Brainstorming is one of the few management practices where the research has been clear for forty years and the practice hasn't moved at all. The finding, replicated many times since Diehl and Stroebe's 1987 review: the same number of people, working alone and then pooling their ideas, generate more ideas and better ideas than they do in a room together.
Not slightly more. Often substantially more. The group session feels more productive, which is exactly the problem.
Why the room loses ideas
Three mechanisms, and they compound:
Production blocking. Only one person can talk at a time. While you wait your turn, you're doing two things badly: holding your idea in memory so you don't lose it, and listening. Both compete for the same few slots of working memory, so by the time it's your turn the idea has either evaporated or flattened into something safer. This is the largest effect of the three, and it's purely structural — nothing about the group's talent changes it.
Evaluation apprehension. People withhold the half-formed idea, which is usually the interesting one. The senior person in the room makes this worse just by being present.
Free riding. In a group, effort is invisible, so it quietly drops. Nobody decides to coast; it just happens.
The fix: write first, talk second
Nearly all of the loss comes from generating ideas out loud, in sequence. So don't.
- State the problem precisely, in advance. A vague prompt produces vague ideas. "How might we cut onboarding from five steps to two?" beats "ideas for onboarding."
- Ten minutes of silent, independent writing. Everyone, alone, one idea per line, no discussion. Quantity is the only instruction. Nobody's idea can block anyone else's, because nobody is waiting for a turn.
- Pool everything unattributed. All lists into one place, duplicates left in — a repeated idea is a signal, not a redundancy.
- Now use the room. Groups are genuinely good at the thing they're usually not asked to do: reacting, combining, spotting the flaw, and choosing. Discussion is a converging tool, not a generating one.
This shape has a name in the literature — brainwriting — and its advantage is that the group's real strength (many perspectives) survives while its structural weakness (one mouth at a time) is removed.
Writing rules that raise the count
The silent phase is a note-taking problem, and a few small rules make a large difference:
- One idea per line. Same instinct as atomic notes: an idea you can move and recombine is worth more than a paragraph you can only admire.
- Don't edit while generating. Judging and producing are different modes, and switching between them every few seconds costs you both. Write bad ideas down deliberately; they're cheap, and they're often the stepping stone.
- Go past the point where it dries up. The first fifteen ideas are the ones everybody already had. The interesting ones tend to show up after the obvious well runs dry, which is precisely when most sessions stop.
- Branch instead of listing when the problem is shaped. For anything with structure, mapping the ideas surfaces relationships a flat list hides. If you're unsure which form fits, mind map vs outline breaks down when each wins.
Solo brainstorming has the same problem
Working alone doesn't exempt you: your own editing voice does what the group does. The same protocol works — timer, one idea per line, no judging until the timer stops, then a separate pass to cut.
And the reason it works at all is that the ideas leave your head as they arrive instead of being held there. Every unwritten idea is occupying capacity you need for the next one, so the habit of capturing fast is not a filing preference, it's what keeps the generator running.
Whatever comes out of the session, it's raw material, not conclusions. Treat the list as fleeting notes — most will be discarded, and the two or three worth keeping earn their place by being rewritten properly afterward.