Empathy Maps: How Founders Use Them to Understand Customers

An empathy map is a simple visual grid, created by Dave Gray at the consultancy XPLANE, that organizes what you learn about a customer into sections: what they say, think, do, and feel, plus what they see and hear. It turns scattered interview notes into one shared picture of a real person.

Quick Answer: An empathy map captures a customer's inner and outer world on a single page — their words, thoughts, actions, and emotions. Fill it from real research, not imagination, and it becomes a shared reference your whole team can point to instead of arguing over who the customer "really" is.

Most founders leave a customer interview with pages of notes and no clear picture. One quote about pricing sits next to a story about a broken workflow, next to an offhand complaint about a competitor. The raw material is there, but it is a pile, not a portrait.

The empathy map is a container for that pile. Dave Gray designed it as part of a workshop toolkit — a way to get a whole room to build one honest, agreed-upon model of a customer instead of each person carrying a different, unspoken version in their head. It does not require new data. It requires you to sort the data you already have into a shape your brain and your team can actually use.

This guide covers what each section captures, how the classic four-quadrant version relates to the extended and updated versions, and how to build one from a real interview transcript without inventing anything. It sits inside our complete guide to customer research for founders, and it pairs naturally with a rigorous interview method like the Mom Test approach to customer interviews, which produces the honest raw material an empathy map needs.

Why an Empathy Map Turns Interview Notes Into Shared Understanding

An empathy map matters because it converts private, scattered impressions into one external artifact the whole team can see and challenge. Instead of five people quietly imagining five different customers, everyone points at the same wall and works from the same evidence.

That shift — from internal assumption to shared reference — is the real value. Interview notes live in one person's document and one person's memory. The map lives on a whiteboard or a shared canvas where a disagreement becomes visible and fixable.

Three things happen once the customer is on the wall:

• Assumptions get exposed. When someone writes "customers feel overwhelmed" in the Feels section, the natural next question is "which interview showed that?" Empty or unsupported cells reveal exactly where your knowledge is thin.
• Emotion stops getting dropped. Founders are trained to track features and requirements. The Feels and Thinks sections force you to keep the human's emotional state in view, which is often where the real buying motivation lives.
• The team aligns on one person. Designers, engineers, and marketers each tend to build a private mental model of "the user." A single shared map replaces those competing models with one you all agreed to.

Gray built the tool for exactly this alignment problem. It is deliberately low-fidelity — a grid, sticky notes, a marker — because the point is fast shared sense-making, not a polished deliverable. The moment it becomes a beautiful document nobody revisits, it has failed at its job.

One caution worth stating early: an empathy map is a synthesis tool, not a research method. It organizes evidence; it does not generate it. Fill it from real interviews, observation, and support tickets, and it sharpens your thinking. Fill it from imagination, and you have built a very tidy diagram of your own biases.

The Six Sections of an Empathy Map: Says, Thinks, Does, Feels, Sees, Hears

An empathy map is organized into sections around a picture of the customer, and the exact number depends on which version you use. The classic version has four quadrants — Says, Thinks, Does, Feels. The extended version adds Sees and Hears, and the updated version adds Pains and Gains at the bottom.

Knowing which version you are using prevents confusion, because the same tool is drawn several different ways across the internet. The table below lays out each section, what it captures, and a prompting question to fill it. No numbers here — an empathy map is qualitative by design.

The takeaway: Says, Thinks, Does, and Feels describe the person directly, while Sees and Hears describe the world pushing on them. Pains and Gains, added in Gray's later "updated empathy map," pull the sharpest insight from all the other sections down into a summary of what the customer is trying to escape and achieve. Start with the four classic quadrants; add the others when your research supports them.

The Says Quadrant Captures the Customer's Actual Words

The Says quadrant records what the customer said out loud, ideally as verbatim quotes rather than your paraphrase. This is the most concrete section and the easiest to fill honestly, because you either heard the words or you did not.

Direct quotes matter more than summaries here. "I've tried three tools and I always go back to my spreadsheet" carries texture — a note of resignation, a hint of what "good enough" means to them — that "user prefers spreadsheets" strips out. Keep the customer's own phrasing.

Useful prompts for this section:

• What did they say when asked how they handle this today?
• What words did they use to describe the problem, in their own vocabulary?
• What did they repeat, or return to more than once?

A word of caution: what people say is the least reliable of the sections when it stands alone, because people say what is socially acceptable or what they think you want to hear. The Says quadrant becomes powerful only when you compare it against the Does quadrant, where their actual behavior lives. A gap between the two is one of the most valuable things an empathy map can surface.

The Thinks Quadrant Captures Unspoken Beliefs and Worries

The Thinks quadrant captures what occupies the customer's mind that they may not say aloud — private beliefs, doubts, and preoccupations. It overlaps with Says on purpose, and the interesting entries are the ones that appear here but never made it into speech.

You cannot read minds, so this section is an inference from evidence, not a guess. If a customer hesitated before answering a question about budget, said "it's fine" in a flat tone, then changed the subject, you have grounds to note "worried this is another expense that won't pay off" in Thinks — flagged as an inference. If you have nothing but your own imagination, leave the cell empty.

Prompts that help:

• What might they be thinking but consider rude or risky to say?
• What doubts or fears sit under their polite answers?
• What do they believe about people who solve this problem the "right" way?

The gap between Says and Thinks is where a lot of buying friction hides. A prospect who says "looks great, send me the details" while thinking "I've been burned by tools like this before" needs a very different follow-up than the words alone suggest.

The Does Quadrant Captures Observable Actions and Behavior

The Does quadrant records what the customer actually does — the observable actions, workarounds, and sequences of behavior. It is the factual counterweight to Says, and behavior is far harder to fake than an opinion.

This is the section that rewards observation over conversation. If you can watch someone work — screen-share while they do the task, or walk their actual process — you fill Does with gold: the tab they keep open, the spreadsheet they paste into, the manual step they do every Monday. Where you cannot observe, reconstruct behavior from stories about the past rather than intentions about the future.

Prompts to use:

• What did they physically do the last time this problem came up?
• What tools or workarounds do they reach for, in what order?
• What behavior contradicts what they told you they do?

That last prompt is the whole point. When Says and Does disagree — "I always plan my week ahead" versus "actually I decide each morning" — believe Does. Behavior is evidence; stated intent is a hope. This is also why an empathy map built entirely from a single interview is weaker than one that incorporates observation, support logs, and usage data.

The Feels Quadrant Captures the Customer's Emotional State

The Feels quadrant captures the customer's emotions — their state of mind, and how the problem makes them feel. Founders often skip or rush this section, and it is frequently where the actual motivation to buy or switch is hiding.

Emotions drive decisions more than features do. Someone does not switch tools because a competitor has one more checkbox; they switch because the current tool makes them feel anxious every time a deadline nears, or embarrassed when a client sees the mess. Name the emotion, and give it a cause: not just "frustrated," but "frustrated that a task they consider simple keeps eating their evenings."

Prompts for this section:

• What emotion showed in their tone, pace, or body language?
• Does this problem make them feel anxious, embarrassed, hopeful, defeated?
• What would relief actually feel like for them?

Look for emotional evidence in how someone speaks, not just what they say — a sigh, a laugh, a sudden burst of energy when a topic comes up. These cues rarely appear in tidy notes, which is one more reason to capture emotion while an interview is fresh. Understanding this emotional layer is also what separates an empathy map from a demographic-driven profile, a distinction we cover in empathy map vs. persona.

The Sees and Hears Quadrants Capture the Customer's Environment

The Sees and Hears sections, added in the extended version of the empathy map, capture the external world acting on the customer — what they observe around them and what other people tell them. They move the map beyond the individual to the environment shaping their decisions.

Sees covers the customer's visual and market environment: the competitors and ads they encounter, what their peers appear to be using, what their day physically looks like, and the options already in front of them. It answers "what is this person exposed to that frames how they think about the problem?"

Hears covers the voices influencing them secondhand: what their boss says matters, what colleagues recommend, what industry figures or communities they trust are saying. A B2B buyer rarely decides in isolation — they hear a CFO worried about cost and a peer raving about a rival tool, and both shape the choice.

Prompts for these two sections:

• Sees: What does their market, workspace, and set of available options look like?
• Sees: Which competitors or alternatives are visibly in their world already?
• Hears: Who influences them, and what are those people saying?
• Hears: What messages reach them from their team, their network, and their industry?

Together, Sees and Hears explain why the same objective problem produces different behavior in different people. Two founders with identical bookkeeping pain will act differently if one keeps hearing "just hire an accountant" and the other keeps seeing ads for do-it-yourself software.

The Pains and Gains Sections Summarize Obstacles and Goals

The Pains and Gains section, from Gray's updated empathy map, sits along the bottom and distills the sharpest insights from every other section into two lists: what the customer struggles against, and what they want to achieve. It is the payoff row — the part you carry directly into product decisions.

Pains are the frustrations, fears, obstacles, and risks standing between the customer and a good outcome. What annoys them, what keeps them from getting the job done, what they are afraid might go wrong. These often echo entries you already wrote in Feels and Does, now stated as barriers.

Gains are the wants, needs, hopes, and measures of success. What outcome they are aiming for, and how they would personally know things got better. Gains are not features you might build — they are the results the customer already cares about, in their terms.

Prompts to draw these out:

• Pains: What obstacles, risks, or fears block their progress?
• Pains: What does "this is too hard" look like for them today?
• Gains: How does the customer define a win?
• Gains: What would make them say "this finally works"?

Because Pains and Gains synthesize the rest of the map, fill them last. Read across the completed quadrants, then pull the recurring threads down into these two lists. Done honestly, they become a checklist you can hold any product decision against: does this reduce a real pain or deliver a real gain the customer named? If it does neither, it is a feature looking for a justification.

How to Build an Empathy Map From Interview Transcripts

Build an empathy map by working backward from real research: gather your transcripts and notes, tag each observation to a section, cluster the patterns, and fill the sections from evidence only. The process is disciplined sorting, not creative writing.

Follow it in order, and resist the urge to fill empty cells with plausible-sounding invention.

1. Pick one specific person or one tight segment. An empathy map for "all our users" collapses into mush, because different segments say, do, and feel different things. Map one persona at a time — a single interviewee if the map is personal, or one clearly-defined segment if it is a composite.

2. Gather real evidence. Pull together interview transcripts, observation notes, support tickets, sales-call recordings, and reviews. If you have not talked to customers yet, stop here — building an empathy map from imagination is the single most common way founders misuse it.

3. Tag observations to sections. Go through your source material and assign each concrete observation to a section: a verbatim line goes to Says, a described behavior to Does, an audible emotion to Feels, an inferred worry (flagged as inference) to Thinks. One good interview can populate most of the grid.

4. Cluster the patterns. Group similar sticky notes. When three interviewees independently describe the same workaround in Does or the same fear in Feels, that cluster is a strong signal — far stronger than a single striking quote.

5. Fill Pains and Gains last. With the quadrants populated, read across them and distill the recurring obstacles into Pains and the recurring goals into Gains.

6. Mark the gaps. Note which sections stayed thin. An empty Hears section is not a failure of the map; it is an honest to-do list telling you exactly what to ask in your next round of interviews.

7. Revisit as you learn. Treat the map as a living artifact. Each new interview either reinforces a cluster or adds a note that challenges it. A map that never changes is a map nobody is using.

For a filled-in reference you can copy the structure from, see our empathy map template and example, which walks through a completed grid built from a realistic set of interviews.

Common Empathy Map Mistakes Founders Make

The most common empathy map mistakes are filling it from imagination, mapping too broad an audience, and treating it as a one-time deliverable. Each one quietly turns the tool from a research synthesis into a confirmation of what you already believed.

Watch for these traps:

• Inventing the content. The fatal error. A map filled with what you assume customers think, rather than what real customers said and did, is just your own bias in a nicer format. Every entry should be traceable to a source.
• Mapping "everyone." A map for an audience too broad to be coherent produces contradictions in every quadrant and insight in none. Narrow to one person or one tight segment.
• Skipping Feels and Thinks. Founders gravitate to Says and Does because they feel objective, and skip the emotional and cognitive sections. But the motivation to buy usually lives in Feels — leaving it blank strips out the most useful signal.
• Confusing it with a persona. An empathy map is not a demographic profile with an age and a stock photo. It is a snapshot of a mindset in a specific context.
• Treating it as a finished poster. A beautiful map that gets made once, hung up, and never touched again has stopped being a thinking tool. Keep it messy, shared, and current.
• Ignoring the Says-versus-Does gap. When stated words and observed behavior disagree, that contradiction is the insight — not a data error to smooth over. Founders who paper over it lose the map's single most valuable output.

The through-line: an empathy map is only as honest as the research behind it. The tool cannot manufacture insight you did not gather. It can only organize — and expose the gaps in — what you actually know.

Key Takeaways

• An empathy map is a shared visual model of one customer, created by Dave Gray at XPLANE, that sorts what a person says, thinks, does, and feels onto a single page.
• The versions build on each other: the classic map has four quadrants (Says, Thinks, Does, Feels), the extended map adds Sees and Hears, and the updated map adds Pains and Gains along the bottom.
• The map organizes research; it does not generate it. Fill every section from real interviews, observation, and support data — content invented from imagination makes the tool worse than useless.
• Says and Does are the honesty check. When a customer's stated words disagree with their observed behavior, that gap is one of the most valuable insights the map can surface.
• Feels is where buying motivation usually hides, so the emotional and cognitive sections deserve as much rigor as the factual ones, not a rushed afterthought.
• Pains and Gains are the payoff row, distilled last from the other sections into a checklist you can hold any product decision against.
• Empty cells are a feature, not a flaw — they tell you precisely what your next round of customer interviews needs to uncover.

Frequently Asked Questions

What is the difference between an empathy map and a persona?

A persona is a fictional profile summarizing a segment's demographics, goals, and background, while an empathy map is a snapshot of a single customer's mindset — what they say, think, do, and feel in a specific situation. Personas describe who someone is; empathy maps capture how they experience a problem. The two are complementary, covered in depth in our empathy map vs. persona comparison.

Who created the empathy map?

The empathy map was created by Dave Gray, founder of the consultancy XPLANE, as a tool for visual thinking and workshop-based sense-making. Gray later published an "updated empathy map" that reorganized the sections and added Goals along with Pains and Gains, refining the original four-quadrant format for clearer team collaboration.

What are the quadrants of an empathy map?

The classic empathy map has four quadrants: Says, Thinks, Does, and Feels, arranged around a picture of the customer. The extended version adds two more sections, Sees and Hears, to capture the customer's environment, and the updated version adds Pains and Gains at the bottom to summarize obstacles and goals.

How many customer interviews do I need before making an empathy map?

You can start an empathy map after a single strong interview, but a reliable one needs enough conversations to reveal patterns — commonly a handful within one segment, until you hear the same pains and behaviors repeat. Map one segment at a time, and treat thin sections as prompts for more research rather than cells to fill with guesses.

Can I build an empathy map without talking to customers?

No — not a useful one. An empathy map organizes real evidence about a customer; without interviews, observation, or support data, you are only diagramming your own assumptions. If you have not spoken to customers yet, run interviews first, then map. Building the grid from imagination is the most common way founders misuse the tool.

When should I use an empathy map versus a journey map?

Use an empathy map to understand a customer's mindset at a single moment — their thoughts, feelings, and motivations — and a journey map to trace their experience across multiple steps over time. Empathy maps go deep on one snapshot; journey maps go wide across a process. Founders often build an empathy map first to ground the human, then map their journey.