The day · Page 45
Alone together, then as one
Give this to the room
The build window, resliced:
9:30 to 11:30 · Build alone. One prototype each, same brief, same tools, two screens each. Sit with your teammate, build without them. Do not converge in minute five; two genuinely different starting points are the whole point of the morning.
11:30 to 12:00 · Synthetic round. Hands off. Every solo build goes through a round of AI user testing. Collect two or three findings per build, written down.
12:00 to 12:30 · The jury half hour. Sixty seconds each: what I built, what the AI users tripped over. Then pick one of three doors for the afternoon: adopt one build, merge the best parts, or start a third idea you both now believe in. The decider, named before the conversation starts, is whoever sits closest to the user in the brief.
12:30 to 15:30 · One build per team. The classic flow from here: lunch lands, the clock keeps running, 14:30 stop adding, 15:15 save links, 15:30 hands off.
For you only
Three guardrails keep the variant honest.
Solo means solo. A pair that agrees a direction at 9:35 has quietly killed the variant and gone back to the classic day with extra steps. When you catch one mid-merge, split them kindly: "build the version you argued for; noon decides."
The merge trap. In Nielsen and Faber's study the merged design won, but they merged like surgeons: the winning skeleton, plus one or two strong organs from the other designs. Stapling two whole prototypes together gives you two products in a trench coat, and you already know how that ends at the pitches. One skeleton, borrowed organs, nothing else.
The my-baby problem. Solo builders arrive at noon attached to their own screens. This is exactly what the synthetic findings are for: "the AI users got lost on my pricing page" is easy to say, and "your pricing page is confusing" never gets said at all. Let the transcripts do the critique. Egos survive, evidence decides.
Four teams used to give you four bets before lunch. This gives you eight.