Segmentation, Personas & Market Potential
Innovation & New Business Proposal - TUHH Institute of Entrepreneurship & Institute of Innovation Marketing, Hamburg · part of my Technology Management MBA · study notes for revision.
This chapter is the recap block that opens the market research part of the module. Before any of the clever tools arrive (customer journey, Kano, conjoint, house of quality), the session goes back to three linked tasks from the first semester: cutting the market into segments, drawing personas that stand for those segments, and putting a defensible number on the market. The obvious question is why we keep repeating this. The answer given is blunt: the number one reason new ventures fail is market and customer related, ahead of liquidity and funding problems, team problems and competition.
The failure logic is worth memorising, because everything in this chapter is a countermeasure to one line of it. Most innovations and new businesses fail because the solution addresses unimportant needs, or because it does not solve the problem effectively and to the point, or because the needs it does solve are relevant only for too few or for unprofitable customers. And even when a real, widely shared need is addressed, many still fail because the solution is neither unique nor better than what competitors already offer. Segmentation and personas attack the “too few or unprofitable customers” line. Market sizing puts a number on it. Opportunity evaluation asks what could still kill it.
That is why the module’s project objectives are phrased the way they are: the market is defined and the market potential is calculated on the basis of a transparent and informed calculation (TAM, SAM), a rough but substantiated sales prediction is made (SOM), the choice of target segments is backed with data, the customer profiles are clearly defined, and the most promising first customers are actually selected by name and type.
1 · Why segment and build personas at all
Section titled “1 · Why segment and build personas at all”Four reasons are given, and they are cumulative rather than alternatives.
| Reason | Why it matters |
|---|---|
| Customer populations are heterogeneous | Real markets hold different needs, different budgets and different habits. An average taken across all of them describes nobody in particular |
| It is the basis for all later market-related decisions | Product and service design, communication, pricing and sales all follow from it. Get the segment wrong and each of those four is optimised for the wrong person |
| It prevents losing focus on the priority customers | It keeps the team pointed at the customers that matter. Without it, every feature request from anyone feels equally important |
| It is a communication device | Internally the whole team argues about the same customer; externally it is exactly what a pitch audience wants to hear - who buys this, and why them |
2 · Segment first, personas second
Section titled “2 · Segment first, personas second”The order matters and is easy to get backwards. A sophisticated segmentation of the market comes first. Only afterwards do you develop personas, and each persona represents a typical customer inside one of the segments you found.
Doing it the other way round is one of the quiet killers. If you write the persona first, you will always find a story that makes your idea sound needed, and you will never learn how many such people there are.
3 · The four questions a persona must answer
Section titled “3 · The four questions a persona must answer”A persona is not a biography. It is an answer to four questions, and the four map onto the decisions you will have to make later.
- Consumers: gender, age, education, profession, family situation, prior knowledge
- Business customers: company size, industry, location by country and region, financial situation
- Buying behaviour: preference for a certain type of product or service, and the risk attitude when buying something new
- Information behaviour: which information channels they actually use
- Usage behaviour: how often they use it, and in which use context
- The current problems of the user or customer, stated as their problems and not as your features
- This is the line your value proposition has to answer
- Preferences for concrete features and functionalities
- Expectations regarding product or service quality, and their willingness to pay
4 · The correction: what dominates, and what should be central
Section titled “4 · The correction: what dominates, and what should be central”This is the single most useful slide of the recap. On the left is what personas are usually full of; on the right is what they should actually be built around. The test is simple: does this characteristic stand in close relation to my idea? If not, it is decoration.
- Demographics: age, gender and so on
- Socioeconomics: occupation, education, income
- Psychographics: attitudes, opinions, lifestyles
- Easy to find and easy to write, and they rarely decide anything about your product
- Pain points: the current problems of users or customers
- Usage behaviour: use frequency, use expertise
- Preferences: features, attributes, product categories
- Purchasing behaviour: willingness to pay, buying volume
- Predicted response: likelihood of early adoption, customer lifetime value
- Preferred communication and distribution channels
5 · Six common flaws
Section titled “5 · Six common flaws”| # | Flaw | What it looks like in practice |
|---|---|---|
| 1 | Crude evidence | The persona rests on good guesses and unproven assumptions. No data behind it, no real understanding of the market |
| 2 | Wishful thinking | You describe the ideal customer for your idea rather than a customer group that actually exists |
| 3 | Neglecting | You never think about the customer groups you cannot address, so you never notice how much of the market you have quietly written off |
| 4 | Overloading | The persona is stuffed with characteristics that are not directly related to the product or service |
| 5 | Simplification | Rough and simple demographics take over: young females, students, elderly people. Broad labels, no behaviour |
| 6 | Heterogeneity | The persona is not homogeneous, meaning the group behind it does not really share similar use behaviour, expectations and preferences |
Flaws 4 and 5 look like opposites but they usually appear together: too much irrelevant detail on the profile side, and almost nothing on the behaviour side.
6 · Limitations: what this exercise can and cannot do
Section titled “6 · Limitations: what this exercise can and cannot do”- Garbage in, garbage out. A segmentation is only as good as its input. Clean-looking segments built on weak data are still weak, they just look convincing on a slide.
- Early segmentation is open to multiple interpretations. At an early stage the same raw material can be cut in several defensible ways. There is rarely one objectively correct set of segments.
- Personas can be reverse engineered. This is the dangerous one: instead of finding the segment in the market, you invent the segment that perfectly matches the business idea you already love.
- It is only a starting point. An early segmentation exercise needs frequent updates as the project moves on. It is a working hypothesis you revise every time you talk to real customers, not a conclusion you file away.
7 · Choosing the first target customers
Section titled “7 · Choosing the first target customers”Segmentation tells you which group to serve. The next question is which customers inside that group to approach first, and here adoption theory does the work. The classic adopter typology splits the market over time into innovators, early adopters, the early majority, the late majority and laggards. That first group carries several names depending on context: innovators, lead users, launching customers, pilot customers. Three characteristics make them attractive: they possess the resources to acquire and implement new technology, so they can deploy an immature product; they have the education, knowledge and experience to understand it plus a high general interest in new technology for its own sake, so they are the first to realise its future potential and you do not have to explain the whole category from zero; and they have a higher risk-taking propensity, so they are motivated to co-develop and are not afraid of fiddling with and debugging something new.
Opinion leaders are not automatically innovators
Section titled “Opinion leaders are not automatically innovators”It has long been argued that opinion leaders and influentials can be used to intensify the diffusion of innovations: mass communication reaches them, and they pass it on through personal communication and word of mouth to the entire market. They typically have more contact with information about innovations, high expertise and high involvement, a higher degree of social interaction through a larger network, and higher status in their community. But the two roles only coincide under one condition. If the customer base is very open to innovation, opinion leaders and innovators are the same people. If it is rather conservative regarding innovations, they are not: the opinion leaders see more that they can lose, and they are difficult to influence.
Value of a first customer beyond the sale
Section titled “Value of a first customer beyond the sale”Likelihood of adoption is not the only thing that makes a first customer attractive. Five further value concepts are given, and they are the reason a small early deal can be worth far more than its invoice.
| Concept | What it measures |
|---|---|
| Customer referral value (CRV) | The lifetime value of the new customers who become customers because this customer referred them |
| Customer influence value (CIV) | How far this customer influences other potential customers through their messages |
| Customer reputation or reference value | How far this customer is perceived by others in the industry as the quality leader and the benchmark |
| Customer knowledge value (CKV) | The value of what you learn from them - about customer preferences, about your quality problems, about competitors |
| Customer innovation value (CInnoV) | How willing they are to co-develop and to test innovative products and services, which overlaps with CKV |
Case: why one venture picked fashion publishers first
Section titled “Case: why one venture picked fashion publishers first”The session’s buyer-persona case is a bootstrapped platform offering a single-click shopping experience: publishers and influencers tag products natively inside their images, videos or text, and consumers buy without leaving the page. It needed a consolidated product catalogue with pricing and inventory, an API to inject orders into retailers’ systems, an intuitive interface for publishers and a frictionless cart for consumers, and it offered a different value proposition to each side - an easy monetisation tool for publishers, higher conversion than affiliate links or display ads plus discovery-journey analytics for retailers and brands, and the best offer across all partners without leaving the page for consumers. The interesting decision is the first market, and each reason maps onto a criterion worth reusing:
| Why fashion publishers first | The criterion behind it |
|---|---|
| Luxury goods are sensitive to online discussion and recommendation, and fashion trends are set by opinion leaders and influencers | The solution’s features match this context best |
| A large and heterogeneous group of bloggers and publishers already exists | The customer group already exists and does not have to be created |
| Fashion bloggers are underdeveloped when it comes to monetising their content | The group is underserved, so the pain is real |
| They are easy to onboard, being individuals rather than firms, and often hobbyists rather than professionals | Low acquisition cost and short sales cycles - but also small deals per head, which the sizing has to expose |
8 · Sizing the market: TAM, SAM and SOM
Section titled “8 · Sizing the market: TAM, SAM and SOM”Once the segments exist, the market potential can be calculated. Three nested figures are used, and the funnel only ever narrows.
- TAM answers “how big is the game”. It sets the ceiling and shows the opportunity is not trivially small.
- SAM answers “how big is my part of the game”. This is where segmentation and persona work enter the arithmetic, because the narrowing steps are your segment definitions.
- SOM answers “what will I actually sell”. It is the number a plan and a budget can be built on.
TAM × share for your segments = SAM · SAM × realistic share = SOM start from a published market total and cut it down with justified percentagesreachable customers × conversion × price × frequency = revenue build the number up from units you can defend one by oneTop-down is fast but fragile: one unexplained percentage can swing the answer by a factor of ten, and it is exactly where wishful thinking hides. Bottom-up is slower but every line is arguable, so a reader can attack an assumption instead of dismissing the whole calculation. Doing both and comparing them is the cheapest quality check available: if they disagree badly, one of your assumptions is wrong and you now know roughly where to look.
Surveys are the natural tool here. Where the customer base is large and heterogeneous and the question is a quantity, a representative survey with clear need hypotheses gives frequencies, mean values and correlations, which is exactly what a status quo analysis of use, a market segmentation and a TAM or SAM estimate need. The advice given is to use existing data rather than running your own survey wherever such data exists, and to use surveys for collecting exact information rather than for exploring.
9 · Evaluating the opportunity
Section titled “9 · Evaluating the opportunity”Sizing tells you how large the prize is. Evaluation asks what stands between you and it. The frame is a simple grid: potential on one axis and challenge on the other, each rated from very small to very high. An idea in the high-potential, high-challenge corner is not automatically bad, but it needs a plan for the challenge before it needs a plan for growth. The exercise itself is deliberately pessimistic and runs in two steps:
-
Grant the value proposition. Assume for a moment that the value proposition is not the crucial problem: target customers really can benefit, and they really do see the benefit.
-
Hunt for the killers. Now list the factors that could kill the business outright, or at least seriously inhibit fast adoption and diffusion in the market. Stay on factors that could really destroy the entire business case, and do not get lost in minor threats and side issues.
The examples used to practise this are instructive because in each one the hard part sits somewhere other than the technology. Clinical proctoring, where a physician is directly observed while carrying out surgery, is required by medical product law when a new medical product or surgical procedure is introduced; a newcomer needs five to ten sessions, proctors often travel from other cities, countries or even continents, one session can add up to ten thousand US dollars, and the whole thing can consume more than five percent of product revenues. The proposed solution is a camera system with browser-based streaming and steering software: low latency through a new streaming architecture, high resolution, remote control of the camera by the proctor, independent of local IT and running in a standard web browser. A second example is charging infrastructure for residential multi-storey buildings with their own parking lot, where the real work is the whole chain of consulting and planning, rollout and installation, starting up into operation, technical operation, customer service and billing. A third is a secure data exchange between a patient’s smartphone and a doctor’s PC, combining a data transfer hub connected to the doctor’s computer by USB with an app on the patient’s phone, linked over NFC or Bluetooth, with highly secured storage and encrypted transfer, able to carry x-rays, MRT or ultrasound images, prescriptions and data from heartbeat monitors or fitness trackers, and monetised by giving the hub away free of cost and charging doctors and clinics a subscription subsidised by health insurers.
The small VR case closes the loop back to segmentation. A young physician in residency, interested in IT and programming and convinced that VR will change medical education through immersive scenarios, wants to build a platform of VR training scenarios for medical students, combining a head-mounted display with input devices such as joysticks or sensor-equipped gloves, teaching content and standard practices according to medical guidelines. The uncertainty is not the hardware. It is whether demand for VR-based education of medical students and nurses exists at all, and in which medical fields or treatment scenarios the need is highest. Those two questions come first, and they are answered by going into the market, not by building.
Worked example
Section titled “Worked example”A transparent sizing for that VR training idea, priced as an annual licence per learner seat. The figures below are my own illustrative assumptions, written out to show the method and the discipline of citing a source for every single line. They are not published data.
| Line | Assumption | Source I would cite |
|---|---|---|
| Institutions in Europe teaching medicine or nursing | 4,000 | National education ministry registries plus Eurostat education statistics |
| Average learners per institution | 600 | Same registries, enrolment figures |
| Learners in Europe (TAM population) | 2,400,000 | 4,000 × 600 |
| Annual licence price per seat | EUR 150 | Own pricing interviews with 12 department heads, plus list prices of two comparable simulation providers |
| TAM | EUR 360 m per year | 2,400,000 × EUR 150 |
| Restrict to the German-speaking market, the language of our content | 110,000 medical students | National statistics offices for Germany, Austria and Switzerland |
| Restrict to the clinical phase, where our scenarios apply (35 percent) | 38,500 | Study regulations, share of semesters in the clinical phase |
| Add residents in emergency medicine and anaesthesia, our two scenario libraries | 9,000 | Medical association membership statistics by specialty |
| SAM | 47,500 seats = EUR 7.1 m per year | 47,500 × EUR 150 |
| Top-down share we claim for year one | 3 percent | Judgement, benchmarked against first-year penetration of two comparable edtech entrants |
| SOM, top-down | EUR 213,000 | EUR 7.1 m × 3 percent |
The same market sized bottom-up, built only from things inside our own control:
| Line | Assumption | Source I would cite |
|---|---|---|
| People selling in year one | 2 founders | Own resource plan |
| Qualified conversations each can run per year | 20 | Own log of the first 6 months, given a decision cycle of about 9 months in public hospitals |
| Share converting to a paid pilot | 25 percent | Own conversion so far, 3 pilots from 12 conversations |
| Paid pilots won | 10 | 2 × 20 × 25 percent |
| Average pilot size, one department | 80 seats | Median department size from the interview sample |
| Revenue per pilot | EUR 12,000 | 80 × EUR 150 |
| SOM, bottom-up | EUR 120,000 | 10 × EUR 12,000 |
Commenting on the gap. The top-down answer (EUR 213,000) is roughly 1.8 times the bottom-up answer (EUR 120,000), and the whole gap sits in the 3 percent line, which is the only figure in either table that nobody can check. The bottom-up chain, by contrast, can be attacked one link at a time: challenge the 20 conversations, or the 25 percent conversion, or the 80 seats, and I can answer with a log or an interview. So the number I would put into the plan is EUR 120,000 for year one, with the top-down EUR 213,000 shown beside it as what the same market would yield if sales capacity roughly doubled. That is a far stronger position than quoting the bigger number and hoping nobody asks how the 3 percent was chosen.
Apply it to your project
Section titled “Apply it to your project”-
Segment before you describe anybody. Cut your market on criteria that relate to your idea - pain points, usage behaviour, purchasing behaviour - and not on age and income just because those were easy to find. Write down how many groups you found and roughly how large each one is.
-
Name the groups you cannot address. Be explicit about which parts of the market are out of reach for reasons of language, regulation, channel or price. This closes the neglecting flaw, and the excluded list is also the first half of your SAM calculation.
-
Build one persona per priority segment. Answer the four questions in order, keep the demographics short, and spend the space on pain points, usage behaviour, preferences, willingness to pay, likelihood of early adoption and the channels they actually use.
-
Check every persona against the six flaws. Is there data behind it, or a good guess? Is it an existing customer or your ideal one? Is it overloaded? Is it just a demographic label? Would the members of that group really behave alike?
-
Size the market top-down and bottom-up. Build the TAM, narrow it to the SAM using your own segment definitions as the narrowing steps, cite a source for every line, then rebuild the first-year number bottom-up from your own sales capacity and reconcile the two.
-
Pick the first target customers by name. Look for lead users, innovators and early adopters, pilot and launching customers. Ask whether your opinion leaders are also the innovators - if the base is conservative, the second tier is the better bet. Weigh referral, influence, reference, knowledge and innovation value alongside deal size.
-
Run the killer hunt, then update. Assume the value proposition holds, list what could destroy the case, and plan the analysis for each. Then treat the whole segmentation as a hypothesis and revise it after every batch of customer conversations.
Key terms
Section titled “Key terms”| Term | What it means in plain words |
|---|---|
| Segmentation | Splitting a heterogeneous customer population into groups whose members need, behave and buy alike |
| Persona | A concrete, readable portrait of a typical customer inside one segment, built from evidence |
| Heterogeneity flaw | The persona does not represent a group that really shares use behaviour, expectations and preferences |
| Reverse engineering | Inventing the segment that perfectly matches the idea you already have, instead of finding it in the market |
| Pain point | A current, real problem of the user or customer, stated as their problem and not as your feature |
| Predicted response | How the persona is expected to react: likelihood of early adoption, customer lifetime value |
| TAM | Total addressable market: everyone who could in principle use this kind of solution |
| SAM | Serviceable available market: the part of the TAM your defined market, product scope and channels can serve |
| SOM | Serviceable obtainable market: the slice of the SAM you can realistically win in a period, and the basis of the sales prediction |
| Top-down vs bottom-up sizing | Cutting a published market total down with justified shares, versus building the number up from reachable customers, conversion, price and frequency |
| Lead user, pilot or launching customer | Names for the first adopters: resourced, knowledgeable, risk-tolerant, willing to co-develop |
| Opinion leader | A high-expertise, high-status, well-networked person who spreads information by word of mouth - not necessarily an innovator |
| Customer knowledge value | The worth of what you learn from a customer about preferences, quality problems and competitors |
| Killer | A factor that could destroy the entire business case or seriously inhibit adoption, as opposed to a minor threat |
Test yourself
Section titled “Test yourself”- Why should segmentation come before persona building, and what specifically goes wrong when the order is reversed?
- Name the four questions a persona should answer, and give one attribute you would record under each for a business customer.
- Which three attribute families usually dominate persona descriptions, which six should be more central, and what is the test that decides whether an attribute belongs in the persona at all?
- Sizing. A venture sells an annual seat licence at EUR 200. The TAM population is 900,000 learners. Language and product scope restrict it to 12 percent of that population, and within that only the 40 percent in the relevant training phase can use it. Compute the TAM and the SAM in euros. If the bottom-up plan is 3 sellers each closing 6 deals of 50 seats, what is the bottom-up SOM, and what share of the SAM is that?
- In a conservative customer base, why are the opinion leaders often the wrong first customers, and who is usually the better choice?
- Two teams present the same market. Team A shows a TAM of EUR 4 billion with no working shown. Team B shows a SAM of EUR 6 million with a source on every line. Which is the stronger presentation, and why?
Revision summary
Section titled “Revision summary”Next: Validating the Value Proposition → - getting real evidence from real customers.