Four-Step Sketch — The Product Playbook

"When it comes to idea generation, quantity is the most predictable path to quality."

That's Adam Grant, the organizational psychologist, in his book Originals. Read it slowly, because it goes against everything most teams believe. You don't get the great idea by putting smart people in a room and asking for the great idea. You get it by making the pile first, then narrowing hard.

You can't pick the one great idea without the pile. There is no version where you skip the pile and land on the winner anyway.

So this page is the machine that makes the pile. One person, alone, in silence, working through four steps with a timer running. It comes straight out of Jake Knapp's book Sprint, where he calls it the Four-Step Sketch. The credit's his, and it's on my slides too.

Knapp built it for the sketching day of his five-day design sprint, day two of the week. But it works just as well pulled out of the week and run on its own, dropped into a Decision Jam (a fast workshop that takes a group from a problem to a decision), or run on a regular Tuesday when a product debate has been circling for too long. That's how I teach it.

First, let's talk about drawing

Here's where every group seizes up. The whole sequence runs on sketching, and "sketch" is the word that makes grown professionals freeze.

When I teach this, I open with two questions. First: "Who in here can draw?" A few hands. Then: "Who in here sucks at drawing?" Now the hands go up. You can guess which question gets the room.

That fear is the thing in your way, so we kill it in about ten minutes before we touch a single step.

I keep a line from Dave Gray, the visual-thinking author, on my slides. "If your ideas can't be drawn, they can't be done." Then I show the room some drawings. Apollo-era NASA diagrams. The loop a spacecraft flies from Earth orbit out around the moon and back. The re-entry sequence, the capsule floating down under its parachutes. Then the last one. A hand-drawn sketch out of NASA's Langley lab, dated 1961, titled "To the moon with C-1's or bust." Boxes, circles, arrows, scribbled weights in the margins. The C-1 was an early Saturn rocket.

The moon landing started as a sketch that wouldn't win a sixth-grade art contest.

Then we practice, and the entire curriculum takes ten minutes. Basic shapes: line, circle, square, triangle, blob. Glyphs, a drill I borrowed from Dave Gray: dots, arcs, hooks, loops. Combine glyphs into objects. Combine boxes, circles, and sticks into people. Combine people and objects into scenes. A stick figure at a laptop. Four people around a table.

Then I ask again: who can draw now?

Everyone. Boxes, arrows, and stick figures are the entire visual vocabulary you need for the rest of this page. Spend the ten minutes. It's the difference between a room that sketches and a room that argues about why it can't.

The Four-Step Sketch

Everyone works alone, in silence, through four steps. The timings are Knapp's, and I teach them exactly as he wrote them. Set a real timer. The clock is not decoration. A brain that knows it has eight minutes puts more into those eight minutes than a brain that thinks it has all afternoon.

1. Notes, 20 minutes. No inventing yet. You're just collecting what you already know. The goal. The question you're trying to answer, framed as a How Might We. A How Might We takes the problem you came in to solve and turns it into an open question that invites solutions instead of describing the pain. "People abandon the cart at checkout" becomes "How might we make checkout feel effortless?" That open question is what every later step aims at. Then keep collecting. Customer quotes. Competitor moves. The products you wish you'd built. Subject-matter-expert input. Your own thoughts on everything you understand about the space.

You've already got a lot of this. This step is you getting it onto the page in one place. It's cheap, and it makes every step after it better.

2. Ideas, 20 minutes. Now turn those notes into rough ideas. Doodles, headlines, diagrams, stick figures acting out the steps. Messy and incomplete is the point, not a problem. They don't all have to be full solutions, either. "I wonder if..." counts. A partnership with something that already exists counts. Write down everything that comes to mind as a way to answer your How Might We. Anything you put on the page is good.

3. Crazy 8s, 8 minutes. Fold a sheet of paper into eight panels first, before the clock starts. Then start the timer. Take your strongest idea from the last step and sketch eight variations of it, one per panel, sixty seconds each. The 8 minutes is the drawing only. Go fast, be messy.

Crazy 8s exists to force you past your first reasonable solution. Your first idea is almost never your best one, it's just the first one your brain handed you. Eight panels in eight minutes drags you into the alternatives you'd otherwise never consider. Sometimes panel seven is a revelation. Sometimes the first idea really was the best, and now you know that instead of assuming it. Either way it's a warm-up for the main event, and it's the most fun eight minutes of the whole sequence.

4. Solution Sketch, 30 minutes or more. This is the main event. Put the Crazy 8s sheet aside and switch to sticky notes. Take your best idea, usually one panel off that sheet, and turn it into a three-panel storyboard on sticky notes. Three notes in a row, left to right, one panel each, showing what your customer sees first, next, and last as they move through your solution. Five rules, and they all matter:

Make it self-explanatory. Someone who has never spoken to you should understand it on sight. Unlike the first two steps, this one has to stand completely on its own, because in the next exercise you will not be standing next to it to explain it. The sketch has to argue for itself.
Keep it anonymous. Don't put your name on it. A name introduces bias. It also makes the room judge the idea instead of the person, which is the entire reason this works.
Ugly is ok. The sketch doesn't have to be fancy. People just need to understand it. Speed and clarity beat polish here, every time.
Words matter. Don't draw squigglies for "text goes here." Write the actual text. The words are half the idea.
Give it a catchy title. Your name isn't on the sketch, so the title is how the room tracks it through the next round. A good title names the bet in a few words, so anyone glancing at the wall knows what this sketch is arguing for.

That's the recipe. Notes, Ideas, Crazy 8s, Solution Sketch. Four steps, about 80 minutes, one person at a time.

What good looks like

You know it worked when you have a stack of solution sketches taped to the wall, and the stack has three things going for it.

Concrete. Every sketch shows what the customer actually sees and does, in real words, panel by panel. Not a bullet point that says "improve onboarding." A drawing of the onboarding screen with the actual button on it.
Comparable. Because everyone ran the same four steps to the same format, you can stand back and read the sketches against each other instead of comparing a polished deck to a scribble on a napkin. Same format, same fighting weight.
Anonymous. No name on any of them. So the next decision gets made on what's on the paper, not on who has the biggest title in the room.

That stack is the whole point of the exercise. It's a pile of real options, drawn by everyone, ready to be decided on.

When to reach for it

Reach for the Four-Step Sketch when the work says "now ideate" and you need ideas worth testing, not a transcript of the loudest person in the room.

In Part 04, this is the step where you move from a problem you've confirmed is real to actual design concepts worth testing. It's day two of a design sprint, if you're running the full week (the Design Sprint page in this part walks all five days). It's the generating half of a Decision Jam, the part where the group makes the pile before it picks. And it's any product debate that's been going in circles, where everyone has an opinion and nobody has drawn anything.

It's not the right tool when you already know exactly what you're building and you're just arguing over details. Save it for the moments when the solution is genuinely open and you want the real range of options on the table before you commit to one.

Supplies: paper, markers, sticky notes, a timer, and a wall. That's it.

Now you decide

Four steps in, you've got a wall of concrete, comparable, anonymous sketches. A pile worth choosing from.

What you don't do next is debate it. The fastest way to waste a great stack of sketches is to open the floor and let the senior voice talk first. There's a separate recipe for narrowing the pile without a debate, the Decide Sequence, and it's the next page in this part. Run that, and you'll walk out with one to three ideas the room actually chose, ready to take into small, cheap experiments and test for real.

The data picks the winner from there. Your job here was just to make sure the winner had a real pile to come from.