Last reviewed on 2026-04-24
Classic open-floor brainstorming has a bad reputation for a reason. Whoever speaks loudest sets the direction, ideas get judged before they are out of people's mouths, and the quiet half of the team leaves the session feeling they did not get to contribute. The fix is not to stop brainstorming but to structure it so that the failure modes have nowhere to hide.
This guide walks through several brainstorming formats that small groups — up to about six people — can actually run without a facilitation background, plus a few rules that keep each one honest.
Most unstructured brainstorms crash because they try to do all three stages at the same time. Someone suggests an idea, someone else evaluates it immediately, a third person starts pushing for action. Separating the stages — even just by a loud "we are still diverging, we will judge later" — fixes half the problems on its own.
Brainwriting is the simplest fix for a noisy room. Instead of calling ideas out loud, everyone writes in silence for a fixed amount of time.
This works especially well in teams with a mix of extroverts and introverts. It also produces a larger raw pool — people generate more ideas on their own than they do waiting for a turn to speak.
Borrowed from design sprints. Each person folds a page (or divides a canvas tab) into eight regions, and has eight minutes to fill every region with a different idea — one minute per idea.
The constraint is the whole point. Once you have to fill eight slots in a minute each, self-censoring stops. Silly ideas sneak in. Some of them turn out to be useful; others open the door for a better idea nearby.
SCAMPER is a checklist of seven prompts for modifying an existing idea. It is useful when the team has a rough direction but cannot see alternatives. The letters stand for:
Walk through the seven prompts out loud. For each one, fill a colour-coded sticky on the canvas. SCAMPER is a generator of variations, not an invention engine — use it once you have at least one idea to riff on.
Ask the group to come up with the worst solutions they can think of. Genuinely bad. Insulting, impractical, expensive, illegal.
Two things happen. First, the atmosphere loosens up, because deliberately bad ideas are funny and low-stakes. Second, inverting a bad idea often produces a useful one. A "worst possible onboarding" that locks the user out of the app until they fill in a 40-field form points clearly at what makes a good onboarding.
This is a good warm-up technique when the team is being too cautious.
Affinity mapping is the clustering stage dressed up with a name. Take every raw idea on the board, read them silently, and drag similar ones together into clusters. Label each cluster with a one-line header. Duplicates collapse. Orphans reveal themselves.
Do this in silence first, then discuss. The silent version produces more honest groupings — people are not talking each other into categories that do not exist.
A fast converging technique. Give everyone a fixed number of votes (three is typical) and ask them to place a coloured marker on their preferred clusters. Whichever clusters get the most votes move forward.
Keep the vote count small. If people have seven votes each in a group of five, more than half the clusters will end up "voted for" and the signal disappears.
The purpose of structure in a brainstorm is not to be fancy. It is to stop the three usual failures: loud voices dominating, ideas judged before they are finished, and the session ending with nothing to act on.
Open a fresh tab. Put the prompt as a large note at the top. Colour-code the note categories — for example, yellow for ideas, pink for concerns, blue for questions. During the diverge stage, everyone in the room (or on a call sharing a screen) dumps ideas onto the canvas in their assigned colour. During clustering, drag related ideas close together. During converge, star the top ideas with a header note in a distinct colour.
When you are done, export the tab as a backup before wiping the canvas for the next session. Finished brainstorms are best kept as short summaries rather than a full archive of sticky notes.
Not every problem needs a brainstorm. A decision between two well-understood options is better made with a short pros-and-cons list. A well-defined execution task needs a plan, not more ideas. Brainstorming is a tool for the stage where the possible solutions are not yet clear. If you already know roughly what you are going to do, spend the meeting deciding and moving instead.