Cost of delay

Use when

• A decision keeps being deferred and nobody has said what that costs.
• You are choosing between deciding now on partial information and waiting for more.
• A group treats more analysis as automatically responsible.

Do not use when

• The value is genuinely not quantifiable, where the number invented will dominate the argument falsely.
• Delay actively improves the decision — some options only appear with time, and the rate misses that.

Inputs required

• The value at stake
• A unit of time
• An estimate of the cost of being wrong

Procedure

1. 01

   Estimate the value

   What the outcome is worth if the decision goes ahead. Order of magnitude is enough; a Fermi estimate is appropriate here.

2. 02

   Convert to a rate

   How much of that value is lost per week of delay? Sometimes it is a share of the total; sometimes it is a fixed carrying cost; sometimes the window closes entirely on a date, which makes the rate a cliff rather than a slope.

3. 03

   Estimate the cost of being wrong

   What a bad decision made now would cost, multiplied by the probability of making one. This is the other side of the comparison.

4. 04

   Compare directly

   Weekly delay cost against expected cost of error. If delay costs more, decide now with what you have. If not, the wait is buying something.

5. 05

   Re-run it as the week passes

   The cost of being wrong falls as information arrives; the delay cost usually does not. The comparison flips at a specific point, and that point is the deadline.