What are Decision Rights?

Every organisation chart answers one question precisely: who reports to whom. Almost no organisation can answer the question that actually sets its speed: who decides what. Ask three executives who owns a particular pricing exception, a scope change or a key hire, and you will often get three confident, incompatible answers.

Decision rights are the explicit allocation of that authority: who may make which decision, with what input, and who must live with the consequences. The idea has a serious pedigree. The economists Michael Jensen and William Meckling argued in the early 1990s that organisations perform best when decision authority sits next to the specific knowledge the decision needs, rather than defaulting upwards to rank. Paul Rogers and Marcia Blenko of Bain & Company turned the principle into management practice with RAPID, published in Harvard Business Review in 2006: five roles in any decision (Recommend, Agree, Perform, Input, Decide) and one iron rule, that exactly one person holds the D. The RACI matrix does related work for tasks. The frameworks differ in their letters; the discipline underneath is identical. Name the owner before the decision arrives, not during it.

A decision right is not a privilege of rank. It is the named ownership of one choice: one person holds the pen, the right people are heard, and everyone can tell the difference.

Consulted is not deciding

The costliest confusion in organisational life is the difference between being consulted and holding the pen. Input is the right to be heard. A decision right is the right to choose. When the two blur, every consulted voice quietly upgrades itself to a veto, meetings stop being decision forums and become negotiations for unanimity, and the organisation falls into the consensus trap: consensus feels safe because nobody can be blamed alone, but it prices every decision at the speed of its most reluctant participant.

Clarity beats consensus, and not because consultation is waste. Consultation is where the evidence comes from; the people closest to the work usually hold the facts the decision needs. The point is what happens after the listening. People accept decisions they argued against far more readily than decisions they cannot trace to an owner. What breeds resentment is rarely being overruled. It is discovering that the decision happened somewhere, at some point, by someone, and nobody can say who.

The slowest decision is the one nobody knows they own

Unclear rights produce a distinctive failure: the orphaned decision. It is not decided badly. It is not decided at all. It sits between functions because each function sincerely believes another holds the pen. Nothing escalates, because escalation requires someone to recognise themselves as the blocker, and nobody does. The decision simply ages while its options expire.

This is the slowest kind of decision precisely because its delay is invisible. A decision with a named owner can be late, everyone can see that it is late, and the owner can be asked why. An orphaned decision emits no signal at all. By the time a deadline drags it into the light, its decision window has narrowed, the better options have expired, and the choice makes itself. The organisation still decided. It decided by default, at the worst available moment, and the latency never appeared on any report.

One concrete example

Clearly illustrative, with no customer implied. A services firm of a few hundred people is negotiating a renewal, and the client asks for a meaningful discount in exchange for a longer term. The account lead believes pricing exceptions belong to finance. Finance believes the account lead owns the relationship and therefore the call. The delivery director, who must staff whatever is promised, is in neither conversation. For weeks the decision belongs to everyone and therefore to no one. The client reads the silence as reluctance and opens a conversation with an alternative provider. The discount is finally approved on a hurried call, steeper than the one originally requested, and nobody has priced what the longer term does to delivery cost, because the person who must live with the decision never held any right in it. Every individual behaved sensibly inside their own understanding. The failure was structural: the right was never named, so the decision was never owned.

Writing rights down, and keeping them honest

The first half of the discipline is naming. For each recurring class of decision, name one owner, name the input the owner must take, and name who must live with the result: the person who bears the consequences deserves a seat in the decision even when they do not hold the pen. The second half is the record. A rights matrix that lives in a slide deck from a reorganisation two years ago is folklore, not governance. Rights stay honest only when each significant decision records, as it is made, who held the pen, who was consulted and what they said, and whether the owner went against a recommendation.

Decision rights in decision intelligence

This is where the old discipline meets decision intelligence. When a decision is captured as a Decision Object, ownership stops being folklore, because the owner is part of the record. In a Decision Room, candidate options are ranked against the constraints that bind them, and then a person decides: a human always holds the pen, and when the decider overrules the recommendation, the override is recorded with who, when and against what evidence. That record accumulates into a decision audit trail: not surveillance, but the end of ambiguity about who decided what.

Decision rights sound like bureaucracy until you watch a company operate without them. Then they look like what they are: the difference between an organisation that makes its decisions and one that lets its decisions happen to it.