How should a startup make decisions under uncertainty?

Discovery does not eliminate uncertainty; at a startup that is impossible. It manages it: by telling apart the kinds of uncertainty, matching each to the validation that can reduce it, preferring decisions that can be undone, and keeping the strategy flexible enough to absorb what the evidence says.

What are the different types of uncertainty in product?

Discovery does not eliminate uncertainty; at a startup that is impossible. It manages it: by telling apart the kinds of uncertainty, matching each to the validation that can reduce it, preferring decisions that can be undone, and keeping the strategy flexible enough to absorb what the evidence says.

What is a reversible decision and why does it matter?

Discovery does not eliminate uncertainty; at a startup that is impossible. It manages it: by telling apart the kinds of uncertainty, matching each to the validation that can reduce it, preferring decisions that can be undone, and keeping the strategy flexible enough to absorb what the evidence says.

Managing uncertainty · Product Discovery

§The third habit is a relationship with not knowing. Founders are told to be confident and are surrounded by things they cannot know. The resolution is not confidence and it is not paralysis. It is a set of moves for acting well inside uncertainty.

§ 14.1Name the kind#

§Uncertainty is not one thing, and the first move is to say which kind you are facing, because each kind is reduced by a different instrument.

§Marty Cagan's four risks are the most useful taxonomy for . Value risk: will anyone want this? Reduced by problem-space work and by putting something in a stranger's hands. Usability risk: can they use it? Reduced by watching them try. Feasibility risk: can we build it? Reduced by a technical spike, a few days of an engineer trying the hard part. Viability risk: does it work as a business? Reduced by pricing conversations, and, eventually, real money changing hands.

§To those four a startup adds a fifth: timing risk. Is the world ready? Reduced least well by anything, which is why it deserves to be named honestly rather than assumed away.

§ 14.2Prefer the door you can walk back through#

§Jeff Bezos's 2015 shareholder letter gave the distinction its name: some decisions are one-way doors, consequential and irreversible, and deserve slow, careful deliberation; most are two-way doors, and should be made quickly by individuals, because if they turn out wrong you walk back through. The failure he described was applying one-way-door to two-way-door decisions, which makes organizations slow and timid.

§At a startup the distinction is survival. Almost every product decision is a two-way door: a feature can be removed, a price can change, a segment can be dropped. The few one-way doors, the core architecture, the data you promised not to collect, the market you told investors you were in, deserve the evidence. Everything else deserves speed and a note about what would make you reverse it.

§ 14.3Decide on sufficient, not complete#

§Rita McGrath and Ian MacMillan's discovery-driven planning, written thirty years ago for corporate ventures, made a point startups still need: in a new venture the plan is a set of assumptions, and the job is to identify which assumption, if wrong, kills the plan, and test that one first. Everything else can stay uncertain for now.

§This is the practical meaning of "informed but not complete." Complete information about a new product does not exist before the product does. Sufficient information is enough to make the next reversible decision and to know what would change it. The Double Diamond's closes are made on sufficiency. So is everything in Product Strategy.

§ 14.4Keep the strategy flexible#

§The last move is at the level above decisions. A that cannot absorb new evidence is a plan, and plans are what uncertainty destroys. Product Foundation's "stubborn on the vision, flexible on the details" is the posture: hold the destination, treat the route as a hypothesis, and expect the evidence to change it.

§ 14.5What this asks of the owner#

§Name the kind of uncertainty. Pick the test that shrinks it. Ask whether the door swings both ways. Decide on enough, write down what would change your mind, and go and get the next piece of . That is the whole of managing uncertainty, and it is most of the job.

14. Managing uncertainty