Which decision skills trends name a real property, not just a preference?

A handful of decision terms have spread out of specialist use and into ordinary meetings. Some of them name a real distinction and are worth keeping. Some are a mood with a technical accent. This page sorts them, using one test: does the term let you say something the plain words could not?

What to take away

One-way and two-way doors

This pair passes the test. It names reversibility, which is a real property, and it makes a practical point that the plain words obscure: the amount of process a decision deserves tracks reversibility rather than cost. That argument is set out in full in the roles and process guide.

The failure mode is treating the two as a fixed classification. Reversibility is a matter of degree and it changes with time, with who else is involved, and with what has already been announced.

Disagree and commit

Names a real and useful state: having argued your case, lost, and chosen to support the decision as though it were yours.

It fails when it is used as an instruction rather than a description. Told to someone who has not yet been heard, it means stop objecting, and it converts a useful norm into a way of closing discussion early. The order matters: the disagreement has to have been genuinely received first.

Pre-mortem

A structured exercise: assume the plan failed, and ask why. It works because it

changes the question from whether something could go wrong, which invites a defense, to why it did, which invites explanation.

This one passes the test comfortably. It is the same move as writing down in advance what would change your mind, and easier to run in a group. It also works against hindsight bias in the other direction.

Having imagined the failure beforehand, you are less able afterward to claim it was obvious, and the decision record fields are the natural home for what it produces.

Role notation

Various schemes exist for writing down who decides, who is consulted and who is informed. The notation is not the value; the conversation the notation forces is.

The test here is whether anyone can say, afterward, what the letters meant for their own participation. Where they cannot, the exercise produced a grid rather than a set of roles.

Terms that do not pass

The right-hand column is the finding. Each of these can appear in a sentence without changing what the sentence asserts, which is the substitution test the corporate jargon guide applies to workplace language generally.

"Data driven" deserves a note. It is not empty when someone says which data, collected how, and what it would have shown if the answer were the other one. Without those, it is a claim of virtue.

The borrowed-assumption problem

Terms imported from other fields bring their conditions with them, and the conditions usually get left behind. A method that works where feedback is fast and outcomes are unambiguous behaves differently where feedback is slow and outcomes are contested.

Before adopting an imported decision method, ask what the original field could rely on. Often, it is repeated trials, a clear measure of success, or a low cost of being wrong.

Methods that assume a known distribution of outcomes, such as any built on an expected value calculation, are only as good as their estimates.

In business, those estimates are usually the disputed part.

Where your situation has none of those, the method may still help, but not for the reason it worked originally. The decision review measures cover which parts of a process are checkable in a slow-feedback setting.

The test, stated once

A term that survives all four is worth using and worth defining on first use. A term that fails the second is worth dropping quietly, without a campaign about it.

Common questions

Should I adopt a decision framework at all?

Adopt the questions, not the framework. The questions travel and the frameworks do not, and a framework adopted whole usually brings a vocabulary your colleagues did not agree to.

Are these terms going to last?

The two that name properties probably will, because the properties are real. The ones that name preferences will be replaced by different words for the same preferences, which is the ordinary way workplace vocabulary turns over.

What about tools that promise to make decisions for me?

Whatever produces a recommendation, the questions are unchanged: what were the options, what were the criteria, and what would have changed the answer. A recommendation you cannot interrogate is not evidence, it is an assertion with a different source. The assessment guide makes the same point about instruments that produce a score.

How do I introduce one of these without sounding like I read a book?

Use the question rather than the term. Asking whether we can undo this cheaply gets the benefit of the one-way door distinction without asking anyone to learn a phrase.