Customer Discovery, Market Insight and Growth Opportunities

Introduction

Growth rarely begins with a clever acquisition tactic. It begins with a defensible understanding of whom the venture serves, what customers are trying to accomplish, which problems are sufficiently important to solve, and where unmet demand exists. Customer discovery is therefore the foundation of growth hacking and scaling. It reduces the risk of investing heavily in a product, channel or market before the underlying demand has been demonstrated.

For an entrepreneurial venture, customer discovery is not merely the collection of opinions. It is a disciplined process of learning from observable behaviour, testing assumptions and translating market evidence into decisions. The process links qualitative inquiry—such as interviews and observation—with quantitative analysis, including segmentation, conversion measurement and market sizing.

This chapter develops an integrated approach to identifying attractive growth opportunities. It examines customer development, jobs-to-be-done, market segmentation, personas, problem validation and market sizing. It also considers the limitations and ethical risks of these methods, particularly when digital data and artificial intelligence are used to infer customer behaviour.

Learning Objectives

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

1. Explain the role of customer discovery in venture growth and scaling.
2. Apply the customer development process and the jobs-to-be-done perspective.
3. Construct useful market segments and evidence-based personas.
4. Design interviews and experiments for problem validation.
5. Estimate total addressable market, serviceable available market and serviceable obtainable market.
6. Evaluate growth opportunities using customer desirability, feasibility, viability and strategic fit.
7. Recognise methodological, ethical and governance limitations in customer research.

1. Customer Discovery as a Growth Foundation

Customer discovery is the systematic investigation of customer problems, contexts, behaviours, alternatives and willingness to adopt a solution. It is particularly important in early-stage ventures because founders normally begin with incomplete information and strong personal beliefs. These beliefs may concern the target customer, the severity of the problem, the expected value of a solution or the most effective route to market.

A useful distinction is between problem discovery and solution discovery. Problem discovery asks what customers struggle with and how they currently cope. Solution discovery asks whether a proposed product can solve the problem sufficiently well and whether customers will adopt and pay for it. Confusing the two often produces leading interviews in which founders ask customers to approve a predetermined product.

Customer discovery also differs from market research conducted solely to describe an existing market. In entrepreneurship, the purpose is not only to measure a market but to identify a promising opportunity under uncertainty. Evidence is therefore used iteratively:

• assumptions are stated;
• evidence is collected;
• assumptions are revised or rejected;
• a small experiment is designed;
• behavioural results guide the next decision.

A growth team should maintain an assumption register. Typical assumptions include:

The strongest evidence usually comes from behaviour rather than stated preference. Saying that a product is “interesting” is weak evidence. Repeatedly using a prototype, sharing internal data, introducing the founder to a decision-maker or paying for a pilot is stronger evidence.

2. Customer Development and the Learning Process

The customer development perspective treats a venture as a set of hypotheses that must be tested in the market. Its central insight is that product development and customer development should proceed together. A technically complete product can fail if it solves an unimportant problem, reaches the wrong segment or requires an unacceptable change in behaviour.

A practical customer-development cycle contains four stages:

2.1 Customer discovery

The venture identifies prospective customers, their circumstances, their current alternatives and the consequences of the problem. At this stage, the objective is learning rather than maximising sales volume.

2.2 Customer validation

The venture tests whether a repeatable sales or adoption process exists. Validation requires more than a few enthusiastic early users. It asks whether customers consistently recognise the value proposition, complete the adoption process and provide economically meaningful commitment.

2.3 Customer creation

Once the venture has evidence of a repeatable value proposition, it develops demand through marketing, partnerships, sales processes and referral systems. Growth efforts should be scaled only when the underlying customer problem and solution have sufficient evidence.

2.4 Company building

The venture formalises roles, systems, budgets, controls and performance management. Premature company building can create fixed costs and bureaucratic processes before product–market fit has been established.

This sequence is not strictly linear. New customer evidence may reveal that the segment is incorrect or that the problem has been misunderstood. A disciplined venture therefore uses a pivot-or-persevere decision: persevere when evidence supports the current assumptions; pivot when a critical assumption is contradicted but an alternative opportunity is credible.

3. Jobs-to-be-Done: Understanding the Customer’s Progress

The jobs-to-be-done (JTBD) perspective focuses on the progress a customer is trying to make in a particular circumstance. Customers do not simply buy product features; they “hire” products to perform functional, emotional and social jobs.

For example, a commuter may hire a bicycle-sharing service to complete a short journey quickly, avoid parking difficulties, feel independent and demonstrate environmental values. The product is only one possible means of achieving these outcomes.

A job statement should identify:

• the situation or trigger;
• the desired progress;
• constraints or anxieties;
• the current solution;
• the outcome by which success is judged.

A useful format is: “When [situation], I want to [motivation], so that [desired outcome].” This is not a final marketing slogan. It is a research hypothesis that must be tested against real behaviour.

JTBD analysis is valuable because it prevents excessive focus on demographic labels. Two customers of different ages may have the same job, while two customers with the same demographic profile may have different jobs. It also reveals competing alternatives. A customer’s alternative may be another software product, a spreadsheet, an employee, a delay, a workaround or a decision not to act.

Functional, emotional and social dimensions

The functional job concerns the practical task. The emotional job concerns how the customer wants to feel. The social job concerns how the customer wants to be perceived by others. A business analytics platform, for instance, may help a manager produce reports (functional), feel confident before a board meeting (emotional) and appear data-driven to colleagues (social).

The JTBD perspective does not replace segmentation. It complements it. Segmentation identifies groups with meaningful differences; JTBD explains the motivation and circumstances within or across those groups.

4. Market Segmentation and Selection of Attractive Customers

Segmentation divides a broad market into groups whose members share relevant characteristics and respond differently to a value proposition, channel or pricing model. The objective is not to create as many segments as possible. It is to identify segments that are sufficiently distinct, reachable and economically attractive.

Common bases of segmentation include:

A segment should be evaluated against at least five criteria:

1. Identifiability: Can the venture recognise who belongs to the segment?
2. Accessibility: Can the venture reach and serve the segment?
3. Substantiality: Is the opportunity large enough relative to the venture’s economics?
4. Responsiveness: Does the segment react differently to the proposed offer?
5. Strategic fit: Does serving it match the venture’s capabilities and objectives?

A small segment can be attractive if it has acute pain, high willingness to pay, low acquisition cost and strong referral potential. Conversely, a large segment can be unattractive if it is difficult to reach or dominated by powerful incumbents.

Beachhead markets

A beachhead market is a narrowly defined initial segment in which the venture can establish credibility and repeatable traction. It should be specific enough to support focused learning but large enough to provide meaningful growth. A beachhead is not necessarily the venture’s permanent market; it is the first market in which the business can win and learn.

5. Personas: Useful Representations, Not Fictional Truth

A persona is a research-based representation of a significant customer type. It generally includes context, goals, behaviours, constraints, decision criteria, alternatives and buying authority. A persona should be treated as a decision aid, not as a stereotype or a substitute for evidence.

A high-quality persona includes:

• role and relevant context;
• job-to-be-done;
• triggering events;
• current workflow and alternatives;
• desired outcomes;
• objections and perceived risks;
• information and purchasing channels;
• authority, budget and other stakeholders;
• evidence base and confidence level.

For a business-to-business product, it is important to distinguish the user, economic buyer, champion, approver and blocker. One person may occupy several roles, but often the product user is not the person who controls the budget.

Personas should be periodically updated. If a persona contains unsupported claims such as “prefers innovation” or “is price sensitive,” those claims should be converted into research questions. A persona may also include an evidence rating—high, medium or low—to make uncertainty visible.

6. Problem Validation and Evidence Quality

Problem validation tests whether a problem is real, important, frequent or expensive enough to justify action. A validated problem has more than emotional resonance; it is connected to observable consequences and a plausible willingness to change.

6.1 Interview design

Effective interviews are exploratory and behaviour-focused. Questions should ask about specific past events rather than hypothetical future intentions. For example:

• “Tell me about the last time this occurred.”
• “What did you do next?”
• “What alternatives did you consider?”
• “What did the problem cost in time, money or risk?”
• “Who else was involved in the decision?”
• “What have you already tried?”

Weak questions include “Would you use this?” and “Do you like our idea?” Such questions invite politeness and speculation. The interviewer should avoid pitching too early, leading the respondent or treating one unusually enthusiastic participant as representative.

6.2 Evidence hierarchy

A practical hierarchy from weaker to stronger evidence is:

1. General opinion or approval.
2. Stated intention to try.
3. Description of a recent problem and existing workaround.
4. Consent to test a prototype.
5. Repeated use or measurable behaviour change.
6. Payment, signed pilot, referral or other costly commitment.

This hierarchy is not absolute. A prepayment may be impossible in regulated procurement, while a signed data-sharing agreement may be a strong signal. Evidence must be interpreted in context.

6.3 Experiments

A problem-validation experiment should specify:

• the assumption;
• the target segment;
• the intervention;
• the behavioural measure;
• the threshold for action;
• the time period;
• possible confounding factors.

For example, a venture might test whether operations managers will request a diagnostic report after seeing a landing page. The experiment does not prove product–market fit, but it can test whether the problem and message attract qualified attention.

Quantitative funnel measures are useful when defined carefully. If 240 qualified visitors produce 18 completed applications, the application conversion rate is:

The measure is meaningful only if “qualified visitor” and “completed application” are consistently defined. It should not be interpreted as causal proof without considering traffic source, audience quality, seasonality and experiment design.

7. Market Sizing and Opportunity Assessment

Market sizing estimates the economic scope of a defined opportunity. It supports strategic choices, investor communication, capacity planning and prioritisation. It does not replace customer evidence.

7.1 TAM, SAM and SOM

Total addressable market (TAM) is the total potential demand if the venture could serve every relevant customer under the defined product and price assumptions. Serviceable available market (SAM) is the portion of TAM that fits the venture’s geography, product scope, regulation and delivery capability. Serviceable obtainable market (SOM) is the realistic portion of SAM that the venture could capture over a specified period, given competition and resources.

A simple bottom-up estimate is:

where N is the number of relevant customers, P is revenue per transaction or period, and F is the number of transactions or periods in the chosen time horizon. Units must be consistent. For an annual subscription, P should be annual revenue per customer and F is 1 annual subscription, not 12 unless P is monthly revenue.

For a subscription business:

These formulas assume that the customers are sufficiently homogeneous and that price, adoption and retention assumptions are credible. They do not automatically account for discounts, churn, capacity constraints or channel margins.

7.2 Worked market-sizing example

Consider an illustrative workflow product for independent dental clinics in one country. Suppose research identifies 8,000 clinics, but only 60% offer the relevant appointment category. The serviceable customer base is:

Assume the annual price is £1,200 per clinic. The SAM estimate is:

If the venture believes it can reach 8% of these clinics within three years, the simple SOM revenue estimate is:

The 8% assumption is not a fact produced by the formula. It requires support from sales capacity, channel access, competitive conditions, adoption rates and retention. A credible analysis should present a range rather than a single falsely precise figure. For example, if plausible capture is 4% to 8%, the corresponding annual revenue range is £230,400 to £460,800.

7.3 Top-down and value-based approaches

A top-down estimate begins with an industry or category total and narrows it using filters. It is useful for context but may conceal whether the venture can actually reach the stated customers.

A value-based estimate starts with the economic value created for the customer and estimates what share of that value could reasonably be captured. For a product that reduces avoidable administrative costs, the analysis might consider the number of affected organisations, the average cost reduction and a defensible price relative to that value.

The strongest market-sizing work triangulates approaches. If top-down, bottom-up and value-based estimates differ greatly, the analyst should investigate the assumptions rather than average the numbers mechanically.

8. From Insight to Growth Opportunity

A growth opportunity exists when a meaningful customer problem intersects with a reachable market, a credible solution and a viable business model. The following opportunity screen is useful:

Growth hacking should be applied after this screen, not instead of it. Tactics such as referral loops, onboarding experiments, content distribution and pricing tests can accelerate learning and adoption. They cannot compensate for an insignificant problem or an incorrectly defined market.

A useful prioritisation method is to score opportunities against explicit criteria using a common scale, such as 1 to 5. Scores should not be treated as objective measurements; their purpose is to expose assumptions and support discussion. The team should record the evidence supporting each score and identify which uncertainty is most important to resolve next.

9. Applied Case: CareRoute Scheduling (Illustrative)

Background

CareRoute is an illustrative early-stage venture developing scheduling software for small home-care agencies. The founders initially assumed that agency owners wanted a sophisticated artificial-intelligence scheduling engine.

Problem

Early demonstrations attracted interest but few commitments. Owners praised the technology, yet staff continued using spreadsheets and messaging applications. The founders needed to determine whether the problem was poor scheduling optimisation, fragmented communication or something else.

Analysis

The team interviewed agency owners, coordinators and care workers about recent scheduling incidents. They discovered that coordinators spent substantial time resolving last-minute absences and communicating changes. The most costly consequence was not merely inefficient route planning; it was missed information, overtime and anxiety about whether vulnerable clients would receive care.

The relevant job was reframed as: “When a care worker becomes unavailable, I need to reassign the visit and communicate the change quickly so that the client receives dependable care and the agency remains compliant.” This revealed several stakeholders and a stronger emotional and regulatory dimension than the original technology-focused hypothesis.

The team segmented the market by agency size, regulatory environment and scheduling complexity. It selected agencies with 20–60 care workers, frequent last-minute changes and a coordinator who could approve a pilot. Large providers had greater budgets but slower procurement, while very small agencies had acute needs but limited ability to pay.

Approach and solution

Instead of building the full optimisation engine, CareRoute created a lightweight pilot that recorded absence alerts, proposed available replacements and generated an audit trail of communications. Interviews were followed by observation of live scheduling work. The team measured time to reassign a visit, missed-visit incidents and weekly active use by coordinators.

The pilot agreement required agencies to provide baseline data and permitted a limited paid trial. This was stronger evidence than collecting favourable opinions about a future product. The team also clarified that automated recommendations required human confirmation, especially where client preferences or safeguarding issues were involved.

Outcome

The pilot showed that coordinators valued rapid communication and traceability more than complex route optimisation. CareRoute narrowed its initial value proposition, delayed advanced automation and focused its growth channel on regional care associations and compliance-focused events. The opportunity became smaller than the founders’ original national estimate but more accessible and strategically coherent.

Learning points

1. The stated solution may conceal the actual job.
2. Observing workflows can reveal costs that interviews alone miss.
3. A narrow beachhead can provide stronger evidence than a broad market claim.
4. Paid or operationally consequential pilots are stronger than expressions of interest.
5. In regulated settings, human oversight and auditability are part of the value proposition.

10. Technology, AI and Data Governance

Digital products provide extensive behavioural data, including search activity, click paths, usage frequency and retention. These data can improve segmentation and identify friction in the customer journey. Artificial intelligence can support interview transcription, thematic coding, propensity analysis and personalised experimentation.

However, automated analysis does not eliminate research judgement. Data may reflect platform bias, unequal access, historical discrimination or measurement errors. Correlation between a customer attribute and conversion does not prove that the attribute causes the behaviour. Models should therefore be tested for validity, stability and fairness.

Responsible practice includes:

• obtaining informed consent where personal data are collected;
• collecting only data necessary for the stated purpose;
• protecting sensitive information and limiting access;
• explaining automated decisions when they materially affect customers;
• checking for discriminatory segmentation or exclusion;
• allowing correction, withdrawal or human review where appropriate;
• distinguishing observed behaviour from inferred identity or intention.

A persona inferred from online behaviour should not be presented as an established fact. In high-stakes sectors such as health, finance, employment and education, customer discovery must comply with applicable privacy, consumer-protection and sector-specific requirements.

11. Challenges and Limitations

Confirmation bias

Founders may selectively notice evidence that supports their preferred idea. Pre-registering hypotheses, using neutral interview questions and inviting independent review can reduce this risk.

Sample bias

Early respondents may be unusually enthusiastic, digitally capable or connected to the founders. Recruitment should cover relevant variation in context and behaviour, not merely convenience samples.

Stated-versus-revealed preference

Customers may sincerely report intentions that do not translate into action. Behavioural tests and commitment mechanisms provide more reliable evidence, although they may introduce their own selection effects.

False precision in market sizing

A market figure with many digits can create unwarranted confidence. Analysts should show assumptions, ranges and sensitivity to price, customer count, adoption and retention.

Changing markets

Technology, regulation, macroeconomic conditions and cultural practices can alter customer needs rapidly. Discovery is therefore continuous rather than a one-time stage completed before launch.

Over-segmentation

Excessively narrow segments can produce fragmented messaging and insufficient scale. Segmentation should be connected to meaningful differences in needs, economics or behaviour.

Best-Practice Checklist

Before committing substantial growth resources, a venture should be able to answer:

• Which customer segment is being prioritised, and why?
• What job is the customer trying to accomplish?
• What evidence shows that the problem is frequent, severe or costly?
• What alternatives are customers using today?
• Who uses, pays for, approves and blocks adoption?
• Which assumptions remain uncertain?
• What experiment would produce the most decision-relevant evidence?
• How were TAM, SAM and SOM calculated?
• Are the market assumptions bottom-up, top-down, value-based or triangulated?
• What ethical, privacy and regulatory risks arise from the research?

Chapter Summary

Customer discovery is the disciplined process of learning about customers, problems, alternatives and adoption conditions before scaling a venture. It links customer development with iterative experimentation and helps entrepreneurs distinguish problem evidence from solution enthusiasm.

The jobs-to-be-done perspective explains the progress customers seek in particular circumstances and includes functional, emotional and social dimensions. Segmentation identifies groups with meaningful differences in needs, behaviour or economics, while personas convert research into a practical representation of priority customers. Problem validation requires behaviour-focused interviews, observation and experiments, with costly commitments generally providing stronger evidence than opinions.

Market sizing uses TAM, SAM and SOM to distinguish total potential demand, serviceable demand and realistically obtainable demand. Bottom-up calculations are useful when their customer, price and adoption assumptions are explicit and unit-consistent. Attractive growth opportunities combine desirability, severity, reachability, feasibility, viability, defensibility and timing.

Finally, discovery is subject to confirmation bias, sample bias, false precision and changing market conditions. Ethical research requires privacy protection, transparency, fairness and human judgement, especially when artificial intelligence is used to infer customer characteristics or make decisions.

Key Terms

Customer discovery: Systematic investigation of customer problems, contexts, behaviours and adoption conditions.

Customer development: An iterative process of testing customer, market and business-model assumptions.

Jobs-to-be-done: A perspective that explains the progress a customer seeks in a particular circumstance.

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

Emotional job: The desired feeling or emotional state associated with an experience.

Social job: The way a customer wants to be perceived by others.

Market segmentation: Division of a broad market into groups with relevant similarities and differences.

Beachhead market: A focused initial segment selected for entry, learning and traction.

Persona: A research-based representation of a significant customer type.

Problem validation: Testing whether a customer problem is real, important and sufficiently consequential.

Assumption register: A documented list of hypotheses requiring evidence.

Revealed preference: Evidence inferred from actual behaviour rather than stated intention.

Total addressable market (TAM): Total potential demand under defined product and market assumptions.

Serviceable available market (SAM): The portion of TAM that fits the venture’s scope and access constraints.

Serviceable obtainable market (SOM): The realistic portion of SAM the venture can capture in a specified period.

Bottom-up market sizing: Estimation based on identifiable customers, price and usage or revenue assumptions.

Value proposition: The specific customer value promised by a product or service.

Pivot: A substantive change to a venture’s customer, problem, solution or business-model hypothesis.

Product–market fit: Evidence that a product consistently satisfies a meaningful market need.

Review Questions

Short Answer Questions

1. What is the difference between problem discovery and solution discovery?
2. Define a beachhead market and explain its strategic purpose.
3. What are the functional, emotional and social dimensions of a job-to-be-done?
4. Distinguish TAM, SAM and SOM.
5. Why is behavioural evidence generally stronger than an expression of interest?

Long Answer Questions

1. Explain the customer development process and discuss how it reduces the risk of premature scaling.
2. Compare demographic, behavioural and needs-based segmentation, using examples of when each may be appropriate.
3. Describe how to design and conduct a high-quality customer discovery interview.
4. Explain bottom-up market sizing, including its assumptions, advantages and limitations.
5. Discuss how personas should be developed and governed so that they remain evidence-based rather than stereotypical.

Analytical / Application Questions

1. A venture claims that all university students are its target market. Design a segmentation and discovery plan that identifies a defensible beachhead market.
2. Using a suitable bottom-up method, estimate the annual SAM for a service with 2,500 eligible organisations and an annual price of £900. State at least three assumptions and explain how you would estimate SOM.
3. A prototype receives many positive interview responses but no users agree to a pilot. Diagnose at least three possible explanations and propose experiments to distinguish among them.

Multiple Choice Questions with Answers

Question 1

Which type of evidence is generally strongest for validating customer demand?

A. A customer says the idea is interesting

B. A customer says they would use the product someday

C. A customer pays for or commits to a pilot

D. A customer agrees that the market is large

Correct Answer: C. A customer pays for or commits to a pilot

Question 2

What does the jobs-to-be-done perspective primarily examine?

A. The demographic profile of a market

B. The progress a customer seeks in a particular circumstance

C. The technical architecture of a product

D. The total number of competitors

Correct Answer: B. The progress a customer seeks in a particular circumstance

Question 3

Which criterion asks whether a venture can identify and reach members of a segment?

A. Accessibility

B. Substantiality

C. Responsiveness

D. Defensibility

Correct Answer: A. Accessibility

Question 4

A market has 2,000 customers and an annual price of £500. Using a simple bottom-up estimate, what is the annual market revenue potential?

A. £100,000

B. £500,000

C. £1,000,000

D. £2,500,000

Correct Answer: C. £1,000,000

Question 5

Which statement best describes SOM?

A. All possible global demand

B. The portion of demand outside the venture’s product scope

C. The realistic portion of the serviceable market the venture can capture

D. The number of people interviewed during discovery

Correct Answer: C. The realistic portion of the serviceable market the venture can capture

Question 6

Which interview question is most consistent with good customer discovery practice?

A. Would you definitely buy our product?

B. Do you like our proposed features?

C. Tell me about the last time this problem occurred

D. Would a discount persuade you to switch immediately?

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

Question 7

In a business-to-business purchase, who is most likely to use the product directly?

A. The user

B. The economic buyer

C. The approver

D. The blocker

Correct Answer: A. The user

Question 8

Why should market-sizing assumptions be stated explicitly?

A. To make the estimate appear more complex

B. To allow the estimate to be evaluated and challenged

C. To guarantee that the forecast will be accurate

D. To eliminate the need for customer research

Correct Answer: B. To allow the estimate to be evaluated and challenged

Question 9

Which is a major risk of using AI-generated customer segments without human review?

A. It always produces too few segments

B. It may reproduce bias or treat inferences as facts

C. It prevents all quantitative analysis

D. It makes customer interviews unnecessary in every context

Correct Answer: B. It may reproduce bias or treat inferences as facts

Question 10

A venture observes that customers use spreadsheets instead of a competitor’s software. What does this most clearly indicate?

A. There is no customer problem

B. The spreadsheet may be a competing alternative or workaround

C. The market is necessarily too small

D. The customers are already loyal to the competitor

Correct Answer: B. The spreadsheet may be a competing alternative or workaround