How do you write a product vision for a startup?

The vision is the world at a specific future point, described from the outside: what a customer sees, what someone in the operation sees, what the founder sees. Two or three sentences, concrete enough that a stranger could say what would count as failure. It is held stubbornly, because it is the level everything below it is allowed to change in service of.

What does 'stubborn on the vision, flexible on the details' mean?

The vision is the world at a specific future point, described from the outside: what a customer sees, what someone in the operation sees, what the founder sees. Two or three sentences, concrete enough that a stranger could say what would count as failure. It is held stubbornly, because it is the level everything below it is allowed to change in service of.

How specific should a product vision be?

The vision is the world at a specific future point, described from the outside: what a customer sees, what someone in the operation sees, what the founder sees. Two or three sentences, concrete enough that a stranger could say what would count as failure. It is held stubbornly, because it is the level everything below it is allowed to change in service of.

Vision · Product Strategy · Product Thinking

§The question: what does the world look like when we have done it?

§The mission is a purpose without a date. The vision puts a date on it, usually two years out, and describes the world at that date from the outside, as it would appear to the people in it. It is the level Bezos meant when he said stubborn: the vision does not move because a feature failed. It moves only when the world itself turns out to be different from what the team believed.

§ 5.1Three points of view#

§Write it from three seats. What does a customer see and do, on an ordinary day, when this is true? What does someone inside the operation see? What does the founder or the group's leadership see when they look at the business? Two or three sentences in total, each concrete enough to be checked.

Weak vision	Strong vision
"To become the most loved dining experience in the region, powered by technology."	"In two years a guest walks into any of our restaurants and is greeted by name, with her preferences already known, having never filled in a form or shown a card. The server saw who was coming before she sat down. The group knows, across all brands, who its regulars are and what they are worth."
"To be the AI-native way to build software."	"In two years a person with an idea and no engineering background describes it in a conversation and has a working, deployed application the same afternoon, and keeps improving it the same way."

§The weak column describes the company's ambition. The strong column describes a Tuesday. Marty Cagan's distinction is the useful one: the mission is why, the vision is what the world is like when the why has been served, and the strong column only works because it can be pictured.

§ 5.2Why stubborn#

§The vision is the level that is allowed to survive a bad month. A prototype fails in the field; the below the vision changes, the roadmap changes, the first task changes, and the vision stays. If the team revises the vision every time a feature fails, the ladder has no fixed point, and every level becomes negotiable at once.

§Stubbornness has a condition. The vision stays until about the world, not about a feature, contradicts it. If the field had shown that guests did not want to be recognized, that they valued anonymity in restaurants and were unsettled by a server who knew their name, the vision itself would have been wrong and the stubbornness would have been a mistake. It showed the opposite: six of six filled in their profiles gladly. The vision held. The features underneath it did not.

§Vince Law's beginner's guide to vision frames the vision as the imagined, idealized world in which your product has solved the problem with the greatest conceivable outcome, and makes one point worth borrowing here: a good vision deliberately suspends constraints. It is written as if feasibility were not a question, precisely so it can pull the roadmap toward something larger than the next sprint. That is why the vision sits above the bets and not among them. The bets are constrained by evidence and runway; the vision is the one place you are allowed to describe the world as it would be if the constraints were gone, and then let the strategy below it negotiate the way there.

§ 5.3The future press release as a check#

§Amazon's working-backwards discipline offers a test that fits a small team: write the announcement of the product as it will be described at the vision date. If you cannot write a paragraph a customer would find interesting, the vision is either too vague to announce or too small to matter. Bryar and Carr's version runs to a page with a FAQ; a startup's can be three sentences and the same three questions. Who is this for? What do they get? Why would they care today?

§ 5.4Where it comes from#

§Unlike the mission, the vision is written with the evidence in view. The field told you what customers valued and what the operation asked for. The vision is the world in which those things are true at scale, and each sentence should be able to point to the row of the tension map that makes it plausible. Recognition without a form: validated by six of six filling in their profiles. The server sees who is coming: validated by five of six operations staff naming the profile card. The group knows its regulars across brands: still open, and the vision says it anyway, because that is what the mission requires.

§ 5.5What comes out#

§Two or three sentences, dated, from three points of view, each traceable to a row of evidence or honestly marked as belief. Next: the plan to get there, which is the first level of the ladder where leaving things out becomes the whole work.

5. Vision