Costs
Part of Decision skills fields compared: statement, constraints, criteria and reversibility
Decision skills mistakes, and the fixes that hold
Nine decision skills mistakes, from choosing criteria after seeing the options to judging a decision by its outcome, each with the tell that gives it away.
Each of these is a way of reaching an answer that feels like reasoning and is not. They are listed with the tell that gives each one away, because the errors are hard to see from inside and easy to see once you know the shape.
What to take away
- Most of these are errors of sequence rather than of judgment. The criteria get written after the options are known, the date gets set after the answer is wanted, the record gets written after the outcome is in.
- Two of them are so common they deserve naming in advancedefending what has already been spent, and treating the absence of objection as agreement.
- Write down what would change your mind at the moment you decide. Almost every error below is caught by that one habit.
1. Choosing the criteria after seeing the options
The options arrive first, one of them is attractive, and the criteria that get written happen to favor it. Nobody notices, because each criterion is defensible on its own.
The tell: the criteria list has an unusual item on it that only one option satisfies.
2. Defending what has already been spent
Money, time and reputation already committed are gone whichever way you now choose, and they feel like reasons. This is the sunk cost error, and knowing its name does surprisingly little to prevent it.
The tell: the argument for continuing describes the past rather than the future.
3. The false binary
Two options, presented as the field, because they are the two somebody prepared. The third option, doing neither and deciding again later, is almost always available and almost never listed.
The tell: nobody can say what happens if both are declined.
4. Treating a forecast as a decision
A projection is produced, everyone agrees it is reasonable, and the meeting ends without anyone choosing anything. The forecast is then referred to as though it were a commitment.
The tell: no one can name the decider.
5. Reading silence as agreement
Nobody objected, so everyone agreed. In fact some people had not understood, some were waiting for someone senior to speak, and some assumed it was already settled.
The tell: the objections arrive after implementation starts. The five states of a message explain why silence carries no information at all.
6. Spending the most process on the most expensive decision
Which sounds right, and is not the rule. Cost and reversibility are different axes. A cheap decision that cannot be undone deserves more care than an expensive one that can.
The tell: the approval threshold is a sum of money, with nothing about reversibility in it.
7. Deciding without a date
The decision is discussed, everyone leaves with a different sense of whether it was made, and the answer emerges weeks later from whoever acted first.
The tell: two people describe the outcome differently a week afterward.
8. Re-deciding
The same decision is reopened whenever someone new arrives or someone dislikes an early consequence. Nothing new has been learned; the discomfort is simply recent.
The tell: the new discussion contains no information that was unavailable the first time. The repair is a stated trigger for revisiting, agreed when the decision is made.
9. Judging the decision by the outcome
A good process can produce a bad result and a careless one can get lucky. Judging by outcome teaches people to avoid uncertainty rather than to reason well, and it makes honest records dangerous to keep. The assessment guide sets out the general form of this: a single result is evidence about one occasion, not about a person.
The tell: the review starts with what happened rather than with what was known at the time.
The nine, sorted
| Error | Where it goes wrong | The one habit that catches it |
|---|---|---|
| Criteria after options | Sequence | Write criteria first, and date them |
| Defending past spend | Frame | Ask: would we start this today? |
| False binary | Options | Always list "neither, and revisit when" |
| Forecast as decision | Roles | Name the decider out loud |
| Silence as agreement | Evidence | Ask for a restatement |
| Process sized by cost | Sizing | Ask about reversibility, not price |
| No date | Timing | Put the date in the statement |
| Re-deciding | Closure | Agree the revisit trigger in advance |
| Judged by outcome | Review | Review the record, not the result |
The right-hand column is short on purpose. None of these needs a framework; each
Nine Errors, Nine Catching Habits
- Criteria after optionswrite criteria first, date them
- Defending past spendwould we start this today?
- False binarylist neither, and revisit when
- Forecast as decisionname the decider out loud
- Silence as agreementask for a restatement
- Process sized by costask about reversibility, not price
- No dateput the date in the statement
- Re-decidingagree the revisit trigger in advance
Common questions
Is it wrong to change my mind?
No. Changing your mind on new information is the point. Changing it on the same information, because time has passed and the discomfort has grown, is error eight, and the difference between the two is exactly whether anything new arrived.
What if the outcome really was somebody's fault?
It can be, and the review still starts with the record rather than the result. Beginning with what happened makes the fundamental attribution error almost unavoidable, because the fastest answer to why is a who.
How long should a decision record be?
Long enough to contain the options, the decider, the reasoning, and what would change it. That is often four sentences, and four sentences written at the time beat four pages written afterward.
Which of these should I fix first?
The date and the decider, because errors four, five and seven all disappear when those two are stated. The roles and process guide covers how to say them without turning a small choice into a procedure.
