Start a new project with nothing on the table and you’ll end up with ten different ideas living in ten different heads. Everyone nodding. Nobody picturing the same thing.
So I don’t start with a conversation. I start with something to see. A rough sketch, a few boxes and arrows, a bad drawing of how the thing might work. It doesn’t have to be right. It usually isn’t. But the second there’s a picture on the table, all ten of those private versions have to come out and reconcile with a shared one. That alone fixes half the meeting.
Why a picture beats a conversation
Talking toward alignment is slow because words are easy to agree with and hard to pin down. Two people can both nod at “streamlined onboarding” while imagining completely different products. The agreement is fake, and you don’t find out until you’ve built the wrong thing.
A diagram kills the fake agreement. You can’t vaguely nod at a picture. It’s specific, it’s in front of everyone, and it forces the real question: is this it or not? Suddenly the quiet objections surface. The thing someone was uneasy about but couldn’t name is now a box they can point at and say “that part’s wrong.” That isn’t conflict. That’s the disagreement finally showing up early, while it’s still cheap to fix.
Make it ugly on purpose
The rougher the diagram, the better it works. That feels backwards, so let me explain.
A polished, high-fidelity mockup reads as decided. It quietly tells the room “this is finished, don’t touch it,” and people hold back. A rough sketch says the opposite. It looks unfinished because it is, and that gives everyone permission to grab a marker and change it. Ugly invites. Polished intimidates. When I actually want input, I bring the ugly version on purpose.
The first draft’s job is to die
Years of design taught me not to love my first drafts. The second, third, and tenth are almost always better, and the first one mostly exists to get killed.
That’s not a flaw in the process. It’s the whole function. The first diagram isn’t a proposal, it’s a target. You put it up so people can shoot at it, and every hole they find is something you didn’t have to discover three weeks later in code. Throwing it away isn’t the failure. Getting attached to it is.
Draw first, decide later
You don’t need to be good at drawing for this to work. You need to be willing to be wrong in public, fast.
Sketch the thing before the meeting so people have something to react to instead of a blank page to fear. Keep it loose. Invite the marks and the objections. Then throw it out and draw the better one you couldn’t have reached without the first.
The image was never the deliverable. It was the thing that got ten heads into the same room.