Chapter 5: Segmentation, Personas & Target Customers

Every product decision downstream -- positioning, messaging, feature priority, pricing, channel strategy -- depends on who you are building for. Segmentation is not a marketing exercise you do once and file away. It is the foundation that determines whether your research, your interviews, and your validation efforts will produce useful signal or useless noise. Get segmentation wrong and everything built on top of it will be wrong too.

---

Why Segmentation Is the Starting Point

Three of the source books treat segmentation as essentially step one of any serious product or business effort.

Bill Aulet opens Disciplined Entrepreneurship with what he calls "The First Commandment": the single necessary and sufficient condition for a business is a paying customer ^de. Before you can find that customer, you must decide which customers to pursue. Segmentation is how you make that decision.

Kazuki Nishiguchi frames it from the measurement side: you cannot develop strategy without understanding who your customers are, how they relate to your brand, and how those relationships differ across segments ^stck. His entire analytical framework -- the 5segs and 9segs maps -- is built on the premise that treating "customers" as a monolith leads to averaged-out strategies that resonate with nobody.

April Dunford approaches segmentation from the positioning side: your positioning only works when it is anchored to your best-fit customers ^{obviously-awesome}. If you do not know who those customers are, you cannot articulate why your product matters or what makes it different.

The common thread: segmentation is not optional. It is prerequisite.

---

Market Segmentation Brainstorm and Matrix

Aulet's approach to segmentation begins with a structured brainstorm ^de:

Step 1: Generate a Wide List

Start by listing every conceivable group of people or organizations that might benefit from your technology, passion, or idea. Cast the net wide. Do not filter yet. Aulet encourages founding teams to generate dozens of potential market segments -- different industries, use cases, demographics, geographies, and contexts where the core value proposition could apply.

Step 2: Build a Segmentation Matrix

Organize the raw list into a matrix with segments as rows and evaluation criteria as columns. Aulet suggests criteria including:

• Is there a well-funded, accessible customer? Can you actually reach them and do they have budget?
• Do they have a compelling reason to buy? Is the pain acute or merely inconvenient?
• Can you deliver a complete product to them today? Or does serving this segment require capabilities you do not have?
• Is there existing competition entrenched? How hard will it be to win?
• Does winning this segment give you leverage into adjacent segments? (This is the strategic criterion -- it connects to the bowling alley metaphor below.)
• Does this segment align with your team's values, passion, and capabilities?

Step 3: Narrow Ruthlessly

Use the matrix to eliminate segments that fail critical criteria. The goal is not to pick the biggest market but to pick the most winnable one -- a segment where you can achieve dominance, build a reference base, and then expand.

See Market Segmentation Matrix

---

The Beachhead Market: Why Narrow Wins

Aulet borrows Geoffrey Moore's bowling alley metaphor: your first market segment is the head pin ^de. Knock it down decisively and it creates momentum that topples adjacent pins (segments). Try to hit all the pins at once and you hit none of them.

The beachhead market is the single segment you will pursue first and dominate before expanding. Aulet is emphatic: this is not a "primary segment" among several you are working on simultaneously. It is the only segment you are working on. All resources -- product, sales, marketing, support -- focus here until you have achieved product-market fit within this segment ^de.

Characteristics of a Good Beachhead

• Small enough to dominate -- If the segment is so large that achieving meaningful market share would take years, it is too big. Narrow further.
• Large enough to matter -- The segment must support a viable business on its own, even if you never expand beyond it.
• Customers within it share word-of-mouth channels -- If customers in the segment talk to each other, your early wins generate organic referrals. If they do not, you have to re-sell from scratch to every new customer.
• Winning it creates strategic leverage -- Ideally, success in the beachhead earns you credibility, technology, or distribution advantages that make adjacent segments easier to enter.

Further Segmenting the Beachhead

Even after selecting a beachhead, Aulet warns that you likely need to segment further ^de. A "beachhead market" of "small law firms in the United States" is still far too broad. You might narrow to "solo immigration attorneys in mid-size Texas cities who currently manage case files in spreadsheets." The more specific your target, the more precisely you can design your product, your messaging, and your go-to-market.

---

Customer Slicing: Drilling to a Specific Who-Where Pair

Rob Fitzpatrick offers a more informal but equally effective narrowing technique he calls "customer slicing" ^mom-test. The idea is simple: keep splitting your target group until you arrive at a pair of answers:

1. Who specifically are they? Not "small business owners" but "freelance graphic designers with 2-5 years of experience who work remotely."
2. Where do they already congregate? Not "online" but "in the Dribbble community, at local AIGA meetups, and in the #freelance channel of Designer Slack groups."

If you cannot answer both questions, your segment is too broad. The "where" is especially important because it determines whether you can actually reach these people for interviews and eventually for sales. A perfectly defined segment you cannot access is useless ^mom-test.

Fitzpatrick's slicing is faster and less formal than Aulet's matrix approach, making it well-suited for very early-stage teams who need to start talking to customers this week rather than spending weeks on segmentation analysis.

---

The "Oh No" Test for Defining Who to Exclude

Michael Margolis and the GV research team use a powerful heuristic for sharpening recruiting criteria: the "Oh No" test ^sprint. Ask your team: "If a participant walked in for our interview, what characteristic would make you immediately say 'Oh no, this is the wrong person'?"

This negative framing is surprisingly effective. Teams often struggle to articulate who their ideal customer is but find it easy to identify who their customer is definitely not. Common "Oh No" triggers include:

• Someone who has never experienced the problem you are solving
• Someone who works in your industry and would evaluate the product as a professional rather than using it as a customer
• Someone who is a power user of a competitor and unlikely to switch under any circumstances

The "Oh No" test produces sharp exclusion criteria that make recruiting dramatically more effective. It complements the positive targeting of segmentation with equally important negative boundaries.

---

The 5segs Customer Pyramid

Kazuki Nishiguchi introduces a quantitative segmentation framework called the 5segs map (Customer Pyramid) ^stck. Rather than segmenting by demographics or firmographics, it segments by relationship to your brand:

Why This Matters

The strategic implications differ dramatically by layer. Turning an Average Customer into a Loyal Customer requires a different strategy than converting an Aware Non-Customer into a first-time buyer. And yet most companies lump all "customers" together and all "non-customers" together, applying one-size-fits-all strategies that serve nobody well ^stck.

The pyramid also reveals the health of your business at a glance. A brand with a large Loyal layer and a shrinking Defected layer is healthy. A brand with a growing Defected layer and a stagnant Loyal layer is in trouble, even if aggregate revenue has not declined yet.

See 5segs Customer Pyramid

---

The 9segs Map: Adding Purchase Intention

Nishiguchi extends the 5segs pyramid into the 9segs map by adding a second axis: Next Purchase Intention (NPI) ^stck. Each of the five layers (except Unaware) is split into two sub-segments based on whether the customer has high or low intention to purchase/use your product next.

This creates nine segments:

• Loyal + High NPI (your strongest advocates)
• Loyal + Low NPI (at risk of defection)
• Average + High NPI (potential to upgrade)
• Average + Low NPI (at risk of churn)
• Defected + High NPI (winback opportunity)
• Defected + Low NPI (likely gone)
• Aware Non-Customer + High NPI (warm prospects)
• Aware Non-Customer + Low NPI (brand awareness without interest)
• Unaware Non-Customers (a single segment, since NPI is not applicable)

Strategic Use

The 9segs map enables far more precise strategy development. For example, "Loyal + Low NPI" customers are your most dangerous segment -- they loved you once but are about to leave. Understanding why they are losing intention (through n1 analysis -- see below) lets you intervene before they defect ^stck.

The map is constructed from survey data, typically asking two questions: (1) a behavioral question about purchase/usage recency and frequency, and (2) an intention question about likelihood of next purchase. This makes it relatively cheap to build and update regularly.

See 9segs Map

---

Building Personas

A persona is a vivid, specific, research-grounded description of a representative customer in your target segment. Done well, personas align the team around a shared understanding of who they are building for. Done poorly, they become fictional characters that reflect the team's biases rather than customer reality.

The Persona Worksheet (Disciplined Entrepreneurship)

Aulet provides a structured persona template with specific fields ^de:

• Demographics: Age, gender, location, income, education, family status
• Psychographics: Values, attitudes, personality traits, lifestyle
• Professional context: Role, company size, industry, seniority
• Priorities: What are the top 3 things they care about most (not necessarily related to your product)?
• Day-in-the-life narrative: What does a typical day look like? Where do they spend time? What tools do they use? What frustrates them?
• Watering holes: Where do they go for information and community? What do they read, attend, subscribe to?
• Proxy products: What products do they already use that signal alignment with your target? (Someone who uses Notion, Figma, and Linear is a different person than someone who uses Word, Photoshop, and Jira, even if their job title is the same.)

The critical requirement: every field should be filled with data from actual customer conversations and observations, not assumptions ^de. A persona built from the team's imagination is worse than no persona at all, because it creates false confidence.

Research-Grounded Personas (Just Enough Research)

Erika Hall argues that the value of personas comes entirely from the research behind them ^just-enough. She warns against "elasticity" -- the tendency for teams to stretch a persona to fit whatever they want to build. To counter this:

• Base every persona attribute on observed patterns from real interviews
• Include direct quotes from research participants
• Specify what the persona does NOT care about, not just what they do care about
• Update personas when new research contradicts them

Hall also advocates for including the persona's context and constraints, not just their goals. Understanding that your persona is a time-pressed parent who does most product evaluation on their phone during a 15-minute commute is more useful than knowing their age and income ^just-enough.

N1 Analysis (Succeed Through Customer Knowledge)

Nishiguchi contributes a complementary technique: n1 analysis ^stck. Rather than building a composite persona from averaged research data, n1 analysis involves deep-diving into a single, specific, real customer who exemplifies a target segment from the 5segs or 9segs map.

The researcher conducts an extensive interview with this one person, mapping their complete journey with the brand: how they first became aware of it, what triggered their first purchase, what kept them coming back (or what drove them away), and what their current relationship looks like. The goal is not statistical generalizability but deep causal understanding of the mechanisms that drive behavior in that segment ^stck.

N1 analysis is especially powerful for the "Defected + High NPI" and "Loyal + Low NPI" segments, where understanding the specific story of one customer can reveal leverage points that survey data alone would miss.

---

Best-Fit Customers as Positioning Anchor

April Dunford approaches the "who" question from a different angle: instead of starting with market segmentation and working toward a product, she starts with your existing product and asks, "Who are the customers that love this the most?" ^{obviously-awesome}.

See Best-Fit Customer Analysis

These "best-fit customers" become the anchor for your positioning. The logic:

1. Identify customers who bought quickly, adopted fully, rarely churn, and actively refer others.
2. Study what they have in common -- not just demographics but characteristics, context, and pain that make your product uniquely valuable to them.
3. Position your product to attract more people like them.

This is a fundamentally different entry point than Aulet's top-down segmentation, and it is most useful when you already have a product in-market with some traction. You are not choosing a segment hypothetically; you are letting your best existing customers tell you who your segment is ^{obviously-awesome}.

Dunford warns against the temptation to position broadly to appeal to everyone. "When you try to be everything to everyone, you end up being nothing to anyone" ^{obviously-awesome}. Best-fit customers give you permission to be specific.

---

The Next 10 Customers Exercise

Once you have a beachhead market and a persona, Aulet recommends a concrete validation exercise: identify and list, by name, the next 10 customers you would sell to ^de. Not a segment description -- actual names of real people or organizations.

This exercise forces brutal honesty. If you cannot name 10 specific prospects, your segment definition is too abstract. If the 10 names you list have nothing in common with each other, your segment is too broad. And if you can name them but have no idea how to reach them, your "where" is missing ^de.

The Next 10 Customers list also becomes your initial interview and sales pipeline. These are the people you will talk to first, sell to first, and learn from first.

---

Common Mistakes

Too-Broad Segments

"Small and medium businesses" is not a segment. "Millennials" is not a segment. A segment must be specific enough that you can design a product, write a message, and choose a channel that resonates with everyone in it. If two people in your "segment" have nothing in common except one demographic attribute, the segment is too broad.

Demographic-Only Personas

Knowing that your persona is "Sarah, 34, lives in Austin, earns $85K" tells you almost nothing about what she needs, what she values, or how to reach her. Demographics are the least useful dimension of a persona. Psychographics (values, attitudes, priorities), behavioral patterns (what she actually does), and context (what constraints shape her decisions) are far more actionable ^{just-enoughde}.

Segments Without a "Where"

A perfectly defined segment that you cannot physically or digitally access is useless for both research and go-to-market. Always pair "who" with "where" ^mom-test. If you cannot find where your target customers congregate, you either need to redefine the segment or invest in discovering their gathering places before proceeding.

Persona Inflation

Creating too many personas dilutes focus. For most products, one to three personas is sufficient. If you have seven personas, you do not have seven user types -- you have a segmentation problem. Narrow your beachhead and reduce your personas to the minimum needed to capture meaningful behavioral variation ^de.

Static Segments

Segments and personas are not fixed. They should evolve as you learn. A team that builds personas in Month 1 and never updates them is building on an increasingly fictional foundation. Treat personas as living documents that get refined with every research cycle ^{cdhjust-enough}.

---

Synthesis: Three Approaches to Segmentation

The source books offer three distinct approaches to segmentation, each suited to different contexts:

These are not mutually exclusive. A startup might use Fitzpatrick's slicing to start interviews this week, then formalize with Aulet's matrix as they narrow toward a beachhead, and eventually adopt Nishiguchi's 5segs/9segs framework once they have enough customers to segment quantitatively. Dunford's best-fit customer analysis ^{obviously-awesome} slots in naturally once you have paying customers to analyze, providing a market-reality check against any top-down segmentation.

The common principle across all approaches: specificity wins. The more precisely you can describe who you are serving, the better every downstream decision becomes -- from what to build, to how to position it, to where to sell it.