Crazy 8s
Eight ideas, one per panel, one per minute, eight minutes — a brutal time constraint that forces you past your obvious ideas to the ones you did not know you had.
Your first three ideas are the ones everyone has. The method exists to drag you, fast, to idea number seven.
What it is
Not a brainstorm. A forcing function — a brutal time constraint designed to exhaust the obvious and reach the non-obvious.
Crazy 8s is a rapid, time-boxed ideation method in which each participant folds a sheet of paper into eight panels and sketches one distinct idea per panel, one every minute, for eight minutes. The brutal time constraint is not an inconvenience; it is the entire mechanism. Eight minutes for eight ideas is deliberately too little time to be careful, and that is the point: it forces quantity over polish, prevents premature attachment to any single concept, and, above all, pushes people past their obvious first ideas.
Its defining insight is counterintuitive and well-documented. The ideas teams actually select for prototyping are disproportionately drawn from panels five through eight, not one through four. The reason is structural: the first few ideas any person sketches are the obvious ones — the solutions already lurking in their mind, the conventional responses everyone in the room would also produce. It is only after those are exhausted, somewhere around the fifth panel, that people are forced to reach for something genuinely new. The late panels, where it gets hard and slightly desperate, are where the breakthrough ideas tend to live.
Crazy 8s is really a forcing function. It uses speed and a rigid constraint to do something people cannot reliably do on their own: exhaust the obvious and keep going into the non-obvious. Jake Knapp, who codified the method in the Design Sprint context, calls it “a forcing function for the ideas you did not know you had.” The rough sketches are almost beside the point; the value is that the constraint drags out ideas that a comfortable, unhurried brainstorm never would.
Fill the eight panels. Feel where it gets hard.
No single panel contains the breakthrough. It lives on the far side of the wall — click through to find it.
Click any panel to see what kind of idea lands there and why. The early panels produce the obvious; the late panels produce what you did not know you had. The wall is around panel five — that is exactly where it starts to get interesting.
When to deploy it
For fast divergence when you need genuine variety. Not for refinement, analysis, or problems with one correct answer.
Use it when
- →You have a well-scoped prompt — ideally a How Might We question — and need to generate a lot of divergent ideas fast.
- →The team is stuck on the obvious solutions and needs to be forced past them into genuine variety.
- →You want broad participation and many ideas quickly, without the polish or commitment of prototyping.
- →You are early in ideation and want quantity and range before converging.
Do not lean on it when
- ×The prompt is too vague to sketch against. Scope it first — Crazy 8s needs a clear challenge, not "improve the product."
- ×You need depth on one concept rather than breadth across many. Crazy 8s is a divergence tool, not a refinement one.
- ×The problem genuinely has one correct answer to be analysed rather than a space of possibilities to be explored.
- ×Participants are so blocked by "I can't draw" anxiety that the sketching itself prevents ideas. Reassure them first — the sketches are rough and private by design.
The honest limit: Crazy 8s generates raw, rough ideas at volume; it does not evaluate, refine, or validate them. Most of the eight panels will be bad — that is expected and fine. The method trades quality-per-idea for quantity and range, on the bet that a few late panels hold something worth developing. It is a divergence engine that must be paired with convergence to produce anything real.
How it works
Six moves, from a clear prompt to a selection of late-panel ideas worth developing.
Start from a clear, well-scoped prompt.
Crazy 8s sketches against a specific challenge, ideally a well-scoped How Might We question. A sharp prompt focuses the divergence; a vague one scatters it. Spend the setup time making sure everyone in the room is sketching against the same clear challenge before the clock starts. A minute clarifying the prompt saves eight minutes of off-target sketching.
Fold the sheet into eight panels and set the timer.
Each participant divides a sheet into eight panels. One minute per panel, eight minutes total. The rigid timing is non-negotiable — it is the forcing function. Protect it: run a visible timer, call the minute, and keep moving. A relaxed, drifting version of the timing loses the effect entirely.
Sketch one distinct idea per panel, fast.
One idea per minute, moving on whether or not the panel feels finished. Distinct ideas, not variations of one concept. Rough is required — polish is impossible in a minute, and that is the point. The sketches are meant to be ugly; they are capturing an idea, not demonstrating drafting skill.
Push through the wall past panel four.
Expect the obvious ideas to come first and run out around panel five. Keep going. The instruction that matters most: do not stop when it gets hard. That is exactly when the original ideas start. The time constraint refuses to let you stop at four ideas and call it done — honour the constraint.
Do it individually, then share.
Everyone sketches their own eight in silence, so the loudest voice does not shape the room. Individual divergence first, collective sense-making after. When the team shares, the full range of ideas becomes visible — including the late-panel ideas that no brainstorm converging on the first voiced idea would have reached.
Select and carry forward.
After sharing, the team identifies the most promising ideas — often, tellingly, from the late panels — and carries them into the next step: further sketching, prototyping, or co-creation. Crazy 8s ends by feeding convergence, not by choosing a winner on its own. The selection is where the method hands off to the next phase.
Best practices
What separates a session that reaches panel seven from one that stops at panel four.
When it goes well
- ✓The prompt is sharp and well-scoped, so eight minutes of divergence stays focused on the right challenge.
- ✓The time constraint is held strictly — one minute per panel — because the pressure is the entire mechanism.
- ✓People push through the wall past panel four and reach for the non-obvious ideas in panels five through eight.
- ✓Everyone sketches individually first, so the room produces genuine variety rather than an echo of the loudest voice.
- ✓The team treats rough as required and judgment as deferred — generating first and evaluating later.
The mistakes, and how to avoid them
Stopping at the obvious ideas.
Quitting mentally around panel four, when it gets hard, forfeits exactly the late panels where the breakthroughs live. Push through the wall — that is the point of the exercise. The instruction "keep going when it gets hard" is not encouragement; it is the mechanism.
Letting people polish.
Trying to make a panel look good burns the minute and defeats the quantity-over-polish mechanism. Keep it rough and fast. Remind participants before the session starts: the sketches are private, ugly is correct, the idea matters and not the drawing.
Softening the time constraint.
Giving "a little more time" removes the forcing function and lets people retreat to comfortable, obvious ideas. Hold the timer. The eight minutes is not a guideline.
Sketching as a group too early.
Diverging together lets the first loud idea anchor everyone. Sketch individually first, share after. The moment the group starts converging before each person has finished their own eight, the method loses its core mechanism.
Judging in the moment.
Evaluating ideas as they are drawn kills the divergence. Defer all judgment to the sharing-and-selection step. The rule is: generate first, evaluate after, always.
Letting "I can't draw" stop ideas.
Sketching anxiety blocks participation. Reassure the team before the session: the sketches are rough, private, and about ideas, not art. Show an example of what an acceptable panel looks like — a few lines and a label is enough.
Logistics
Keep it dead simple. The beauty of Crazy 8s is that it needs almost nothing.
Crazy 8s needs paper, a pen, and a timer. That is almost everything. Resist adding complexity. The low setup is part of why it is so reliable and repeatable — it can run anywhere, with any team, against almost any prompt, in eight minutes plus the time to share and select.
Scope the prompt before the clock starts
Spend the setup time making sure everyone is sketching against the same clear, well-scoped challenge. A minute clarifying the prompt saves eight minutes of scattered, off-target sketching. The most common failure mode in Crazy 8s is a prompt so vague that the eight panels produce eight different interpretations of the problem rather than eight different solutions to it.
Protect the timing ritual
Run a visible timer and call the minute. The strict cadence is what creates the productive pressure; a relaxed, drifting version loses the effect entirely. One minute per panel, no negotiation. Calling the time — even when participants are still mid-sketch — is not harsh; it is the mechanism.
Set expectations about roughness and volume
Tell participants up front that the sketches will be ugly, that most of the eight will be bad, and that both are expected and fine. This defuses the "I can't draw" block and gives permission to reach for the strange late ideas. A participant who understands that roughness is required produces more varied panels than one who is trying to make each one presentable.
Plan the share and selection in advance
Decide in advance how the team will share sketches and select promising ideas afterward. Silent review with dot-voting is common. Crazy 8s produces raw material; the share-and-select step turns it into direction. Without a planned selection process, the session ends with a pile of sketches and no clear next step. The selection is part of the method.
In person or remote, analog or digital
Crazy 8s works in person (paper, folded) and remotely (a shared digital canvas such as a collaborative whiteboard tool). In remote sessions, the digital canvas allows the team to share sketches in the same space without the delay of a photo. Named as common examples, not endorsements: the method is the constraint and the structure, not the tool.
AI and this method
AI can generate a hundred ideas in seconds. That is not what Crazy 8s was ever for.
Toggle between modes to see what AI produces when it replaces the eight minutes — and why more ideas is not the same as reaching idea seven. The value of Crazy 8s was never the raw count; it was the constraint and the ownership.
In-depth example
The robotics company: the panel-seven idea that almost wasn’t drawn.
A team running Crazy 8s for a robotics product pushes through the wall and discovers a personality-forward direction in the late panels that they had nearly self-censored. The one-minute constraint forced it onto the page. It became the product’s defining characteristic. Toggle to see what a hypothetical AI-first approach would have produced.
The session: eight panels, eight minutes
The team ran Crazy 8s against a well-scoped prompt. Each person folded a sheet into eight panels and sketched one direction per minute for eight minutes. No polish, no judgment in the moment, one distinct idea per panel.
The early panels went as they always do. The obvious robotics directions came fast and easily — sensible, functional, the things the whole category was already building. Reliable to sketch, unremarkable to select.
The wall: around panel five
Around panel five, the obvious directions ran out. It got hard. For a moment, nothing came. In a comfortable, unhurried session, this is where most people stop — satisfied with their first four, never reaching what they did not know was in there.
The one-minute constraint refused to let the team stop. The pen kept moving. Panel five produced something slightly desperate; panel six something strange. And in one participant’s late panels, something unexpected appeared.
The late-panel idea: panel seven
Forced past the conventional robotics directions, a participant sketched a personality-forward design direction — a playful, character-driven take that they had almost not bothered to draw. “Too playful for a serious robotics product.” The one-minute constraint forced it onto the page before judgment could kill it.
The share-and-select step: that sketch stood out
When the team shared their eight panels, the personality-forward sketch stood out immediately. It was the only direction in the room that was genuinely differentiating — not because it was more polished or more logical, but because it was the only one that had not come from the obvious zone. It went on to become the product's defining, differentiating characteristic.
The forcing function and the ownership
The direction was a late-panel idea that a comfortable brainstorm would never have surfaced — and that the participant themselves would have discarded if given the chance to judge it first. The time constraint dragged out an idea the team did not know it had. And because a team member had sketched it themselves, they understood it, believed in it, and championed it.
Jake Knapp: a forcing function for the ideas you did not know you had
The panel-seven sketch was the right idea. It was also exactly the idea that would have been self-censored out of any comfortable brainstorm. The eight-minute constraint was the only mechanism that could have produced it.
Frameworks
Where Crazy 8s shows up.
A fast divergence method for the ideation moments of frameworks where teams need to generate many ideas quickly. It appears at the start of phases where genuine variety matters and settling on the first comfortable direction is the main risk.
Related methods
What to pair with Crazy 8s.
Sources & further reading
The work behind this method.
Sprint
Jake Knapp, John Zeratsky, and Braden Kowitz (2016)
The defining account of Crazy 8s in the Design Sprint context. Knapp codified the method as the core rapid ideation activity for Tuesday in a sprint, and his description of the late-panel insight — that the ideas teams actually select come disproportionately from the final panels rather than the first — is the insight that makes the method worth understanding rather than just running. The book's framing of the method as a "forcing function for the ideas you did not know you had" is the clearest articulation of why the time constraint is the whole point, not an inconvenience to be managed.
Gamestorming
Dave Gray, Sunni Brown, and James Macanufo (2010)
A comprehensive reference for rapid ideation methods, including the broader family of time-boxed sketching and divergence practices that Crazy 8s belongs to. Gray, Brown, and Macanufo's treatment of the generative value of constraints — and of why a well-designed game produces better outcomes than an open brainstorm — provides the theoretical grounding for why Crazy 8s works as a structural intervention rather than merely a fun activity. The book's broader catalogue situates Crazy 8s within a family of related practices.
Creative Confidence
Tom Kelley and David Kelley (2013)
On getting past the fear that blocks fast, generative sketching. The Kelleys' treatment of creative confidence — the belief that you can generate good ideas and make them real — is directly relevant to the "I can't draw" block that stops participants from committing the strange late-panel ideas to paper. Their core argument, that creative capacity is learned and sustained by practice rather than innate, is the foundation for the reassurance that rough sketches are required and that the drawing is beside the point. The book's cases show repeatedly that the constraints that appear to limit creativity are often the ones that produce it.