners export sales data into spreadsheets every Monday. The pattern may be that several owners do this because their existing systems do not provide a clear weekly view. The interpretation is that reporting effort, rather than data collection, is the pain point. The implication is to test a reporting service. The test might measure whether qualified owners connect their data or pay for a manual reporting pilot.

Triangulation

Triangulation compares evidence from different methods, sources, researchers, or theories. Convergence increases confidence, while divergence is analytically valuable. If interviews describe high frustration but usage logs show little engagement, researchers should investigate whether the problem is important but poorly served, or whether interviewees were polite or atypical.

A simple prioritization score can help compare opportunities, but it is a decision aid rather than a scientific truth. For example, teams may rate problem severity, frequency, willingness to pay, reachability, and strategic fit on a common scale. The scoring criteria and weights must be documented, and qualitative reasoning should accompany the total.

9. Digital Experiments and Validation

A landing-page experiment can test whether a message attracts qualified interest. The page should specify the audience, problem, promise, call to action, and qualification mechanism. Traffic sources must be recorded because conversion rates differ by intent and audience.

A basic conversion rate is:

where CR is conversion rate, C is the number of desired conversions, and V is the number of eligible visitors or exposures, defined consistently. The metric is meaningful only when the denominator, traffic quality, time period, and conversion definition are clear.

Suppose 60 of 1,000 qualified visitors request a demonstration. Then:

This does not establish product-market fit. It indicates response to a particular message and offer under particular conditions. Further evidence is required: attendance, activation, payment, retention, and customer outcomes.

A minimum viable product should be designed around the riskiest assumption. A manual “concierge” service may test whether customers value an outcome before engineering automation. A fake-door test can measure interest in a proposed feature, but it must clearly disclose availability and must not collect payment deceptively.

10. Ethical, Legal, and Governance Issues

Digital research involves people, identities, opinions, behavioral traces, and sometimes sensitive information. Ethical practice requires more than legal compliance.

Core principles include:

• Purpose limitation: collect information for a defined, legitimate purpose.
• Data minimization: collect only what is necessary.
• Transparency: explain who is collecting data, why, and how it will be used.
• Consent and choice: obtain valid consent where required and provide meaningful refusal or withdrawal mechanisms.
• Confidentiality: restrict access and remove identifying details where possible.
• Security: protect data through appropriate technical and organizational controls.
• Fairness: assess whether recruitment, models, or targeting exclude or disadvantage groups.
• Retention control: delete or anonymize data when it is no longer needed.

Legal requirements differ by jurisdiction and may include privacy, consumer-protection, marketing-communications, copyright, platform terms, and research-ethics rules. Public availability does not automatically mean unrestricted ethical use. A post in an online support community may be publicly visible but contextually private. Researchers should avoid quoting identifiable content without a defensible basis, especially when discussing health, financial difficulty, children, political beliefs, or other sensitive matters.

AI tools can accelerate transcription, clustering, summarization, sentiment classification, and synthetic concept generation. They can also hallucinate, reproduce bias, expose confidential information, and obscure the source of a conclusion. Human researchers must verify outputs against source material, document prompts and transformations, protect confidential data, and avoid presenting generated text as participant evidence. AI-generated personas or simulated interviews are hypotheses, not substitutes for real customer research.

11. Applied Case: GreenCart’s Subscription Hypothesis

Background

GreenCart, an illustrative early-stage venture, proposed a subscription service delivering low-waste household goods to urban professionals. The founders initially targeted “environmentally conscious consumers” and planned to acquire them through social advertising.

Problem

Early advertisements produced clicks but few purchases. The team assumed the issue was insufficient environmental messaging. Before increasing the advertising budget, it conducted discovery research.

Analysis

The team interviewed 18 potential customers, reviewed 600 public product reviews across competing services, analyzed search queries, and observed how six households purchased cleaning products. The evidence showed three patterns:

1. Participants supported waste reduction but were frustrated by the inconvenience of refills and uncertain product performance.
2. Competitor reviews frequently mentioned delayed deliveries, excessive packaging, and unclear cancellation procedures.
3. Search behavior included substantial interest in “best refill products,” but commercial queries were concentrated around convenience, price, and allergy safety rather than environmental identity.

The original segment was too broad. The most promising early group was urban households with recurring purchases, limited time, and a demonstrated willingness to buy premium allergy-conscious products.

Approach and solution

Using JTBD, the team reframed the job as: “When household supplies are running low, help me replenish reliable, allergy-safe products without researching options or managing complicated subscriptions.” It tested three propositions through separate landing pages: ecological impact, convenience, and allergy-safe reliability.

The prototype was a manual monthly replenishment service. Customers selected a product profile, received a curated order, and could pause or change deliveries through a simple messaging channel. The team measured qualified sign-ups, completed first orders, reasons for cancellation, and repeat orders.

Outcome

The convenience-and-safety message generated fewer total visits than the environmental message but a higher proportion of qualified sign-ups and completed orders. Interviews after delivery revealed that transparent substitution rules and easy pausing were more persuasive than additional sustainability claims. The venture therefore postponed full automation, narrowed its initial segment, improved cancellation controls, and used environmental benefits as supporting rather than primary positioning.

Learning points

The case demonstrates that attention is not the same as demand, and that customers may endorse a social value while buying for a more immediate functional reason. It also shows the value of combining search intelligence, competitor reviews, interviews, observation, and behavioral experiments. Most importantly, research changed not only the advertising message but also the service design.

Challenges and Limitations

Market research cannot remove uncertainty. Respondents may misremember, seek social approval, or describe ideals rather than behavior. Online data overrepresent people who post, complain, or belong to particular platforms. Search trends may reflect curiosity rather than willingness to pay. Competitor reviews often come from unusually satisfied or dissatisfied users. Small samples can generate deep insight but should not be presented as population estimates.

Researchers also face confirmation bias. Founders may selectively notice evidence that supports their preferred idea. Pre-registering key questions, recording disconfirming evidence, using independent reviewers, and separating discovery from advocacy can improve rigor. Research must remain proportionate: excessive analysis can delay learning, while superficial research can create false confidence.

Best-Practice Research Workflow

A disciplined startup workflow is:

1. State the strategic decision and riskiest assumption.
2. Define the customer, context, and job without embedding a solution.
3. Gather diverse evidence through qualitative, quantitative, and behavioral methods.
4. Record source, date, method, limitations, and confidence for each finding.
5. Analyze observations into patterns, interpretations, and implications.
6. Seek disconfirming evidence and investigate contradictory signals.
7. Build the smallest ethical experiment that tests the critical assumption.
8. Define success and failure criteria before observing results.
9. Update the segment, value proposition, product, and channel strategy.
10. Repeat the cycle as new evidence changes the venture’s understanding.

Chapter Summary

Market intelligence converts information into decision-relevant understanding. Customer discovery begins with problems and contexts, not with a predetermined product. Segmentation identifies a strategically meaningful group, while personas summarize evidence about that group. Jobs-to-be-done broadens analysis from product categories to the progress customers seek, including functional, emotional, and social dimensions.

Design thinking supports iterative movement from empathy and problem definition to prototyping and testing. Interviews, observation, surveys, reviews, social listening, search intelligence, and experiments each provide different kinds of evidence and each has limitations. Triangulation improves confidence by comparing sources rather than relying on a single metric.

Competitor analysis must include substitutes and non-consumption, and digital experiments must be interpreted as evidence about a particular message or behavior—not automatic proof of product-market fit. Ethical research requires purpose limitation, transparency, consent where appropriate, security, fairness, and respect for contextual privacy. AI can support analysis but cannot replace primary research or human accountability. The central entrepreneurial discipline is to convert uncertain assumptions into explicit hypotheses and test them with increasingly strong evidence.

Key Terms

Market intelligence: Systematic collection and interpretation of market information for decision-making.

Customer discovery: Research used to understand customer problems, contexts, behavior, and alternatives.

Market segment: A group sharing characteristics relevant to an offering or decision.

Persona: A research-based representation of a meaningful customer pattern.

Problem statement: A focused description of an undesirable customer condition without prescribing a solution.

Jobs-to-be-done: A perspective that views customers as hiring solutions to make progress in particular circumstances.

Functional job: The practical task a customer seeks to accomplish.

Switching forces: Push, pull, anxiety, and habit influencing adoption or non-adoption.

Design thinking: An iterative human-centered approach involving empathy, definition, ideation, prototyping, and testing.

Social listening: Monitoring and analyzing digital conversations to identify patterns and signals.

Search intelligence: Analysis of search behavior, query language, intent, and trends.

Competitor analysis: Systematic examination of direct competitors, substitutes, and alternative behaviors.

Triangulation: Comparison of evidence from multiple sources or methods.

Qualitative saturation: The point at which additional qualitative data yield little conceptual novelty.

Conversion rate: The proportion of defined eligible exposures that produce a specified action.

Prototype: A representation or early version used to test an assumption.

Data minimization: Collecting only the personal data necessary for a defined purpose.

Confirmation bias: The tendency to favor evidence supporting existing beliefs.

Product-market fit: A condition in which a defined market repeatedly values and adopts a product sufficiently to support sustainable growth.

Review Questions

Short Answer Questions

1. What distinguishes market intelligence from raw market data?
2. Why should a problem statement avoid embedding a proposed solution?
3. Identify the functional, emotional, and social dimensions of a job-to-be-done.
4. What is the purpose of triangulation in customer research?
5. State two ethical principles relevant to digital research.

Long Answer Questions

1. Explain how design thinking and jobs-to-be-done can be combined in startup customer discovery.
2. Compare interviews, surveys, social listening, and search intelligence as research methods. Discuss the strengths and limitations of each.
3. Explain why competitor analysis should include substitutes, workarounds, and non-consumption.
4. Discuss how a startup can distinguish stated interest from meaningful demand.
5. Evaluate the opportunities and risks of using AI in digital market research.

Analytical / Application Questions

1. A startup receives high engagement on an environmental advertisement but very few purchases. Design a research plan to determine whether the problem concerns segment selection, positioning, trust, price, or the product itself.
2. Using the conversion-rate formula, calculate the conversion rate for 84 sign-ups from 1,400 qualified visitors. Explain what the result does and does not establish.
3. Develop a JTBD statement, interview guide, and low-cost experiment for a digital service intended to help graduate students manage research references.

Multiple Choice Questions with Answers

Question 1

Which statement best describes market intelligence?

A. The collection of as much market data as possible

B. The systematic interpretation of relevant information for decisions

C. The process of publishing advertisements to a target audience

D. A financial forecast based only on competitor revenue

Correct Answer: B. The systematic interpretation of relevant information for decisions

Question 2

Which question is most appropriate for a customer-discovery interview?

A. Would you definitely buy our convenient platform?

B. Do you agree that our product is innovative?

C. Tell me about the last time you encountered this problem.

D. Which of our proposed features do you prefer?

Correct Answer: C. Tell me about the last time you encountered this problem.

Question 3

In jobs-to-be-done analysis, a customer hires a product primarily to:

A. Copy the market leader’s feature set

B. Make progress in a particular circumstance

C. Increase the number of competitors

D. Eliminate the need for customer research

Correct Answer: B. Make progress in a particular circumstance

Question 4

Which is an example of a substitute in competitor analysis?

A. Only a firm selling an identical product

B. A customer’s existing workaround for accomplishing the job

C. A government census report

D. An unrelated social-media account

Correct Answer: B. A customer’s existing workaround for accomplishing the job

Question 5

What does social listening primarily analyze?

A. Private customer passwords

B. Public or legitimately accessible digital conversations

C. Only paid search advertisements

D. Internal accounting records

Correct Answer: B. Public or legitimately accessible digital conversations

Question 6

If 60 of 1,000 qualified visitors request a demonstration, what is the conversion rate?

A. 0.6%

B. 6%

C. 16.7%

D. 60%

Correct Answer: B. 6%

Question 7

Which evidence generally represents the strongest commitment to a startup’s proposed value?

A. A customer saying the idea sounds interesting

B. A customer liking a social-media post

C. A customer repeatedly paying for and using the service

D. A customer completing an unrelated survey

Correct Answer: C. A customer repeatedly paying for and using the service

Question 8

Which principle means collecting only the personal information necessary for a defined purpose?

A. Data minimization

B. Market expansion

C. Feature parity

D. Sentiment scoring

Correct Answer: A. Data minimization

Question 9

Why can high search volume be misleading as evidence of demand?

A. All searches are generated by existing customers

B. Search volume may reflect curiosity or free information seeking rather than willingness to pay

C. Search engines never record query language

D. Search volume is always lower than social engagement

Correct Answer: B. Search volume may reflect curiosity or free information seeking rather than willingness to pay

Question 10

What is the principal purpose of a minimum viable prototype?

A. To launch the final product with every planned feature

B. To test a critical assumption with the least costly credible artifact

C. To imitate a competitor without research

D. To maximize design complexity before customer contact

Correct Answer: B. To test a critical assumption with the least costly credible artifact