This empathy map template places one person's words and actions beside possible thoughts and feelings. Keep a source beside each note, and label your interpretation so someone else can question it. The worked example turns an apparent contradiction into a question to explore.
Start with one person and a short evidence set
The original teaching example is a fictional train-ticket kiosk walkthrough, U1. It supplies two quotes and two observations:
| Source | Supplied record |
|---|---|
| Q1 | “I usually use this machine because I know the steps.” |
| O1 | The participant opens Help before confirming a changed fare. |
| Q2 | “I worry I might buy the wrong date when I am rushing.” |
| O2 | The participant rereads the displayed date before confirming. |
This is a map of U1's supplied experience. It gives the team something to discuss without turning one person's account into a claim about all commuters.
The Nielsen Norman Group guide to empathy mapping describes the Says, Thinks, Does, and Feels quadrants, with evidence traceable to research. It also recommends investigating contradictions and treating missing evidence as a reason for further research.
Fill the quadrants without changing the evidence
The completed map uses these entries:
| Quadrant | Map entry | Basis |
|---|---|---|
| Says | “I usually use this machine because I know the steps.” | Direct quote Q1. |
| Says | “I worry I might buy the wrong date when I am rushing.” | Direct quote Q2. |
| Does | Opens Help before confirming the changed fare. | Observation O1. |
| Does | Rereads the displayed date before confirming. | Observation O2. |
| Thinks | The selected date must match my journey. | Researcher's interpretation of Q2 and O2. |
| Feels | Worried about choosing the wrong date while rushing. | Feeling reported in Q2. |
Select a note in the example to inspect its source and the reason for its placement. The interpretation in Thinks remains visibly different from a direct quote. A teammate can accept the observation while questioning that interpretation.
A note can fit more than one area. The example lets you move an interpretation between Thinks and Feels while keeping its wording, source, and interpretation label. Discuss what the note adds; the quadrant label should not turn an inference into something the person actually said.
Replace an unsupported label with a useful note
Choose Compare wording to inspect this change:
| Unsupported entry | Evidence-based replacement |
|---|---|
| Hates technology. | Worried about choosing the wrong date while rushing, as reported in Q2. |
Nothing in the supplied record says the participant hates technology. The replacement names the concern and the context in the participant's own account. It gives the team a specific issue to investigate without inventing a general attitude.
If a teammate suggests another explanation, record it as a question or interpretation. Keep Q1, Q2, O1, and O2 available so the team can see which parts are established by the teaching record.
Use the apparent contradiction to plan a follow-up
The participant describes familiarity in Q1 and opens Help in O1. That contrast does not tell you why Help was needed. Perhaps the changed fare raised a new question; the supplied record has no answer.
The example's Follow-up view puts the two records together and offers:
What were you checking when you opened Help?
Leave the actual response blank until you ask. “Was the fare change confusing?” would suggest an explanation before the participant has given one.
After a real follow-up, add the response as a new source. Revise the interpretation if needed, while retaining the earlier notes so a reader can understand what changed.
Make a map from your own session
Create one complete, self-contained HTML empathy map.
Person or participant code and session context: [FILL IN]
Session date and research reference: [FILL IN]
Permitted direct quotes with source IDs: [PASTE]
Observed actions with source IDs: [PASTE]
My interpretations and their supporting source IDs: [PASTE]
Reported feelings and their source IDs: [PASTE]
Apparent contradictions or gaps to explore: [PASTE]
Build a Says, Thinks, Does, Feels map for the supplied person.
Keep direct quotes unchanged. Distinguish observed actions, reported
feelings, and researcher interpretations. Put a source ID beside each
note and mark an unsupported entry as an assumption to review.
Do not invent motives, personality traits, emotions, or participant data.
Let a reader select each note to inspect its source and rationale.
Allow an interpretation to move between Thinks and Feels without
changing its source or making it a direct quote. Keep the original
source record visible separately from map edits.
Show supplied contradictions beside neutral follow-up questions.
Leave answers blank unless provided. Do not assign sentiment scores
or generalize the map to a population. Label teaching data fictional.
Include a separate blank map and temporary notes. Name missing context
or note evidence before building a review copy. Preserve notes when
switching views, mark an old review stale after edits, and ask before
resetting typed work. Keep input in the open page only, without claims
that research records were saved or a participant was contacted.
Use text of at least 16px, labeled keyboard-accessible controls, and a
layout readable at 360px without horizontal scrolling. Put CSS and
JavaScript inline, with no external dependencies. Return the complete HTML.
Read the result beside the original notes. You should be able to point to the source of every quote and observation, and identify each sentence the team has inferred.
Let the team review the interpretation
For U1, ask whether “the date must match my journey” accurately expresses the interpretation the evidence supports. Then keep the Help question open for the next conversation with the participant.
If the team uses an HTML map for that discussion, NOS lets you clarify the map's opening instruction while viewing the map. Share the saved reading copy so colleagues can inspect the notes together; a correction is visible when they reopen the existing active link. Keep new participant responses and original research artifacts in your maintained research record.