When to use it
Use when the team has anchored too early on one idea, when ideation has stalled, or when you want unexpected directions quickly. Works especially well after a How Might We session.
It's particularly effective in The Spark — you have a seed and need to push it in multiple directions.
How it works in The Studio
Here's how a session works with WAiDE:
Sample output
Here's what a Crazy 8s session looks like in practice:
What you get
Eight rough ideas, typically in note form. One idea selected to carry forward — not safest, but most interesting. Session followed by a brief reflection on patterns across all eight.
Foundation
Created by Jake Knapp, John Zeratsky, and Braden Kowitz at Google Ventures. Published in "Sprint" (2016). Used by Google, Slack, Airbnb, Lego, Uber, and the NHS.
Why it works
Crazy 8s exploits a well-documented feature of creative cognition: the first ideas in any brainstorm are drawn from the most accessible mental pathways — common solutions, industry conventions, things you've seen before. The time pressure in Crazy 8s is specifically designed to exhaust these obvious options fast. By idea five or six, you've outrun your existing repertoire and must access something genuinely novel.
The eight-minute constraint is not arbitrary. It's short enough to prevent self-editing — the inner critic doesn't have time to shut down ideas — but long enough to move past the obvious. Jake Knapp and the GV team found this duration reliably pushes participants through the "I've run out of ideas" wall, which is actually the productive threshold where the interesting territory begins.
There's also a social science dimension. Studies on brainstorming in groups show that in unstructured settings, people anchor on the first idea voiced and evaluate rather than generate. Crazy 8s is deliberately solitary and simultaneous — everyone generates independently, preventing anchoring. The "most interesting, not safest" selection criterion then redirects group energy toward exploration rather than consensus.
The mechanism: Volume forces quality. When you must produce eight ideas and have no time to filter, you discover that your creative range is wider than you thought — and that the most interesting territory always starts after the comfortable options run out.
Frequently asked questions
What if I can't come up with eight ideas in eight minutes?
That's normal, and it's the point. The goal isn't eight polished ideas — it's eight attempts, including half-formed ones, wild ones, and ones you'd never actually pursue. WAiDE will keep you moving when you stall, and the rule is: if you're stuck, write down the worst idea you can think of. Bad ideas are data too, and they often surface better ones.
Why pick the "most interesting" idea rather than the "best" one?
"Best" is a convergent filter that rewards safety — ideas close to what the group already knows and trusts. "Most interesting" is a divergent filter that rewards novelty and surprise. At the ideation stage you don't yet have enough information to evaluate what's best; you need to explore a wider space first. The most interesting idea is often the one that makes you slightly uncomfortable, which is a reliable signal that it's genuinely new.
Can we do Crazy 8s as a team rather than individually?
The original format is individual and simultaneous for a reason — group brainstorming is vulnerable to anchoring on early ideas, status effects (deferring to the senior person's idea), and production blocking (only one person can speak at a time). Individual generation prevents these biases. If you want to use the outputs collaboratively, do a share-out and dot-vote after the solo round.
How does Crazy 8s fit into a broader ideation process?
Crazy 8s works best after a reframing step — a How Might We question or Five Whys session that has sharpened the problem statement. Without a clear challenge, the eight ideas tend to scatter in unrelated directions. After Crazy 8s, the selected idea typically moves into SCAMPER for further development, or directly into a rapid experiment to test its core assumption.