There's a pattern that shows up in almost every organization, and it looks like this: three weeks of debate about which project management tool to use, decided by committee. And a 30-minute all-hands meeting announcing a major architecture change.
The first is trivially reversible. The second is not. The organizational energy is almost exactly backwards.
This isn't universal — some teams have it figured out. But the inversion is common enough that it's worth naming: most organizations treat all decisions as equally weighty, when the actual distribution of reversibility is wide and consequential.
The Inversion Problem
Most organizations have the default posture inverted. Reversible decisions get slowed down by consensus requirements, alignment meetings, and process. Irreversible decisions get rushed through because they feel routine, or because the urgency of the moment overrides the structural consequence.
The result: the organization moves slowly on things that don't matter much and quickly on things that matter enormously.
Jeff Bezos described this distinction well in his 2003 letter to shareholders — distinguishing Type 1 (mostly irreversible) and Type 2 (mostly reversible) decisions. His point: most decisions are Type 2, and treating them as Type 1 is where organizations slow down unnecessarily. The corollary, which he didn't belabor but is equally important: treating Type 1 decisions as Type 2 is where organizations take irreversible damage.
How to Handle the Irreversibles
For decisions that genuinely can't be easily reversed, the goal is to create optionality before committing. This means:
Defer commitment until the point of necessity. Don't make the irreversible decision before you have to. Buy time by identifying what information you'd need to make a better call, and what's preventing you from getting it.
Stage the commitment. Structure the decision so the irreversible step is last, not first. A contract with a 90-day exit is less irreversible than a multi-year lock-in. A pilot before a full rollout is less irreversible than a full commitment. The sequence matters — don't make the hard-to-undo move before you have signal on whether the easy-to-reverse move was worth it.
Buy the option if you can. In high-uncertainty environments, paying for optionality is rational. A right-of-first-refusal, a prototype, a proof-of-concept, a beta customer agreement — these cost something upfront and create the ability to walk away later if the signal is bad. The cost of the option is the price of not being locked in.
Get the critical inputs. For decisions that genuinely can't be reversed, the information gathering is different. You're not trying to eliminate uncertainty — you're trying to identify the single biggest risk, the scenario that would make you regret this, and whether you can structure the decision to reduce that scenario's likelihood.
