Design the decision, not the org chart.

Ask an organization to describe how it is set up and it will show you an org chart. Ask it to describe how a specific important decision gets made, who proposes, who is consulted, who decides, who can veto, and you will usually get silence, or a description that three people in the room would each dispute.

This is backwards. The org chart tells you who reports to whom, which matters far less to execution than how decisions actually get made, which most organizations have never deliberately designed at all. Structure is drawn with great care. Decision architecture is left to emerge, and what emerges is usually slow, contested, and unaccountable.

The invisible system

Every consequential decision runs through an implicit process. Someone frames it, some people weigh in, someone decides, and the decision either sticks or gets relitigated. In most organizations this process is undocumented and inconsistent, varying by decision, by personality, and by who happens to be in the room. The same class of decision might be made three different ways in a month.

Because the process is invisible, its failures are invisible too. Decisions stall and no one can say why. Choices get made and then unmade because it was never clear who actually held the authority. Work waits on approvals that no one owns. The organization experiences this as friction, slowness, or politics, without recognizing that it is the absence of designed decision architecture.

You can have a perfect org chart and still be unable to make a decision, because the chart was never the thing that made decisions.

What designing the decision looks like

Decision architecture makes the implicit explicit. For the decisions that matter most, it defines four things clearly. Who has the authority to decide. Who must be consulted before the decision but does not hold a veto. Who is simply informed after. And what threshold moves a decision from one level to another.

This is not bureaucracy. Bureaucracy is process for its own sake. This is the opposite: a deliberate reduction of the ambiguity that makes decisions slow and reversible. When the four roles are clear, a decision can be made once, by the person who owns it, and stay made. That single property, decisions that stay decided, is worth more to execution than almost any structural change.

Decide the decisions in advance

The leverage comes from designing the architecture before the pressure arrives, not during it. In the heat of a contested choice, everyone has an incentive to claim the authority that favors their position, and sorting out who really decides becomes its own political fight layered on top of the actual decision.

Designed in advance, calmly, the decision rights are a neutral fact rather than a live negotiation. When the hard choice comes, the question of who decides is already settled, and the organization can spend its energy on the substance instead of on the procedure. This is why decision architecture belongs to the design phase of a transformation, alongside the operating model, not to the moment a decision is stuck.

The sponsor owns the hardest ones

Some decisions cannot be delegated. The tradeoffs between scope, time, and cost, the choices that create winners and losers across the organization, the calls that require spending political capital, these belong to the sponsor. A decision architecture that pushes them downward guarantees they will be made badly or not at all, because the people below lack the authority and the standing to make them hold.

Designing decision architecture therefore forces a healthy conversation about what the sponsor must personally own. That conversation is uncomfortable, which is exactly why it is valuable. A sponsor who has explicitly accepted their decisions is a sponsor who cannot later go quiet without it being visible.

Draw the decisions before the boxes

If a transformation is stalling on friction that no reorg seems to fix, the missing design is probably not structural. It is the architecture of how decisions get made, which no one ever drew. Map the decisions that matter, assign each one an owner, settle the thresholds, and put the hardest ones where they belong. The org chart can wait. The decisions cannot.