Chapter III
Building
Six uninterrupted weeks on the bet — a team that owns the whole pitch, from the first working piece to the deployed result.
The betting table said yes. The focused day view is now a bet, and Ada, Don, and Jiro own it for the next six weeks. No one will hand them a task list. No one will pull them onto something else in week three. The bet bought the team an unbroken stretch of time and full responsibility for what happens in it: what to build first, what to trade away, what the screens finally look like, and what counts as done. This chapter is those six weeks from the inside.
What makes this hard
Building inside a bet asks three things that cut against how most teams are used to working. Watch for them as the weeks go by.
No one hands out tasks.
The team gets a pitch, whole — not a ticket queue. The real task list can't be written up front; it's discovered by doing the work, and it belongs to the people doing it.
Done means deployed.
Not code-complete, not “in review,” not waiting on a launch plan. Until the work is live, the bet is still open.
Scope bends. Time doesn't.
The deadline is fixed. When the work doesn't fit, the team trims what's optional instead of asking for more time.
The clock
Six weeks, uninterrupted
Six weeks is a deliberate size. Long enough to build something meaningful start to finish; short enough that the deadline is felt from day one, not discovered in the final stretch. And the time is uninterrupted by rule: no drive-by requests, no “quick” side asks, no borrowing a builder for a week. Whatever comes up mid-cycle waits for the next betting table, and almost nothing is too urgent to wait.
Fig. 1 · The cycle
W1
W2
W3
W4
W5
W6
Ship
All six weeks belong to the bet, and the bet gets nothing past week six. There are no extensions.
The first move
Get one piece done
The temptation in week one is to lay foundations: set up the data model, rough in all the screens, wire everything loosely and promise to tighten it later. Resist it. The team's first goal is to pick one small but meaningful piece of the pitch and finish it — front to back, clickable, real — within the first week or so. For the day view, that's a short ranked list rendering real events with the crudest possible ordering. Ugly, narrow, and complete.
A finished slice does two things no amount of parallel groundwork can. It forces the layers to actually meet, which is where the surprises live. And it converts imagined work into discovered work — the real tasks, the ones nobody could have listed at the betting table, start surfacing on day two instead of week four.
The map
Carve the work into scopes
As the first piece comes together, the true structure of the pitch shows itself — and the team carves the work along it into scopes. A scope is a slice of the pitch that can be finished on its own: one part of the problem, with its interface work and its plumbing together in the same bucket. Not “frontend” and “backend.” Not “Ada's tasks” and “Don's tasks.” Parts of the thing itself.
Scopes get names, and the names become the language of the pitch. “The signals are done except ties” tells the whole team something true.
Fig. 2 · The day view, in three scopes
The list
The ranked day view itself — the screen the whole bet is for.
- Ranked list layout
- Empty & one-event days
- Focus entry point
The signals
The rules that decide what makes the day's short list.
- Soonest-first ordering
- Has-guests signal
- Marked-important flag
The way back
Leaving Focus — the full calendar is always one motion away.
- Return affordance
- Remember last view
- Keyboard escape
Each scope cuts through every layer of the stack, so each one can be finished — truly finished — on its own. The task lists under them were discovered by building, not written at the start.
Showing progress
Where things stand: the hill
Every piece of work is really two kinds of work. First the uphill stretch — figuring out the approach, where the unknowns live and estimates mean nothing. Then, past the crest, the downhill — executing a solved problem, where the remaining work is simply doing. The hill chart plots each scope on that journey, and the team nudges the dots as their confidence changes.
This replaces the status meeting. Anyone — a teammate, a shaper, whoever placed the bet — can glance at the hill and know the state of the pitch without interrupting the team. And the chart says the one thing a task list can't: not how much has been done, but how much is still unknown.
Fig. 3 · The hill, mid-cycle
The way back is nearly home and the list is over the top — but the signals have sat on the left slope for a while now. A dot stuck uphill in week three is the most important signal on this chart: something unsolved is still hiding in there.
The trade
Cut scope, on purpose
Somewhere past the middle of the cycle, there's more work left than time left. That's normal. The team sorts what's left into must-haves — the bet isn't honored without them — and nice-to-haves, which get cut when time runs short. Tie-breaking in the ranking is a must-have. Remembering which view you left from can go.
The team makes these cuts itself. Mid-build, it knows more about the pitch than anyone ever has, and deciding what to drop is part of the work.
Week six ends with the day view deployed — smaller in places than the fat-marker sketch, but whole and live. The bet is closed, and cool-down begins.
Terms from this chapter
One piece done
The first milestone: a small but meaningful slice finished end to end — front, back, and the seam between.
Discovered work
The real task list, surfaced by building rather than written up front. Any list written in advance is mostly guesses.
Scope
A slice of the pitch that can be finished on its own — cut along the structure of the work, not by layer or by person.
The hill
Uphill is figuring it out; downhill is making it happen. A scope's position says how much is still unknown, not how much is done.
Must-have / nice-to-have
The sorting that lets scope bend. Must-haves define an honored bet; nice-to-haves are pre-approved cuts.
Done means deployed
The bet closes when the work is live — not merged, not reviewed, not waiting. Shipping is inside the six weeks.