What is product strategy at an early-stage startup?

Strategy is choice, not planning. A plan lists what you will do; a strategy says what you will not, and the exclusions are what make it one. This whitepaper takes the written position from the sprint and turns it into a ladder from mission to first unit of work, where each level is the input to the next, so that the first month of building is a consequence of the mission rather than the result of a vote.

What is the difference between a strategy and a plan?

Strategy is choice, not planning. A plan lists what you will do; a strategy says what you will not, and the exclusions are what make it one. This whitepaper takes the written position from the sprint and turns it into a ladder from mission to first unit of work, where each level is the input to the next, so that the first month of building is a consequence of the mission rather than the result of a vote.

How does a startup go from customer evidence to a roadmap?

Strategy is choice, not planning. A plan lists what you will do; a strategy says what you will not, and the exclusions are what make it one. This whitepaper takes the written position from the sprint and turns it into a ladder from mission to first unit of work, where each level is the input to the next, so that the first month of building is a consequence of the mission rather than the result of a vote.

Introduction · Product Strategy

§The third whitepaper ended with a written position: a , a two-act memo, a room that had let twelve strangers overrule its own ranking. This one begins there and asks the question the position leaves open. Now what?

§Most teams answer with a plan. They list the things the says are good, put them in an order that feels right, and call the list a roadmap. The list is honest and the order is a vote, and six months later nobody can say why the first month was spent on what it was spent on. Evidence without order is a list of good things. Strategy is the order, and the order is made of exclusions.

§ 0.1Choice, not planning#

§Richard Rumelt's diagnosis is still the sharpest: most of what is called is a list of goals with adjectives, and bad strategy is recognizable because it excludes nothing. Michael Porter said it a decade earlier in different words: the essence of strategy is choosing what not to do. Both were writing about large companies. At one to five people the point is more urgent, not less, because a large company can afford to pursue three directions badly and a startup cannot afford to pursue one.

§A plan says: we will build recognition, then loyalty, then payments, then reservations. A strategy says: we will let the customer build her own profile before we ask anything of the operation, and we will not launch payments before check-in works. The first is a sequence. The second is a bet, with a reason, and a thing it forbids.

§ 0.2Why the vote loses#

§The 's most uncomfortable finding was that the team's own ranking lost. Payments was pain number one in the room, and the written position proposed leaving it out of version one because six of six strangers had failed the step it depended on.

§That loss is not a failure of the team. It is the job. The purpose of gathering evidence is to have something that can outvote you, and a strategy process that lets the room's original ranking survive contact with the evidence has not processed the evidence at all. This whitepaper is, in part, a set of mechanisms for making sure the loss happens when it should: an evidence framework that scores honestly, a ladder that cites its sources, a strawman that arrives already filled from the field, and rules for a small group that put writing before speaking and voting before debate.

§ 0.3The ladder#

§The center of this whitepaper is a of six questions, borrowed openly from the framing Lenny Rachitsky and others have used: what are we trying to achieve, what does the world look like when we have, what is the plan to win, how do we measure progress, what do we build, what is the first unit of work. Mission, vision, strategy, goals, roadmap, task. Each answer is the input to the one below. If one stays vague, everything under it becomes opinion.

§The ladder is not a form to fill in. It is a discipline for making the first month of work traceable to the reason the company exists, so that when a stakeholder asks why the team is building a profile screen instead of a payment flow, the answer is a chain of reasons rather than a shrug.

§ 0.4How to read this whitepaper#

§Every chapter takes one question and shows a weak answer next to a strong one. The weak answers are not strawmen; they are the answers most teams actually give, and the point of putting them beside strong ones is that the difference is visible when they are side by side and invisible when they are alone. A running example threads through Part II: the ladder that came out of the 2026 field sprint for a multi-brand hospitality group, filled row by row from its evidence.

§Part I sorts the evidence and makes the iterate-or-pivot decision. Part II climbs the ladder. Part III is about how a group of five decides without a facilitator's department behind it. Part IV is how it fails, and the cases. The conclusion is what AI changes, which in strategy is less than in the sprint and more than teams expect.

§How the evidence was gathered is the third whitepaper. Whether the thing closes as a business is the fifth. What experiments move the goals once the product exists is the sixth. This one is the choosing.