The decisions that define companies are almost never made with complete information. If the information were complete, the decision would not be hard — it would be obvious, and everyone would make it, which means it would no longer be a source of competitive advantage. The decisions that matter are the ones where reasonable, intelligent people looking at the same situation reach different conclusions. That gap — between what the data says and what the founder decides — is where company outcomes are determined.
The regret minimization framework (Bezos)
Jeff Bezos describes leaving his hedge fund job to start Amazon by imagining himself at 80, looking back on his life. The question was not "what is the expected value of this decision?" It was "which path will I regret not taking more?" He concluded he would not regret failing at Amazon — he would regret not having tried.
The framework is useful for irreversible decisions where the cost of inaction compounds over time. It reframes risk from "probability of bad outcome" to "cost of permanent regret." For founders facing major pivots, market timing decisions, or career-defining bets, it provides clarity that expected value calculations cannot — because the inputs to expected value are unknowable, but the inputs to regret minimization are internal and accessible.
Reversibility sorting (two-way doors)
Bezos also articulated a distinction that most founder decision frameworks underweight: the difference between one-way doors (irreversible decisions that require careful analysis) and two-way doors (reversible decisions that should be made quickly by the smallest appropriate team). Most organizations treat every decision like a one-way door, which slows them down on the decisions that should be made quickly and focuses insufficient attention on the ones that actually require it.
The practical application is a sorting habit: before any major decision, explicitly categorize it as reversible or irreversible. If reversible — decide fast, with less information, at a lower level of the organization. If irreversible — slow down, gather more data, get the right people in the room. Most founders get this backwards, spending their attention on the reversible decisions and underweighting the irreversible ones.
Pre-mortems (Klein)
Gary Klein's pre-mortem technique asks decision-makers to imagine that the decision has already been made and the project has failed. The task is to explain why it failed. Research consistently shows that pre-mortems surface risks that standard risk analysis misses — because the imagined failure state activates different cognitive processes than the forward-looking analysis mode. For founders making high-stakes bets, running a pre-mortem with the founding team before committing produces a different quality of risk identification than asking "what could go wrong?"
The two-track model (Kahneman adapted)
Daniel Kahneman's System 1 / System 2 framework has an underused application in founder decisions: the insight is that the feeling of confidence is not correlated with the accuracy of the judgment. Founders who have operated in a domain for years develop strong intuitions — System 1 responses — that feel like certainty. Sometimes they are. But the domains where intuitions are reliable (chess, weather, firefighting) share a feature that most startup markets do not: fast, clear feedback that allows the intuition to calibrate over time.
In a new market, or in a market undergoing rapid change, the founder's intuition may be processing patterns that no longer hold. The discipline of slowing down to check whether the intuition is operating in a domain with the feedback characteristics that make intuition reliable is one of the most valuable habits a founder can develop. The question is not "what does my gut say?" It is "is my gut calibrated to give reliable answers in this specific type of situation?"
The commitment device
The best founders recognize that the hardest part of a difficult decision is often not making it — it is preventing themselves from relitigating it. Once a direction is chosen, the organizational cost of reversal compounds quickly. The most effective founders build commitment devices into their decision process: they announce the decision to stakeholders whose buy-in matters, they allocate resources before doubt can accumulate, and they set a specific review date rather than leaving the decision open-ended.
This is not inflexibility. It is the recognition that second-guessing a decision before the data has had time to come in is not analysis — it is anxiety. The commitment device creates the space for the decision to generate real feedback, which is the only information that should change it.
What to do when the information will never be complete
The founders who make the best decisions under uncertainty are not the ones with the best information. They are the ones who are clearest about what information they actually need, fastest at acquiring it, and most disciplined about the difference between the information they have and the information they wish they had. That distinction — between what is known and what is assumed — is the single most important skill in founder decision-making. Building it is a practice, not a trait.