The signal
The meeting was well prepared. That was never the problem.
The dashboard had an owner. The KPI had a steward. The process had a process owner. The system had an application owner. The programme had a sponsor. Two more people were there because the outcome touched their area, and both had a legitimate reason to be in the room.
Seven responsible people. One decision.
Someone asked the obvious question: so who decides?
What follows is usually not conflict. It is a small, polite pause. Then one of three answers.
"That depends."
"We should align first."
"Let me take it back internally."
Everyone in the room agreed on what the data said. Most agreed on what should probably happen. And the decision still did not get made that afternoon — or the following one.
This is the contradiction worth noticing: ownership everywhere, decision authority nowhere.
The pattern
Organisations are good at assigning ownership to things.
Things are stable. They can be listed, catalogued and audited. A dataset has a name. A system has a lifecycle. A process has a diagram. A programme has a budget line. Each of those can be given an owner, and that owner can be held to a standard.
Recurring decisions are not things. They arrive repeatedly, in slightly different shapes, across the boundaries of several of those owned objects. Nothing in a governance model naturally catches them, because they do not sit inside any single domain.
So the ownership map grows in one direction only. Data ownership is defined. Process ownership is defined. System ownership is defined. Risk ownership is defined. And the decisions that all of this exists to support remain, quietly, unassigned.
The result looks mature from the outside. There is a governance forum. There are named accountabilities. There are minutes. An audit would find very little wrong.
What is missing is not documentation. It is an answer to a specific question: who is allowed — and required — to make this call?
What is actually stuck
It helps to separate four things that organisations routinely treat as one.
Information ownership. Someone is accountable for definition, quality, availability, lineage and compliance. This is the layer most organisations have invested in most heavily.
Decision ownership. Someone is accountable for making a specific recurring choice, within a defined scope, by a defined moment. Not for having a view. For making the call.
Action ownership. Someone is accountable for converting the decision into something that changes in the operating reality — a price, a schedule, a rule, a resource, an intervention.
Outcome accountability. Someone is accountable for whether the intended effect actually materialised, and for what happens if it did not.
Friction rarely comes from any one of these being absent. It comes from the boundaries between them being ambiguous.
When decision ownership is unclear, information ownership starts to compensate. Data owners are asked to produce more evidence, more granularity, more scenarios — as if the missing ingredient were certainty rather than authority. The analysis improves. The decision does not arrive.
When action ownership is unclear, decisions degrade into recommendations. A recommendation is a decision that nobody has to carry, and it can be revisited indefinitely without anyone being wrong.
When outcome accountability is unclear, nobody discovers that the action never happened. The organisation still believes the matter was settled, because the meeting concluded.
Each of those gaps has the same visible symptom: the time between knowing and acting gets longer. But the cause is not slowness. It is that the chain of ownership stops one link short.
Why consensus does not close the gap
None of this is an argument against collaboration.
Cross-functional input is often the only way a consequential decision can be made responsibly. Consultation surfaces constraints that no single function can see. In regulated environments, consultation is not optional — it is the control. The organisations that skip it do not move faster; they simply move twice.
The problem is not that people collaborate. It is that the moment when collaboration ends and ownership begins is left undefined.
Without that moment, consensus stops being a way of reaching a decision and becomes a substitute for one. A few things follow, all of them familiar.
Meetings multiply, because the only mechanism for progress is getting more people in the room. Escalation becomes routine, not because the matter is exceptional, but because escalation is the only reliable way to end a discussion. Positions soften, because nobody wants to own a choice they were not clearly asked to own. And the decision, when it eventually happens, happens somewhere unrecorded — in a corridor, in a chat thread, or by default, because the deadline passed.
Consensus is a legitimate decision method. It is not a decision owner. Someone still has to be the person for whom the choice was theirs to make.
What changed
The intervention that helps is smaller than it sounds, and it is not a methodology.
It starts by making recurring decisions visible as objects in their own right. Not every decision — the handful that repeat and that materially affect an outcome. Prioritisation. Pricing. Exception handling. Capacity. Intervention thresholds. Resource allocation.
For each of those, five things get written down, once:
who provides the information; who is consulted; who owns the decision; who owns the resulting action; and how anyone knows the action occurred.
That is it. No new forum. No new role. In most cases the answers already exist informally — they are simply held differently by different people, which is precisely why the room goes quiet when the question is asked out loud.
Two effects tend to show up quickly. The decision itself gets faster, because the debate about who decides no longer runs in parallel with the debate about what to decide. And the evidence requirement gets smaller, because a named owner asks for what they need to act, not for everything that might protect them afterwards.
What we learned
Governance is usually judged by how much it covers. It is more useful to judge it by how far it reaches.
An organisation can own its data completely and still not own its decisions. The ownership map has to continue past the information layer — through the decision, into the action, and out to the evidence that something actually changed.
Governance becomes useful at the point where ownership stops describing what the organisation holds, and starts describing what the organisation does.
Seven owners is not the problem. Seven owners and no decision is.