Kinds of hard
A leader facing a hard decision reaches, almost by reflex, for the same shelf: the pros-and-cons list, the decision matrix, the scenario exercise, one more round of analysis. The shelf's promise is plain; with the right tool, the decision will get easier. The pattern we keep meeting in the coaching room says something different: the tools mostly fail, because every one of them is built on the same assumption — that what makes a decision hard is missing information.
Hardness is not one substance. Three hard decisions arriving at the same desk are usually hard for three different reasons, and each reason answers to a different move. The three kinds we meet most often in practice:
- Hard because information is missing: nobody can yet know which option works.
- Hard because values collide: both options are right by their own measure, and they cannot live together.
- Hard because the cost has a face: the decision is correct, and its consequence lands on a person the leader knows.
The first question is a diagnosis, not a strategy
So the first question we ask in the room is not what the leader will do; it is what kind of hard this decision is. Get the kind wrong and the strategy arrives wrong with it: data is gathered for a values conflict, a matrix is built against a cost that has a face, and the decision circles the same spot for weeks. Naming the kind is usually half the decision; the rest is a single move matched to it.
When information is missing
The first kind is the only one the shelf was built for. When the missing thing is information, analysis is legitimate: a market can be tested, customers can be asked, a small pilot can be run. In these decisions strategy means organising the search for evidence, and well-organised evidence really does close decisions.
Analysis needs a closing move
The trap in an information problem is not the analysis; it is analysis without an end. Past a certain point, new information feeds the waiting rather than the decision. We described the mechanics of that end in what to do when a decision will not close: the threshold and the date are set before the evidence, and what an in-between result will mean is said in advance. What this essay adds is a condition of order — that method works only when the decision is hard in the information way.
One confusion returns constantly in practice: a large share of the decisions presented as information problems turn out, on a closer look, to be one of the other two kinds. The leader asks for data because asking for data is a respectable request; putting the conflict, or the face, on the table is not. The simplest sign that a decision is genuinely waiting for information is that one sentence can say which evidence would settle it, and in which direction. When that sentence cannot be written, information is not what is missing.
The simplest sign that a decision is genuinely waiting for information is that one sentence can say which evidence would settle it, and in which direction.
When values collide
The second kind does not yield to data, because both sides are right. Growth against margin, speed against care, loyalty against competence: the decision between them is not a question of arithmetic but a question of precedence. Arithmetic asks which number is larger. Precedence asks who the company is; no spreadsheet answers that.
Data becomes a way of waiting
In a values conflict, asking for one more round of analysis is the most respectable available form of not deciding. The more data arrives, the more each side finds proof of its own rightness, because each side is right. What closes the decision is not a new number but a stated order of precedence — which value leads in this period — and that is not an analysis. It is a declaration.
The declaration is usually larger than the decision that forced it. What looks like a single pricing call is a sentence about what the company will be for the next two years, and where that sentence has never been said, every similar decision restages the same conflict. The sentence work we described in the fear behind an unclear direction returns here; in a company where the direction sentence exists, values collide once per era, not once per decision.
When the cost has a face
The third kind is the least discussed. The decision itself is clear; the leader has usually known for months what has to happen. The hardness is not in the choice but in its consequence: the role being taken away belongs to a name, the unit being closed holds somebody's years. The delay this produces is not analysis but protection; the leader is not postponing the decision, only the conversation.
Separating the decision from its delivery
The move that works here is not making the decision easier; it is separating the making of the decision from the delivery of its consequence. The decision is made today, alone, in writing. The conversation — how, when, with what care — is planned as its own piece of work. Left unseparated, the two hold each other hostage: the conversation is hard, so the decision is not made, and the decision is not made, so the conversation is never prepared.
This kind announces itself differently, too. In an information problem the leader asks questions, and in a values conflict the leader weighs sides; in this one, the leader changes the subject. When the decision is mentioned and the conversation slides elsewhere, what is missing is rarely courage. What is missing is that the care owed to that face has not yet been designed. Once it is designed — the day, the order, the words — the decision tends to come down under its own weight.
From the field
Seen from the field, the three kinds join into one pattern: in hard decisions, the search for a strategy is often the most industrious available way of avoiding the name of the difficulty. The shelf keeps a leader busy, the tools feel like progress, and the weeks fill. The leader who names the kind usually discovers two things at once; the decision is smaller than it looked, and the real work sits just before or just after it.
One strategy
Of the strategies available for hard decisions, the first — and usually the sufficient one — is to name, in one written sentence, what kind of hard this decision is. If information is missing, a threshold is set. If values collide, precedence is declared. If the cost has a face, decision and delivery are separated. All three moves fit inside an afternoon; none of them makes the decision easier, and each of them makes it closable.
How we work begins with this distinction. The first session does not argue the content of the decision; it names the kind. Once the kind is named, the question in the leader's hands usually becomes, for the first time, a question that can be answered.