Playbook
Deciding Under Uncertainty
Four frameworks for decisions the spreadsheet cannot settle: work out what kind of problem you are actually facing, rehearse the futures it could land in, price the choice branch by branch, then let the team fail the plan on paper while it can still be changed.
The worst decisions of my twenty years in business shared a pattern: I brought the wrong kind of thinking to the problem in front of me. I analysed things that needed a quick experiment, and I went with my gut on things that deserved a spreadsheet. Uncertainty punishes that mismatch harder than anything else, which is why this sequence starts a step earlier than most decision advice, with the question of what kind of problem you are actually holding. Cynefin sorts that out first, so the method fits the problem rather than your habits. Scenario planning then builds a small set of genuinely different futures and lets you rehearse inside them, which loosens the grip of the one official future everyone privately doubts. Decision trees put prices on the branches, so the choice and its assumptions sit on the table where they can be challenged one number at a time. And the pre-mortem comes last, after all that rigour, precisely because all that rigour breeds confidence, and confidence before commitment is the most dangerous asset in the room. None of this removes the uncertainty. It gives you a route through it that you can defend afterwards, including to yourself.
Decision-making & prioritisation
Cynefin Framework
For working out what kind of problem you are facing, so you decide the way the problem needs rather than the way you are comfortable with. Leave it alone when you want a label to file and forget; situations move. Instead Nothing to file; come back to it when the situation moves and you need to know what kind it has become.
Decide how to decide before deciding what to do. Cynefin sorts situations by the relationship between cause and effect: some problems reward analysis, some reward expertise, some only reveal themselves through cheap experiments, and a genuine crisis needs stabilising action before assessment. Most of us default to the method we are comfortable with rather than the one the problem needs. When your engineers call something Complicated and your frontline calls it Complex, treat the disagreement as data and run the cheap experiment before commissioning the expensive analysis.
Watch forPeople place problems by professional preference: engineers find Complicated everywhere, agilists find Complex, and executives under pressure find Chaotic.
Strategy & competition
Scenario Planning
For building a few different, plausible futures and rehearsing what you would do in each, so the future you quietly assume stops being the only one. Leave it alone when the uncertainty is a risk you can put numbers on; forecasting and sensitivity analysis are sharper. Instead Decision Trees, when the odds can be put on the branches; scenarios are for the uncertainty that cannot.
For the strands that are genuinely uncertain, build two to four futures that differ in structure rather than in degree, and rehearse your strategy inside each one. The question to ask of every scenario is 'what would we wish we had done?', never 'which one do we believe?'. You are sorting your moves into three groups: the robust ones that work everywhere, the options worth pricing, and the bets. The exercise has worked when the room describes the present differently afterwards.
Watch forA good year, a bad year and the budget is sensitivity analysis wearing a costume, and it leaves the official future untouched.
Decision-making & prioritisation
Decision Trees
For laying out a decision with branches for what you control and what you do not, with rough odds and payoffs, so you can see which first move is worth most. Leave it alone when nobody can put even a rough number on the odds; it will manufacture precision. Instead Scenario Planning, which works without odds you cannot defend.
With the futures rehearsed, price the actual choice. A decision tree lays out the sequence of choices and chance events, puts probabilities and payoffs on the branches, and folds the values back to tell you which first move is worth most. Its deeper service is honesty: every assumption sits on a branch where somebody can challenge it, one number at a time. Give the sensitivity analysis equal billing with the answer, because if the recommendation flips inside the honest range of an estimate, the tree is telling you to go and improve that estimate.
Watch forA positive expected value that risks the survival of the firm is still a bad bet, because you will experience one branch, never the average.
Decision-making & prioritisation
Pre-mortem
For before you commit, imagining the plan has already failed and having everyone write down why, in silence; it costs an hour. Leave it alone when the decision is already signed; it produces regret, not improvement. Instead A Risk Matrix on the signed decision, to manage what you can no longer avoid.
After the domains, the scenarios and the tree, the team will feel rigorous, and feeling rigorous is exactly when overconfidence does its best work. So, before commitment, declare the decision a year old and failed outright, and have everyone write down why, independently and in silence. The silence matters; it flattens the hierarchy for twenty minutes and surfaces what junior people knew and would never have volunteered against a senior sponsor. If the confidence in the room drops afterwards, the exercise has worked. That deflation is the product.
Watch forSoftening the frame to 'imagine things might go wrong' collapses the exercise into an ordinary risk discussion and loses the effect entirely.
In the end
By the end you should hold an honest reading of what kind of problem you faced, a set of named futures with indicators that tell you which one is arriving, a priced policy for the choice itself, and a ranked list of failure modes with tripwires attached. You will still be deciding without certainty, because that is what deciding is. What this sequence buys is a decision whose reasoning survives contact with hindsight, and early warning when the world starts down a branch you did not choose.