Design Thinking and Customer Discovery

Introduction

Entrepreneurship begins with an opportunity, but an opportunity becomes a viable venture only when it addresses a real and sufficiently important customer problem. Many new products fail not because they are poorly engineered, but because their creators misunderstand customers, solve a weak problem, or build features that people do not value. Design thinking and customer discovery provide practical methods for reducing these risks.

Design thinking is a human-centred approach to understanding needs, framing problems, generating possibilities, creating representations of solutions, and learning through testing. Customer discovery is the systematic process of identifying potential customers, investigating their behaviours and problems, and testing whether a proposed solution creates meaningful value. Together, these approaches help entrepreneurs move from assumptions to evidence.

The objective is not to eliminate uncertainty. Early-stage ventures operate under uncertainty. The objective is to make uncertainty visible, learn economically, and improve decisions before committing substantial resources.

Learning Objectives

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

1. Explain the principles and stages of design thinking.
2. Conduct basic empathy research using observation and interviews.
3. Define a customer problem without prematurely prescribing a solution.
4. Develop customer personas and evidence-based problem statements.
5. Generate, screen and prioritise ideas.
6. Create low-fidelity prototypes and design meaningful tests.
7. Conduct customer interviews and experiments ethically.
8. Interpret evidence relating to problem–solution fit.
9. Recognise limitations, biases and ethical risks in customer discovery.

1. Design Thinking as an Entrepreneurial Capability

Design thinking emerged from traditions in industrial design, participatory design and design research. It is commonly associated with an iterative process involving empathy, definition, ideation, prototyping and testing. These stages are useful as a learning structure, but they should not be treated as a rigid sequence. Entrepreneurs frequently return to earlier stages when new evidence changes their understanding of the problem.

Design thinking combines three forms of feasibility:

• Desirability: Do customers experience the problem and want the proposed improvement?
• Feasibility: Can the solution be built and delivered with available technology, skills and resources?
• Viability: Can the venture create sufficient economic value to sustain operations?

A technically feasible product may be undesirable. A desirable product may be too expensive to deliver. A viable business model may create unacceptable social or environmental consequences. Effective entrepreneurial design considers all three dimensions, together with safety, legality and ethical responsibility.

Core principles

Human-centredness. Decisions begin with the experiences, goals, constraints and contexts of people affected by the problem.

Learning through action. Direct observation, prototypes and experiments generally produce more useful learning than speculation alone.

Divergence and convergence. Teams first expand the range of possible interpretations and ideas, then narrow choices using explicit criteria.

Iteration. A prototype or test is not a final examination; it is an instrument for learning and improvement.

Tolerance for ambiguity. Early evidence is incomplete and sometimes contradictory. Entrepreneurs must remain open to revising assumptions.

Systems awareness. A customer problem may involve suppliers, regulators, families, employees, platforms and communities, not merely one end user.

2. The Design Thinking Process

Empathise: understand lived experience

Empathy research seeks to understand what people do, say, feel and struggle with in a particular context. It is more reliable to study actual behaviour than to ask people what they might do in a hypothetical future. Useful methods include:

• contextual observation;
• semi-structured interviews;
• diary studies;
• journey mapping;
• analysis of complaints, reviews and support records;
• shadowing or guided task completion.

A good observation records actions, sequence, time, workarounds and environmental constraints. An interview should explore recent experiences rather than invite general opinions. For example, “Tell me about the last time you missed a medical appointment” is usually more informative than “Would you use an appointment reminder application?”

Researchers should distinguish observations from interpretations. “The participant checked three messaging applications before confirming delivery” is an observation. “The participant is disorganised” is an interpretation that may be unsupported or unfair.

Define: frame the problem

Problem definition converts a broad concern into a focused opportunity for investigation. A useful problem statement identifies a target group, context, unmet need and consequence without prescribing a particular product.

A practical structure is:

[Specific users] need a better way to [perform a task or achieve an outcome] when [relevant context], because [evidence-based difficulty or consequence].

For example: “Independent food vendors need a dependable way to forecast daily ingredient demand during irregular trading periods because inaccurate estimates cause avoidable waste and stock-outs.” This statement does not assume that the answer is an application, spreadsheet or consulting service.

A “how might we” question can support ideation: “How might we help independent food vendors make affordable, reliable purchasing decisions when demand is uncertain?” Such questions should be broad enough to permit alternative solutions but specific enough to guide research.

Ideate: generate alternatives

Ideation separates the generation of possibilities from their evaluation. Techniques include brainstorming, brainwriting, mind mapping, analogy, role reversal, morphological analysis and “worst possible idea” exercises. Teams should first produce a range of alternatives, including service, process, partnership and behavioural solutions—not only physical products or software.

Ideas can be screened using criteria such as customer value, reach, feasibility, cost, differentiation, risk and alignment with the venture’s capabilities. A scoring matrix may be helpful, but numerical scores are judgements rather than objective facts. Teams should record the reasons behind scores and identify assumptions requiring testing.

Prototype: make the idea tangible

A prototype is a representation designed to answer a learning question. It may be a sketch, paper interface, storyboard, role-play, landing page, physical mock-up, service script or manually delivered service. The appropriate level of fidelity depends on the question.

If the question is whether users understand the basic workflow, a paper prototype may be sufficient. If the question concerns physical comfort, a three-dimensional mock-up is more appropriate. If the question concerns willingness to pay, a realistic offer and purchasing process may be required.

A prototype should be deliberately incomplete. Building more detail than necessary increases cost and can cause teams to defend features that have not been validated.

Test: learn from evidence

Testing involves exposing a prototype, message or proposed behaviour to relevant people and recording what happens. Testing is not equivalent to asking whether people “like” an idea. Stronger evidence includes whether they attempt a task, provide access to existing information, commit time, sign up, make a purchase, return, recommend the solution or change behaviour.

After a test, the team should compare results with the original assumption. It may:

• persevere with the current direction;
• modify the customer segment, problem or solution;
• pause and conduct additional research; or
• pivot to a substantially different proposition.

3. Customer Discovery

Customer discovery is an evidence-seeking activity, not a sales presentation disguised as research. Its central question is whether a clearly identifiable group experiences a sufficiently important problem and whether the proposed solution may address it better than current alternatives.

Customers, users and beneficiaries

The user operates or experiences the solution. The customer may select or pay for it. A beneficiary gains value but may not participate in purchase or use. In a school transport service, for example, a child may be the user, a parent the customer and the school or community a beneficiary. Confusing these roles can produce inaccurate research and an incomplete business model.

Assumptions and hypotheses

Entrepreneurs should make assumptions explicit. Common assumptions concern:

• who has the problem;
• how often it occurs;
• how serious it is;
• what people currently do;
• who decides and who pays;
• what barriers prevent adoption;
• which benefit matters most;
• whether the venture can reach customers economically.

A hypothesis states an assumption in a form that can be investigated. For example: “Small independent retailers experience stock-counting errors at least weekly and would allocate staff time or money to reduce them.” A useful hypothesis specifies a population, behaviour, frequency and consequence.

Segmentation

Segmentation divides a broad market into groups with meaningful differences. Segments may be based on role, location, industry, behaviour, life situation, purchasing authority or intensity of need. Demographic description alone is often insufficient. Two customers of the same age may have very different problems because their work, resources and contexts differ.

An early venture should select a narrow initial segment in which the problem is especially urgent. This does not mean the eventual market must remain narrow; it means learning is easier when the initial population is clearly defined.

Customer journey mapping

A journey map describes stages before, during and after a task or service. It can show goals, actions, emotions, pain points, touchpoints, delays and opportunities. The map should be based on evidence from customers rather than the team’s assumptions. It is particularly useful for service businesses because value may be created across multiple interactions.

4. Personas and Problem Statements

A persona is a research-based representation of a significant customer type. It is not a fictional biography created from stereotypes. A practical persona includes:

• role and relevant context;
• goals and desired outcomes;
• behaviours and routines;
• constraints and resources;
• frustrations and workarounds;
• decision criteria;
• purchasing authority;
• representative evidence and quotations;
• implications for design.

A persona should state its evidence base. For example, “Based on eight interviews and two weeks of observation” is more responsible than presenting unsupported demographic details as fact. Teams should avoid treating one person as representative of an entire market.

A strong problem statement describes a meaningful unmet need and its consequences. It should not include the proposed feature. “Students need an AI timetable” is a solution statement. “Students who combine employment and study need a reliable way to identify workable weekly schedules because timetable conflicts are discovered too late” is a problem statement.

5. Conducting Effective Interviews

Preparing the interview

Before interviewing, define the learning objective, target participant, recruitment method, consent process and note-taking procedure. Explain the purpose honestly. Participation should be voluntary, and sensitive information should not be collected unless necessary.

A semi-structured guide provides consistency while allowing probing. Questions should focus on past or current behaviour:

• “Walk me through the last time this happened.”
• “What did you do first?”
• “What alternatives did you consider?”
• “What was difficult or costly?”
• “How did you solve it?”
• “Who else was involved?”
• “What happened afterwards?”

Avoid leading questions such as “Wouldn’t an automated system save you time?” Avoid selling during discovery. Praise and enthusiasm from participants can be polite rather than predictive of adoption.

Recording and analysing data

With permission, interviews may be recorded; otherwise, take detailed notes. Immediately after each session, capture observations, surprises, quotations, contradictions and new questions. Coding can group evidence into themes such as delays, trust, cost, workarounds and decision authority.

A useful evidence table distinguishes the following:

The number of interviews should not be selected mechanically. Continue until additional interviews produce little new insight for the specific research question, while remaining alert to differences between segments. Small samples can reveal important problems but cannot automatically establish market size or statistical prevalence.

6. Experiments and Problem–Solution Fit

An experiment links an assumption to an observable test and a decision rule. Its components are:

1. Hypothesis: what the team believes.
2. Target participants: who will be studied.
3. Method: how the proposition will be exposed.
4. Measure: what behaviour or result will be observed.
5. Threshold: what result supports action.
6. Decision: persevere, revise, pivot or stop.

Examples include a concierge test, where the service is delivered manually; a landing-page test, where visitors respond to a clear proposition; a prototype usability test; a pre-order; or a limited pilot with a defined customer group.

Suppose 40 relevant visitors view a landing page and 8 request a trial. The observed conversion rate is:

Thus, the conversion rate is 8 ÷ 40 × 100 = 20%. This result is descriptive, not proof of future success. Its interpretation depends on how visitors were recruited, whether the offer was credible, whether duplicate or accidental actions were excluded, and what happens after the trial request.

If a team sets a threshold of 15% qualified trial requests for a first test, the result exceeds that threshold. The appropriate conclusion is that the proposition merits further investigation—not that product–market fit has been proven.

Problem–solution fit is early evidence that a defined customer group has an important problem and that a proposed solution plausibly addresses it. It precedes broader market validation, operational scaling and sustainable business-model validation.

7. A Practical Workflow for Student Entrepreneurs

A disciplined discovery cycle can be organised as follows:

1. Write a one-page assumption map.
2. Identify the riskiest assumptions, especially those concerning customer urgency and payment.
3. Recruit participants who genuinely experience the problem.
4. Conduct interviews and observations without pitching.
5. Synthesise evidence into themes and a provisional persona.
6. Write several alternative problem statements.
7. Select one problem based on urgency, frequency, reachability and strategic fit.
8. Generate multiple solution concepts.
9. Build the cheapest prototype that can answer the next important question.
10. Define a behavioural experiment and decision threshold.
11. Run the test, document results and update assumptions.
12. Repeat until evidence supports a focused proposition or indicates the need to pivot.

A research log should record date, participant type, assumption tested, evidence, interpretation, confidence and next action. This prevents selective memory and helps teams distinguish evidence from enthusiasm.

8. Illustrative Case: FreshRoute for Independent Food Vendors

Background

FreshRoute is a fictional student venture exploring a service for independent food vendors who purchase ingredients daily. The initial idea was a mobile application that would forecast demand and automatically recommend purchases.

Problem

The team initially assumed that vendors needed more sophisticated forecasting software. Six interviews revealed a more complex situation. Vendors used notebooks, messaging groups and personal experience. Their main difficulties were not only forecasting: supplier minimum orders, unreliable delivery times, fluctuating events and limited cash created risk. Several vendors said they would not trust an automated recommendation without understanding its basis.

Analysis

Observation of three vendors showed that purchasing decisions were made quickly, often while serving customers. The team mapped the journey from estimating demand to ordering, receiving ingredients and disposing of unused stock. The most urgent pain point was a lack of a simple way to adjust purchases when local events or weather changed expected demand.

The team revised its persona to “Amina, an independent lunch vendor operating with two assistants, limited storage and variable weekday demand.” This persona was based on interviews and observations, not merely demographic assumptions.

Approach and solution

Instead of immediately building an application, the team created a spreadsheet and manually sent daily purchasing suggestions through a messaging channel. Vendors could change assumptions and report actual sales. The team tested whether vendors would provide yesterday’s sales data and use a recommendation before ordering.

The first pilot involved five vendors for two weeks. Three supplied data consistently, two used the recommendations several times, and one vendor paid a small fee for continued support. Participants valued the adjustment process and explanations more than an automated forecast. The team therefore considered a service-assisted purchasing tool rather than a fully automated application.

Outcome

The evidence did not validate the original software concept. It did reveal a narrower problem, a reachable customer group and a promising service behaviour. Further tests were required to assess savings, reliability, willingness to pay and scalability.

Learning points

• The original solution was not the original problem.
• Workarounds revealed both pain and customer constraints.
• A manual prototype generated learning at low cost.
• Payment was stronger evidence than positive comments, but one payment was insufficient to establish a business.
• Trust and explainability were part of the value proposition.

9. Challenges, Ethics and Limitations

Bias and weak evidence

Recruiting friends, highly enthusiastic early adopters or people who receive incentives can distort results. Interviewer expectations may influence questions and interpretation. Confirmation bias causes teams to remember supporting evidence while discounting contradictions. Use neutral questions, varied participants, shared coding and explicit records of disconfirming evidence.

Privacy and informed consent

Customer research may involve personal, financial, health or location data. Participants should know what is collected, why it is needed, how it will be stored, who can access it and when it will be deleted. Collect the minimum data necessary. Anonymise reports where possible and secure files appropriately. Sensitive research may require institutional or legal approval.

Vulnerable participants

Children, patients, people experiencing poverty and employees subject to managerial pressure may be vulnerable to coercion or harm. Researchers should use appropriate consent procedures, avoid exploitative incentives and ensure that refusal has no penalty.

Deception and dark patterns

A landing page should not falsely claim that a product exists or imply scarcity that is not real. Prototypes should be clearly described as prototypes when misunderstanding could cause harm. Design should not manipulate users into sharing data, making unwanted purchases or accepting unfair terms.

Limits of design thinking

Design thinking can become superficial if it is reduced to sticky notes and workshops without serious research. It may overemphasise individual experiences while ignoring structural inequality, regulation, environmental effects or power relationships. Customer preference alone does not justify a harmful solution. Entrepreneurs must combine human-centred insight with technical analysis, financial discipline, legal compliance and social responsibility.

10. Technology and AI in Customer Discovery

Digital tools can assist transcription, coding, survey administration, prototype production and pattern identification. Generative AI can help produce interview guides, alternative hypotheses and rough interface concepts. However, AI-generated interpretations may reproduce bias, invent themes or expose confidential data to third parties. Human researchers must verify summaries against source material, protect personal information and disclose appropriate uses.

AI should support—not replace—direct contact with customers. Automated sentiment scores, click-through rates and generated personas can create false confidence when the underlying sample is biased or the measure is poorly connected to value. The key question remains: what decision will this evidence improve?

Best-Practice Checklist

Before concluding a discovery cycle, ask:

• Have we identified a specific customer segment?
• Have we observed or discussed recent real behaviour?
• Have we separated users, customers and beneficiaries?
• Is our problem statement free of a predetermined solution?
• What evidence contradicts our assumptions?
• Does the prototype answer a clearly stated question?
• Are we measuring behaviour rather than approval alone?
• Is our decision rule defined before seeing the result?
• Have we protected participants and their data?
• What will we do next if the evidence is weak?

Chapter Summary

Design thinking is an iterative, human-centred approach to entrepreneurial problem solving. Its main activities are empathy research, problem definition, ideation, prototyping and testing. Customer discovery complements design thinking by systematically examining customer segments, behaviours, needs, current workarounds, purchasing roles and willingness to adopt.

Effective research begins with observation and open questions about real experiences. Evidence is then synthesised into personas, journey maps and problem statements that describe needs without prematurely specifying solutions. Ideas should be generated broadly and evaluated against explicit criteria. Prototypes should be inexpensive and designed to answer particular learning questions. Experiments become more useful when they specify a hypothesis, participant group, measure, threshold and decision.

Problem–solution fit is early evidence that an important problem exists for a defined group and that a proposed solution may address it. It is not the same as product–market fit or business success. Ethical discovery requires informed consent, privacy protection, honest communication and special care with vulnerable participants. AI and digital tools can improve efficiency but cannot substitute for responsible human judgement and direct customer understanding.

Key Terms

Design thinking: A human-centred and iterative approach to framing problems and developing solutions.

Customer discovery: The process of investigating customer problems, behaviours, needs and responses to proposed solutions.

Desirability: The extent to which a solution is wanted or valued by users or customers.

Feasibility: The extent to which a solution can be built and delivered with available resources.

Viability: The extent to which a solution can support a sustainable venture.

Empathy research: Investigation of people’s experiences, behaviours, goals and constraints.

Problem statement: A focused description of a customer need and its context without prescribing a solution.

Ideation: The deliberate generation of alternative solution concepts.

Prototype: A representation used to learn about a concept or interaction.

Low-fidelity prototype: A quick, inexpensive representation with limited detail.

Experiment: A planned test linking an assumption to observable evidence and a decision.

Customer segment: A group of potential customers sharing relevant needs or characteristics.

Persona: An evidence-based representation of a significant customer type.

Customer journey map: A visual or structured account of stages, actions and experiences in a process.

Workaround: A solution customers currently use when preferred options are unavailable or inadequate.

Problem–solution fit: Early evidence that a meaningful problem and plausible solution are aligned.

Conversion rate: The percentage of eligible people who perform a specified desired action.

Pivot: A significant change in customer segment, problem, solution or business direction based on learning.

Informed consent: Voluntary agreement to participate after receiving relevant information about the research.

Confirmation bias: The tendency to favour evidence that supports existing beliefs.

Review Questions

Short Answer Questions

1. What is the difference between desirability, feasibility and viability?
2. Why should a problem statement avoid prescribing a solution?
3. Identify two methods of empathy research.
4. What is the purpose of a prototype?
5. Distinguish a user from a customer.

Long Answer Questions

1. Explain the stages of design thinking and describe why the process is iterative rather than strictly linear.
2. Discuss how a student entrepreneur can plan and conduct a semi-structured customer interview.
3. Explain how personas and customer journey maps support opportunity identification.
4. Describe the components of a well-designed customer discovery experiment.
5. Discuss ethical risks in customer research and recommend practical safeguards.

Analytical / Application Questions

1. A team proposes an application for students who miss assignment deadlines. Design an interview plan, identify three assumptions and write a solution-neutral problem statement.
2. A prototype receives enthusiastic verbal feedback, but no participant signs up for a trial. Analyse at least four possible explanations and recommend the next experiment.
3. A venture observes 12 purchases among 60 qualified visitors to a pilot offer. Calculate the conversion rate, explain what the result does and does not establish, and propose two follow-up tests.

Multiple Choice Questions with Answers

Question 1

Which question best supports customer discovery research?

A. Would you buy our application if we added more features?

B. What happened the last time you experienced this problem?

C. Do you agree that our solution is innovative?

D. Would you recommend our idea to your friends?

Correct Answer: B. What happened the last time you experienced this problem?

Question 2

Which statement is a solution-neutral problem statement?

A. Students need an AI timetable.

B. Vendors need a forecasting application.

C. Students combining work and study need workable schedules because conflicts are discovered too late.

D. Customers need a dashboard with predictive analytics.

Correct Answer: C. Students combining work and study need workable schedules because conflicts are discovered too late.

Question 3

What is the primary purpose of a prototype?

A. To make the venture appear fully established

B. To answer a learning question at an appropriate cost

C. To guarantee product–market fit

D. To replace all customer interviews

Correct Answer: B. To answer a learning question at an appropriate cost

Question 4

Which provides stronger evidence of customer value than a positive opinion?

A. A participant says the idea is interesting

B. A participant compliments the researcher

C. A participant makes a relevant purchase or commitment

D. A participant agrees with the team’s description

Correct Answer: C. A participant makes a relevant purchase or commitment

Question 5

A landing page receives 15 trial requests from 75 eligible visitors. What is the conversion rate?

A. 5%

B. 15%

C. 20%

D. 60%

Correct Answer: C. 20%

Question 6

Which item should normally be included in an experiment plan?

A. A guarantee that the idea will succeed

B. A decision threshold established before interpreting results

C. Only positive customer quotations

D. A complete final product specification

Correct Answer: B. A decision threshold established before interpreting results

Question 7

What is a research-based persona?

A. A fictional character based entirely on the founder’s assumptions

B. A demographic profile copied from a competitor

C. A representation of a significant customer type supported by evidence

D. A list of all possible market segments

Correct Answer: C. A representation of a significant customer type supported by evidence

Question 8

Which practice reduces confirmation bias in customer discovery?

A. Interviewing only enthusiastic friends

B. Ignoring contradictory evidence

C. Recording disconfirming findings and using neutral questions

D. Changing the results to match the original hypothesis

Correct Answer: C. Recording disconfirming findings and using neutral questions

Question 9

What does problem–solution fit indicate?

A. The venture has achieved guaranteed profitability

B. A defined group has an important problem and the solution may address it

C. The product is ready for global scaling

D. All competitors have been eliminated

Correct Answer: B. A defined group has an important problem and the solution may address it

Question 10

Which is an appropriate ethical practice in customer research?

A. Collecting unnecessary personal information

B. Using hidden pressure to obtain participation

C. Explaining data use and obtaining voluntary informed consent

D. Presenting a prototype as an available product when it is not

Correct Answer: C. Explaining data use and obtaining voluntary informed consent