Empathy Mapping

Empathy Mapping

Reconstruct what another party sees, hears, thinks, feels, and needs, from their position rather than from your own assumptions about them.

Why

Predicting others' behavior or designing for their needs from your own vantage systematically imports your own priorities and blind spots. Deliberately reconstructing their actual situation catches what a self-referential guess misses.

Use when / Don't use when

• Use when: designing for users; preparing for a negotiation or difficult conversation; predicting a stakeholder's reaction.
• Don't use when: direct data — an actual interview, actual observed behavior — is available; use that over a constructed map.

Inputs → Outputs

• Inputs: a specific other party — an individual, role, or persona.
• Outputs: a structured picture of what they perceive, think, feel, and need, distinguished from what you assume or wish they thought.

Principles

• Ground the map in actual evidence about this party where available — a map built on stereotype is worse than useless.
• Separate what they say from what they likely think or feel — these often diverge.
• Their needs are usually more stable and more legitimate than their stated positions — map both, but weight needs higher for solution design.
• The map is a hypothesis to test against real behavior, not a conclusion.

Procedure

1. Specify the party precisely — a real individual or a well-evidenced persona, not a vague stereotype.
2. What do they see and hear: their actual environment, information, and context.
3. What do they think and feel: their likely internal state, informed by evidence, not projection.
4. What do they say and do: their observable behavior and statements, including where these seem to conflict with the thinks/feels layer.
5. What do they need: the underlying need behind their stated wants.
6. Use the map to generate hypotheses; verify against real interaction where possible, and flag which parts of the map remain unverified.

Common mistakes

• Building the map from assumption or stereotype instead of evidence.
• Conflating what they say with what they need.
• Treating the finished map as settled fact rather than a testable hypothesis.

Examples

• A UX research persona built from interview data.
• Preparing for a negotiation with an unfamiliar counterparty.
• A policy designed around a mapped constituency's actual situation.

Related

• [[active-listening]] — the real-time counterpart used once actual interaction begins.
• [[stakeholder-mapping]] — the group-level version of this same reconstruction.
• [[interview-synthesis]] — the mechanism for grounding this map in real evidence.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: skyf0xx
• Source: skyf0xx/better-thinking
• License: MIT
• Homepage: https://zenbin.org/p/better-thinking

Install and usage instructions live in the source repository linked above.