Affinity diagram tool, affinity diagram maker and affinity mapping software compared

An affinity diagram is a chart that sorts a large pile of separate notes into groups of related items, with each group given a name that says what ties it together. It is used to turn something unmanageable, sixty interview quotes or a wall of brainstorm output, into a handful of themes a team can actually act on.

The terms around it get used loosely, and it is worth being precise. Affinity mapping is the process of doing the sorting. An affinity diagram, affinity map or affinity chart is the artifact that comes out of it. The KJ method is the same thing under its original name. All four labels point at one activity: group what is similar, name each group, then decide what the names are telling you.

The reason it matters is that raw output is not evidence. Sixty notes on a wall look like a productive session and support no decision at all, because nobody can hold sixty items in mind at once. Six named themes with counts attached is a finding: it tells you where the room's attention actually went, which is frequently not where anyone assumed it went.

Name a theme after the shared underlying cause, not the surface topic, and write it as a short statement rather than a category label. "Users cannot tell whether their upload worked" is a theme name. "Onboarding" is a filing cabinet drawer. The test is simple: if the name would fit half the other clusters too, it is too vague to be useful.

Group first, name second. Nielsen Norman Group's guidance is explicit on this, and it matters more than it sounds: deciding the labels before you sort means you spend the session filing notes into categories you already believed in, which guarantees you find what you expected. Sort by similarity while resisting the urge to name anything, and only once the piles have stopped moving do you ask what each one is about.

Four practical rules for the naming pass. Use the participants' own words where you can, because a theme named in customer language survives the trip to a stakeholder deck intact. Keep names to under about eight words. Give every cluster a count, since a theme of eleven notes and a theme of two are not equally interesting and the label alone hides that. And leave outliers alone rather than forcing them into the nearest group, because the single note that fits nowhere is quite often the most valuable thing on the wall.

This is the step Brainstormer does not do for you, and it is worth being straight about that. The studio files every idea under the persona that produced it, so the wall arrives sorted by angle rather than as one undifferentiated pile. Grouping those ideas into themes and naming them stays your judgment call. What changes is that you are naming piles of real material instead of working out where to start on a blank wall.

Five steps: get every idea onto its own separate note, sort the notes into piles by similarity in silence, name each pile after what unites it, count the notes in each, then rank the themes and decide what happens next. Working silently through the sorting stage matters, because discussion during sorting turns the exercise into a negotiation about categories.

Two mechanics decide whether the output is any good. One idea per note is non-negotiable: a note carrying three thoughts cannot be placed in one pile honestly and will distort whichever group it lands in. And the sorting should be done by multiple people moving notes at the same time rather than by one person narrating, which spreads the interpretive work and surfaces disagreements early instead of at the readout.

Done by hand on a wall or a canvas, a fifty-note affinity diagram takes a competent team about an hour, most of which is spent hovering, squinting at duplicates and arguing about whether something is onboarding or activation. The full manual method, including when doing it by hand is genuinely the right call, is in our affinity mapping guide.

The table above splits the market by the only distinction that changes your afternoon: whether the tool gives you a canvas to sort on, or does the sorting. Every whiteboard in the list, Miro, Mural, FigJam and Lucidspark, is in the first category. They are good canvases, they have affinity templates, and you still do all the dragging and all the naming yourself.

Pick a whiteboard when the sorting itself is the point, which is a real scenario: a team that argues its way through a wall of research notes together ends up with shared understanding that no automated grouping can hand them. Pick a generative tool when you do this weekly, when the volume is high enough that the manual pass is the bottleneck, or when you are working alone and there is nobody to argue with anyway. If you are already paying for one of the whiteboards, our verified pricing breakdowns are on Miro pricing, Mural pricing, FigJam pricing and Lucidspark pricing.

Jiro Kawakita, a Japanese cultural anthropologist, developed the method in the 1960s as a way to make sense of unstructured field data gathered across different cultures. It is still widely called the KJ method after his initials, particularly in quality management, where it became one of the standard seven management and planning tools.

The origin explains the method's shape. Kawakita was not trying to run a workshop; he was trying to find structure in a mass of observations without imposing a taxonomy on them first. That is why the discipline of sorting before naming sits at the heart of it, and why the technique transferred so cleanly from anthropology to Six Sigma to UX research. All three deal with the same problem: a pile of specific observations, and a need to find the pattern without deciding the pattern in advance.

Affinity mapping is the workhorse of every design sprint: get everything on the wall, then silently move stickies until themes emerge. It works, and it costs an hour of hovering, duplicate-squinting and "is this onboarding or activation?" debates. The insight was never the sorting; it was the shape of the sorted wall.

Brainstormer does not do this sort for you, and it is worth saying plainly because several tools in this category imply they will. What it does is remove the reason the wall is usually a mess: it returns around twenty ideas already tagged with the persona that produced them, so you start the sort with structure rather than with forty undifferentiated stickies. The full method, including when to do it by hand, is in our affinity mapping guide.

Counts tell you where the wall's energy went: eleven ideas about onboarding and two about pricing is itself a finding. Cluster names become the language of the follow-up ("we are betting on the win-back theme"), and the structure feeds directly into idea prioritization: themes compete first, then ideas within the winning theme, which is how good facilitators run convergence by hand.

The same loop works by hand on any wall: diverge, group what is there, then push the strongest group through another diverge round if it deserves one. That rhythm is the divergent-convergent shape the double diamond is drawn from, and every Brainstormer run is one pass of it.

Solo, grouping is how you find out what you actually think: twenty scattered ideas become three named positions. Brainstormer does that narrowing for you at the end of every run, lifting a shortlist out of the ideas and naming one winner with reasons.

The studio above shows a sample run with its winner lifted. A free account gets 3 runs a month; paid plans start at $16 per month with no cap.