Build a Decision Machine
The Practice
Make a f*cking decision. Building a decision machine means treating decision-making as an organizational capability to be engineered: deciding rather than drifting, orienting to reality before acting (orient or die), pushing authority to whoever has the training, awareness, and time (choosing the chooser), deciding with percent of the information you wish you had when decisions are reversible (70 percent is close enough), and imagining failure before execution (90 percent? it's time for a premortem).
How It Works
The practice works by matching decision process to decision type. Reversible decisions move fast; only consequential, irreversible ones get heavy deliberation. Before major decisions, you check your orientation: What do I think is happening? What is that based on? Who has better situational awareness than I do? Authority flows to the person closest to the problem who can understand it, see it, and act within the window. And after plans are formed but before execution, the team imagines the plan has failed and works backward to prevent what's preventable.
Why It Matters
John Boyd's insight was that orientation is everything: in his actual sketch it is the largest component and occupies the central position, and if your read of reality is wrong, everything downstream fails. Meanwhile, indecision is itself a decision, usually the worst one. When you don't decide, you let circumstances decide for you. And timing is unforgiving: doing the right thing at the wrong time is just as bad as doing the wrong thing at any time. Most decisions should probably be made with somewhere around percent of the information you wish you had. What It Looks Like Done Well In organizations with a decision machine, teams make decisions within their windows. Approval chains are lean; sign-offs that never produce a 'no' have been eliminated. People closest to problems act without waiting for permission, and their reasonable judgment is backed even when outcomes are imperfect. Mental models get updated when the environment changes. And plans proceed with known risks accepted rather than unknown risks ignored.