Customer Discovery, Segmentation and Market Sizing

Introduction

A business plan is credible only when it demonstrates that a clearly defined group of customers has a meaningful problem, recognizes the proposed solution as valuable, and represents a market large enough to support the venture. Customer discovery, segmentation, and market sizing provide the evidence for that claim. They are not independent activities. Discovery generates evidence about customer needs; segmentation organizes customers according to meaningful differences; targeting selects the group the venture can serve most effectively; and market sizing estimates the commercial opportunity.

These activities also reduce a common entrepreneurial error: confusing an attractive idea with a validated opportunity. Entrepreneurs frequently begin with a product and then search for customers. A disciplined process begins with customer circumstances, problems, desired outcomes, purchasing behavior, and constraints. The resulting business plan should distinguish observed evidence from assumptions and estimates.

Learning Objectives

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

1. Design and conduct ethical customer-discovery research.
2. Develop evidence-based personas and jobs-to-be-done.
3. Segment a market using relevant and measurable criteria.
4. Select a target market and explain positioning choices.
5. Map a customer journey and identify points of friction.
6. Estimate TAM, SAM, and SOM using defensible top-down and bottom-up methods.
7. Evaluate assumptions, data quality, uncertainty, and ethical risks in market analysis.

1. Customer Discovery as Evidence-Based Learning

Customer discovery is the systematic process of learning about prospective customers before committing substantial resources to a business model. It is associated with lean-startup practice, but its intellectual foundations also include market research, consumer behavior, design thinking, and opportunity evaluation. Discovery is not simply asking people whether they like an idea. Positive reactions are weak evidence because respondents may be polite, curious, or unwilling to pay.

A stronger discovery process investigates:

• the customer’s current situation;
• the problem or desired outcome;
• how often the problem occurs;
• the consequences of leaving it unresolved;
• current alternatives and substitutes;
• who influences and authorizes purchase;
• budget, switching costs, and adoption barriers; and
• observable evidence of commitment, such as a trial, referral, deposit, pilot, or purchase.

Discovery questions and evidence quality

Good questions elicit past or current behavior rather than hypothetical intentions. “Tell me about the last time you managed this task” is generally more informative than “Would you use our application?” The first question can reveal frequency, existing workarounds, costs, and emotional context. The second often produces an unreliable prediction.

Evidence can be arranged in an approximate hierarchy:

This hierarchy is not absolute. A paid commitment may be small or experimental, while a large survey may reveal a robust pattern. Nevertheless, entrepreneurs should avoid treating all data as equally persuasive.

A practical discovery protocol

A capstone team can use the following sequence:

1. State assumptions. Identify who is believed to experience the problem, what the problem is, and why the proposed solution might be preferred.
2. Recruit varied participants. Include current users of alternatives, non-users, experts, and people who rejected existing solutions where relevant.
3. Conduct semi-structured interviews. Use a consistent topic guide while allowing unexpected themes to emerge.
4. Observe context. Where appropriate, examine the task in its natural setting rather than relying only on memory.
5. Record evidence systematically. Capture quotations, behaviors, frequency, costs, and circumstances without exposing personal identifiers.
6. Code and compare. Group observations into recurring problems, motivations, constraints, and behaviors.
7. Test risky assumptions. Use prototypes, landing pages, concierge services, pilots, or pricing experiments.
8. Update the business model. Evidence should change the product, segment, channel, pricing, or even the decision to proceed.

Interview samples need not be statistically representative when the objective is exploratory learning. However, the team must explain whom it interviewed, how participants were recruited, and what bias may result. A small sample can identify a serious problem; it cannot establish market prevalence by itself.

2. Personas and Jobs-to-be-Done

Personas

A persona is a research-based representation of a meaningful customer type. It normally includes goals, context, behaviors, constraints, triggers, and decision criteria. A persona is not a fictional biography designed to make a report attractive. It should summarize patterns found across several individuals and should identify the evidence supporting those patterns.

A useful persona contains:

• a descriptive label, such as “time-constrained independent retailer”;
• relevant demographic or firmographic characteristics;
• context and triggering events;
• tasks and desired outcomes;
• current alternatives and frustrations;
• purchasing authority and budget;
• objections and adoption barriers; and
• evidence sources and confidence level.

Demographic variables can be useful, but they rarely explain behavior on their own. Two people of the same age may have entirely different needs, budgets, or buying situations. Behavioral and situational characteristics are often more predictive.

Jobs-to-be-done

The jobs-to-be-done perspective asks what a customer is trying to accomplish in a particular circumstance. A job may be functional, emotional, or social. For example, a commuter may be trying to “arrive reliably without spending time planning,” while also wanting to feel in control and appear responsible to colleagues.

A job statement can follow this structure:

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

The job is not identical to a product feature. “Use an app” is a solution; “coordinate a reliable delivery window” is a job. This distinction helps entrepreneurs identify substitutes and avoid defining the market too narrowly.

The value of a job depends on several dimensions: importance, frequency, dissatisfaction with current alternatives, urgency, and willingness or ability to pay. These dimensions should be investigated rather than assumed. A frequent problem may still be commercially weak if customers tolerate it and do not control the budget.

Persona–job example

Consider an illustrative venture offering scheduling software for small medical practices. Research may produce the following persona:

Maya, practice administrator: manages a five-person clinic, loses appointment capacity through cancellations, uses spreadsheets and phone calls, and values predictable implementation more than advanced customization. Her job is: “When appointments change at short notice, I want to refill vacant slots quickly so that the clinic protects revenue without increasing administrative workload.”

The persona identifies the decision maker and context; the job identifies the outcome. The clinic’s patients may be users, but Maya may be the economic buyer. This distinction is essential in business-to-business markets.

3. Market Segmentation and Target-Market Selection

Segmentation divides a broad market into groups that differ in needs, behavior, economics, or response to marketing. A useful segment is not merely a group that can be named; it must be sufficiently identifiable, reachable, substantial, and responsive to a distinct value proposition.

Major segmentation bases

Segmentation should be relevant to the decision being made. If a firm is deciding how to deliver a service, usage intensity and location may matter more than age. If it is selling enterprise software, procurement process, number of sites, and integration requirements may be decisive.

Selecting a target segment

Targeting is a strategic choice, not a claim that every potential customer should be served. A segment can be evaluated using criteria such as:

1. Problem intensity: Is the need important and urgent?
2. Reachability: Can the venture identify and access buyers efficiently?
3. Economic attractiveness: Is there sufficient willingness and ability to pay?
4. Competitive position: Can the venture offer a defensible advantage?
5. Strategic fit: Does the segment match capabilities, mission, and resources?
6. Growth and learning potential: Can initial customers support expansion or learning?
7. Risk and compliance: Are legal, ethical, or operational risks manageable?

A simple scoring matrix can structure judgment. Suppose a team scores each segment from 1 to 5 on problem intensity, accessibility, economics, fit, and competition advantage, with weights that sum to 1. The weighted attractiveness score is:

where S_j is the score for segment j, w_i is the weight assigned to criterion i, and r_ij is the rating of segment j on criterion i. Ratings are assumed to use the same scale, and weights must be non-negative and sum to 1. The result is a decision aid, not an objective truth; teams should perform sensitivity analysis because scores may reflect subjective assumptions.

Positioning

Positioning specifies how the venture intends to be understood by the target customer relative to alternatives. A positioning statement commonly identifies the target customer, need, category, benefit, and reason to believe:

For [target customer] who [need], [offering] is a [category] that [primary benefit] because [evidence or capability].

A strong position is specific enough to guide product and communication decisions. “High quality for everyone” is not a useful position because it does not distinguish the offering or identify a priority customer.

4. Customer Journey Mapping

A customer journey map represents the sequence of experiences through which a customer moves before, during, and after an interaction with a venture. It should be based on evidence, not merely on the organization’s internal process.

A typical journey includes:

1. Awareness: the customer notices a problem or encounters the venture.
2. Consideration: the customer searches, compares, and evaluates alternatives.
3. Purchase or adoption: the customer commits, pays, or begins a trial.
4. Use and onboarding: the customer attempts to achieve the desired outcome.
5. Support and recovery: the customer seeks help or encounters failure.
6. Retention, expansion, or exit: the customer renews, recommends, upgrades, or switches.

For each stage, document the customer’s goal, actions, questions, emotions, touchpoints, obstacles, and evidence. The map should distinguish the customer from the organization. A company may regard a form as “completed,” while the customer experiences uncertainty about price, privacy, or next steps.

Journey maps help identify moments of truth: interactions that disproportionately influence trust, conversion, satisfaction, or retention. They also expose handoff failures between marketing, sales, operations, and support. A business plan should connect these insights to measurable improvements, such as reducing onboarding time, increasing activation, or clarifying refund policies.

5. Market Sizing: TAM, SAM, and SOM

Market sizing estimates the scale of a commercial opportunity. The three common levels are:

• TAM (Total Addressable Market): the total annual revenue or volume that could theoretically be generated if the relevant market were served without the venture’s practical constraints.
• SAM (Serviceable Available Market): the portion of TAM matching the venture’s geographic, product, regulatory, channel, and capability boundaries.
• SOM (Serviceable Obtainable Market): the portion of SAM the venture can realistically capture within a stated period, given competition and execution capacity.

These are not automatically market shares. TAM and SAM should be defined by a unit and time period, such as annual revenue, annual subscriptions, or annual transactions. SOM must state the forecast horizon and assumptions.

Top-down estimation

Top-down sizing begins with an external market total and narrows it using defensible filters. A simplified structure is:

where p_g, p_c, p_r, and p_f are proportions associated with geography, customer or product category, regulatory eligibility, and other explicit filters. Each proportion must refer to the same market definition, and multiplying proportions assumes the filtering logic is appropriate; overlapping or dependent filters should not be treated as independent without justification.

Top-down research may use government statistics, industry associations, audited company reports, regulator publications, and reputable research providers. The analyst should record publication date, definition, geography, unit, methodology, and limitations. A headline such as “the digital health market is worth billions” is not sufficient unless the venture’s product and customer are included in that definition.

Bottom-up estimation

Bottom-up sizing builds the estimate from customers, usage, price, and capacity. A basic revenue formulation is:

where N is the number of potential customer units, q is annual quantity purchased per customer, and P is price per unit. Units must be consistent. If N is clinics, q is subscriptions per clinic per year, and P is annual price per subscription, the result is annual revenue.

For a subscription venture with monthly price P_m and N customers, annual recurring revenue capacity is:

This assumes every customer pays the stated price for twelve months and excludes discounts, churn, taxes, refunds, and variable usage fees. A realistic forecast should adjust these assumptions rather than presenting the formula as a guarantee.

Worked market-sizing example

An illustrative scheduling venture initially serves independent clinics in one country. Secondary research identifies 18,000 eligible clinics. Interviews and industry data suggest that 40% use a paid scheduling solution or have a sufficiently urgent need to be considered serviceable. The proposed annual price is $1,200 per clinic.

The bottom-up SAM estimate is:

The venture’s first three-year plan includes a sales team and implementation capacity sufficient to serve 180 clinics. If the average realized annual revenue is $1,050 after discounts, the capacity-constrained SOM is:

The implied share of the calculated SAM is:

The estimate is defensible only if the 18,000 clinics, 40% filter, price, and 180-clinic capacity are documented. It would be misleading to claim that the venture will capture 2.19% merely because the arithmetic produces that number. The figure represents a stated scenario, not a probability.

Reconciling top-down and bottom-up estimates

A useful practice is to calculate both methods and investigate major differences. Top-down analysis provides context and comparability; bottom-up analysis exposes operational assumptions. If a top-down estimate is much larger, the venture may be defining the market too broadly. If the bottom-up estimate is larger, the customer count, price, or purchase frequency may be unrealistic.

Scenario analysis is preferable to a single false-precision estimate. Define conservative, base, and optimistic cases by varying the most uncertain inputs. For a multiplicative estimate, relative uncertainty in any major input can materially affect the result. Teams should identify which assumptions are most sensitive and test those first.

6. An Integrated Capstone Method

A coherent capstone chapter or business plan can use the following workflow:

1. Define the initial problem hypothesis and relevant market boundary.
2. Conduct discovery interviews and observations across plausible segments.
3. Synthesize patterns into personas and jobs-to-be-done.
4. Identify alternative solutions, including non-consumption and manual workarounds.
5. Segment customers using variables linked to need, behavior, and economics.
6. Score segments and select a beachhead target market.
7. Map the journey from trigger to retention.
8. Build top-down and bottom-up TAM, SAM, and SOM estimates.
9. Test pricing, conversion, and commitment assumptions.
10. Record evidence, uncertainty, and decisions in the business plan.

The sequence is iterative. New evidence may redefine the persona, reveal a better segment, or show that the original market is not attractive. Entrepreneurs should therefore maintain an assumption register containing the assumption, evidence, confidence, test, owner, and decision consequence.

Applied Case: RefillRoute (Illustrative)

Background

RefillRoute is a fictional venture proposing a reusable-container delivery service for household cleaning products. The founders initially believe their target market is “environmentally conscious consumers.” This description is too broad to guide product design or market sizing.

Problem and discovery

The team conducts 24 interviews: 10 apartment residents, 8 parents in suburban households, and 6 owners of small hospitality businesses. It finds that apartment residents like waste reduction but have limited storage; parents value convenience and safety; hospitality businesses have predictable volume but require invoices, reliable delivery, and compliance documentation. The strongest evidence of urgent demand comes from hospitality businesses that currently buy single-use containers and spend staff time managing deliveries.

The team observes that the hospitality buyers do not primarily purchase environmental benefits. Their job is: “When cleaning supplies run low, I need a predictable replenishment process that reduces staff interruption and waste.” The founders revise the target from environmentally conscious consumers to independent hospitality businesses with regular supply usage.

Segmentation and choice

The team compares cafés, small hotels, and short-term rental managers. Cafés are accessible through local associations, have recurring demand, and can trial the service quickly. Hotels have larger potential volume but longer procurement cycles and more stringent documentation. Short-term rental managers are fragmented and have variable demand. The team selects cafés as its beachhead segment while retaining hotels as a later expansion segment.

Journey analysis

The café journey reveals a major friction point during onboarding: owners do not know whether reusable containers will fit their existing storage and dispensing systems. The venture therefore adds a compatibility checklist, a sample-kit trial, and a delivery calendar. These changes address the customer’s operational risk rather than merely improving promotional messaging.

Market sizing

The team identifies 3,200 independent cafés in its launch region. It estimates that 35% purchase sufficient cleaning supplies through accessible channels and could use the container system. At an average annual contribution of $900 per café, the regional SAM scenario is:

Its first-year logistics plan can support 120 cafés at an expected realized annual contribution of $780 per café:

The team labels these estimates as scenarios, identifies the 35% eligibility assumption as uncertain, and tests it through a local customer list and pilot. It also measures container return rates because the business model fails if containers are not returned reliably.

Learning points

The case illustrates five principles. First, a broad value such as sustainability does not automatically define a useful segment. Second, jobs are situated in operational contexts. Third, the buyer’s economic and logistical constraints may matter more than demographic identity. Fourth, journey mapping can reveal product requirements. Fifth, SOM is constrained by operational capacity, not merely by the size of the market.

7. Ethics, Data Governance, and the Use of Technology

Customer discovery involves personal information, sometimes including health, financial, location, or employment data. Researchers should collect only information necessary for the stated purpose, obtain informed consent, protect recordings and transcripts, and remove identifying details from reports. Applicable privacy, consumer-protection, accessibility, and sector regulations must be considered. Participants should not be misled about the venture, pressured to disclose sensitive information, or promised benefits that are unavailable.

Sampling and interpretation also create ethical risks. Recruiting only friends, affluent early adopters, or digitally active users can produce exclusionary conclusions. Teams should report who was not represented. Personas should not reinforce stereotypes, and automated tools should not infer sensitive characteristics without a legitimate and transparent basis.

Artificial intelligence can assist with interview transcription, thematic coding, competitor monitoring, and scenario calculations. However, AI-generated summaries may omit context, misclassify statements, or reproduce bias. Raw recordings should not be uploaded to a tool unless confidentiality, consent, security, and contractual conditions permit it. A researcher must verify quotations, preserve an audit trail, and distinguish machine-generated interpretation from participant evidence.

Contemporary market analysis increasingly combines first-party behavioral data, digital experiments, social listening, and public administrative data. These sources can improve timeliness but may also create surveillance, consent, and representativeness concerns. Technology should support judgment, not replace direct engagement with customers.

8. Common Errors and Best Practices

Error: treating a large population as a market. A population becomes a market only when a relevant need, access route, purchasing mechanism, and value exchange are plausible.

Error: counting competitors as proof of demand. Competitors may indicate demand, but they may also reveal saturation, weak economics, or regulatory barriers.

Error: using survey interest as revenue. Stated interest must be tested against behavior, price, and commitment.

Error: mixing units. Customer counts, transactions, monthly prices, and annual revenues must be converted consistently.

Error: presenting SOM as a guaranteed share. SOM is a scenario constrained by resources, competition, conversion, retention, and time.

Error: hiding uncertainty. A range with explicit assumptions is more professional than an unsupported precise number.

Best practice is to triangulate sources, retain a research log, separate facts from assumptions, use conservative definitions, test the riskiest assumptions first, and revise the business plan as evidence accumulates.

Chapter Summary

Customer discovery is an evidence-based process for understanding customer circumstances, problems, alternatives, decision processes, and willingness to commit. Strong discovery emphasizes past behavior, observation, and commitment rather than hypothetical enthusiasm. Personas summarize evidence-based customer patterns, while jobs-to-be-done express the outcomes customers seek in particular situations.

Segmentation divides a market using meaningful geographic, demographic, psychographic, behavioral, firmographic, or situational differences. Targeting selects a segment on the basis of problem intensity, accessibility, economics, strategic fit, competitive position, growth, and risk. Positioning then explains how the offering should be understood relative to alternatives.

Customer journey maps connect customer goals and experiences from awareness through retention. They reveal friction, moments of truth, and operational requirements. TAM, SAM, and SOM must be defined with consistent units and time periods. Top-down estimates provide market context, whereas bottom-up estimates expose assumptions about customer counts, quantities, prices, and capacity. A rigorous business plan uses both methods, reports uncertainty, and treats SOM as a scenario rather than a promise.

Ethical research requires consent, privacy protection, fair representation, and responsible use of data and AI. The central capstone principle is integration: customer evidence should determine the segment, the segment should shape the journey and value proposition, and the resulting market estimate should be traceable to documented assumptions.

Key Terms

Customer discovery: Systematic learning about customer problems, behavior, alternatives, and purchasing conditions.

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

Jobs-to-be-done: The functional, emotional, or social outcome a customer seeks in a particular situation.

Segment: A group of customers sharing relevant needs, behaviors, or characteristics.

Target market: The segment selected for focused service and marketing.

Beachhead market: A narrowly defined initial segment used to establish traction and learning.

Positioning: The intended place of an offering in the customer’s comparative understanding.

Customer journey: The sequence of customer experiences before, during, and after adoption.

Touchpoint: An interaction between a customer and the venture or its offering.

Moment of truth: A touchpoint with disproportionate influence on trust or behavior.

TAM: Total addressable market under a broad theoretical market boundary.

SAM: Serviceable available market within the venture’s practical boundaries.

SOM: Serviceable obtainable market that can realistically be captured in a stated period.

Top-down sizing: Estimating a market by narrowing an external market total.

Bottom-up sizing: Estimating a market from customer units, usage, price, and capacity.

Firmographic data: Organizational characteristics such as industry, size, and revenue.

Triangulation: Comparing multiple evidence sources to improve credibility.

Assumption register: A documented list of assumptions, evidence, confidence, tests, and consequences.

Review Questions

Short Answer Questions

1. Why is reported past behavior generally more useful than hypothetical purchase intention in customer discovery?
2. Distinguish a persona from a job-to-be-done.
3. Identify three criteria for evaluating a target segment.
4. What is the difference between SAM and SOM?
5. Why should a market-sizing estimate specify units and time period?

Long Answer Questions

1. Explain a complete customer-discovery process and discuss how a venture should evaluate the strength of different types of evidence.
2. Compare demographic, behavioral, situational, and firmographic segmentation, giving an appropriate entrepreneurial example for each.
3. Explain how customer journey mapping can influence product design, marketing, and operations.
4. Compare top-down and bottom-up market sizing, including their advantages, limitations, and appropriate data sources.
5. Discuss ethical and governance issues in interviews, customer data collection, and AI-assisted market research.

Analytical / Application Questions

1. A venture identifies 10,000 eligible organizations, estimates that 30% are serviceable, and expects annual revenue of $800 per organization. Its first-year capacity is 150 organizations at an average realized revenue of $700. Calculate the SAM and capacity-based SOM, and state two assumptions that should be tested.
2. Design two contrasting personas and jobs-to-be-done for a venture selling remote bookkeeping software. Explain which segment you would target first and why.
3. A survey shows that 75% of respondents like a proposed product, but only 8% currently use a workaround and none has paid or requested a trial. Evaluate the evidence and propose three next research actions.

Multiple Choice Questions with Answers

Question 1

Which type of evidence is generally strongest for validating willingness to pay?

A. A favorable opinion about the concept

B. A statement that the respondent might buy later

C. A paid pilot or deposit

D. A demographic match with the target profile

Correct Answer: C. A paid pilot or deposit

Question 2

What does a jobs-to-be-done statement primarily describe?

A. The technical features of a product

B. The outcome a customer seeks in a situation

C. The total number of competitors

D. The organization’s internal workflow

Correct Answer: B. The outcome a customer seeks in a situation

Question 3

Which segmentation basis is most directly concerned with usage frequency and benefits sought?

A. Geographic

B. Behavioral

C. Demographic

D. Firmographic

Correct Answer: B. Behavioral

Question 4

Which market measure represents the portion of the market a venture can realistically capture within a stated period?

A. TAM

B. SAM

C. SOM

D. Population size

Correct Answer: C. SOM

Question 5

A market has 5,000 potential customers, 20% are serviceable, and annual price is $600. What is the bottom-up SAM revenue estimate?

A. $60,000

B. $600,000

C. $1,000,000

D. $3,000,000

Correct Answer: B. $600,000

Question 6

What is a primary purpose of a customer journey map?

A. To replace all financial forecasts

B. To identify customer goals, friction, and touchpoints

C. To calculate tax liability

D. To classify competitors by ownership

Correct Answer: B. To identify customer goals, friction, and touchpoints

Question 7

Why should a venture use both top-down and bottom-up market sizing?

A. They always produce identical results

B. Top-down gives context while bottom-up exposes operational assumptions

C. Bottom-up methods do not require data

D. Top-down methods guarantee market share

Correct Answer: B. Top-down gives context while bottom-up exposes operational assumptions

Question 8

In the formula Market\ revenue = N × q × P, what does q represent?

A. The number of potential customers

B. The annual quantity purchased per customer

C. The regulatory eligibility proportion

D. The venture’s market share

Correct Answer: B. The annual quantity purchased per customer

Question 9

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

A. Would you use our product someday?

B. Do you like environmentally friendly products?

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

D. Would your friends recommend this idea?

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

Question 10

What is an appropriate practice when using AI to analyze customer interviews?

A. Accept all generated themes without checking them

B. Upload any recording regardless of consent

C. Verify summaries and protect confidential participant data

D. Replace interviews with synthetic respondents

Correct Answer: C. Verify summaries and protect confidential participant data