Decisions.

The framework

1. 01

   Identify where loss aversion is making the decision.

   The symptom is a decision that's been open too long. A vendor relationship you haven't terminated. A strategy you've kept running past its evidence. A hire you haven't let go. A product you won't sunset. In each case, the sunk cost — money already spent, time already invested, relationship already built — is generating a psychological loss that dominates the expected-value calculation. Name the loss explicitly: 'I haven't fired this agency because I'd have to admit the last six months of spend were wrong.' Making the psychological mechanism visible is the first lever.

2. 02

   Separate sunk costs from forward costs.

   The sunk cost fallacy is loss aversion applied to money already spent. The $2M already invested in the platform rebuild is gone regardless of what you decide next. The only relevant question is: given where we are today, does continuing create more value than stopping? The forward cost is what matters. The brutal reframe: if you were handed this project fresh today — no history, no investment, just current state and future options — would you start it? If the answer is no, you're holding because of loss aversion, not expected value.

3. 03

   Reframe losses as information, not failure.

   The organisations and individuals with the highest decision quality treat losses as feedback, not verdicts. A pricing experiment that loses money proves what customers won't pay — that's valuable data. A failed hire reveals a hiring criterion you hadn't made explicit. A closed door points toward an open one. The frame shift: a realised loss that produces learning is better than an unrealised loss that produces paralysis. Jeff Bezos's standard for Amazon experiments: the downside is bounded and the learning is not. That's a loss aversion antidote built into the decision architecture.

4. 04

   Pre-commit to exit criteria before you start.

   The cleanest antidote to loss aversion is setting the exit criterion before you enter. Define in advance the conditions under which you'll stop: 'If this channel doesn't produce 50 qualified leads in 90 days, we stop.' 'If churn exceeds 5% in month three, we reassess the model.' Pre-commitment removes the decision from the emotional state that will exist when the loss is being felt. Written exit criteria are enforced by System 2 before System 1 gets involved. They're also the fastest way to learn — because you've defined what failure looks like before you're invested in the outcome.