13. The written position

§The question this chapter answers: what do we now believe, what do we still not know, and what do we propose to do about it?

§Everything before this point produced pages. This chapter produces a position: a document short enough to be read in one sitting and specific enough to be wrong. It is the last thing the sprint makes and the first thing the strategy work in the fourth whitepaper reads.

§ 13.1The tension map#

§Take every assumption you have written down since the research plan in chapter 1. The ones about the problem, the ones from the solution canvases, the ones the brief for the machine encoded as rules. Sort each into one of three columns.

§Validated. More than one stranger, without prompting, behaved or spoke in a way that supports it. Next to it: the quote, and the code of the person who said it.

§Open. The evidence is mixed, or thin, or comes from the wrong setting. Next to it: the quote that made it open, an owner, and a date by which it will be closed, and how.

§Discarded. Strangers behaved or spoke in a way that contradicts it, more than once. Next to it: the quote that killed it, and the code. Discarded does not mean forgotten. It means not in version one, and the map says why.

§ 13.2Open is a state. Undecided is not.#

§The distinction the whole document turns on. An open assumption has a name next to it and a date, and a plan for how it will be closed. An undecided one has a conversation next to it that will resume, louder, the next time the team meets.

§Sprints end with open items. That is expected and healthy. In the field sprint, the closing session ended with six open decisions written on the board, each with an owner. None was resolved that day, and none needed to be, because each was a task rather than a mood.

§ 13.3Reading the map: iterate or pivot#

§Eric Ries framed the decision at the end of a learning loop as pivot or persevere, and the tension map is what makes it a decision rather than a feeling.

§Read the validated column. If the problem is confirmed, the segment is real, the current solutions fail the way you said, and the solution's core assumption held, then the solution needs refinement and you iterate: keep the problem and the core, change what the open and discarded columns tell you to change.

§Read the discarded column. If the problem itself is there, or the segment, or the willingness to pay, or the assumption on which the solution's value rested, then you pivot: keep what validated and go back to chapter 7 or chapter 8 with a different statement or a different canvas. There is no shame in this. The shame is in the team that read the map, saw the discarded column, and built anyway.

§Bezos's distinction between one-way and two-way doors helps with the open column. An open assumption that would be cheap to reverse can be carried into the build and closed by shipping. One that would be expensive to reverse has to be closed before the build, and the map should say which is which.

§ 13.4The memo that proposes, not decides#

§Short. A few pages at most, in prose rather than slides, in the tradition of the narrative memo that Colin Bryar and Bill Carr describe. It proposes; the strategy work decides. Its parts:

§The proposal in one line. What to build first and what to leave out, in a sentence a stranger could disagree with.

§Act one. What is in, what is out, and why it goes first. The strongest reason for a first act is that it depends on nobody else and answers the most important open question. In the field sprint, act one was a customer profile and a visible tier, nothing else, because it could ship in a month and depended on no one.

§Act two. Why it goes second and how it starts. In the sprint, act two was letting the operation read that profile before serving, in one location, read-only, because the operation had asked for exactly that in five of six interviews.

§Deferred and why. What is not in version one and why it is deferred rather than discarded. In the sprint: payments, which the team had entered ranking first and left proposing to leave out, because it depended on a check-in that had failed six of six.

§What must be defined this week. The open items whose date is now, each with a number to hit.

§ 13.5The handoff#

§The written position leaves the sprint and enters strategy. The fourth whitepaper takes a tension map and a two-act memo and turns them into bets, critical questions and a roadmap. The fifth asks whether the thing closes as a business. The sixth owns what you test once it has shipped. None of them can start from a feeling. All of them can start from this.

§The team entered with payments as pain number one. It left proposing to leave payments out of version one. Nobody in the room decided that. Twelve people outside it did, and the map is how the room let them.

§ 13.6What you leave with#

§The tension map, complete: every assumption sorted, every row with its quote and its code, every open row with an owner and a date. The two-act memo, a few pages, proposing and not deciding. And the sprint's own record: the research plan you started with, marked up with what the field did to it. Read that last one twice. It is the most honest account you will ever have of how much you knew on day one.