A surprising amount of organizational friction comes from one missing sentence: who gets to decide?
Not who has opinions. Not who should be consulted. Not who will be affected. Who can say yes, who can say no, who can say not yet, and who has to live with the consequences.
When decision rights are unclear, power does not disappear. It becomes informal. The person with the most persistence wins. The executive who speaks last wins. The function with the scariest risk wins. The team that owns the dependency wins. The person who can delay without saying no wins.
That is not collaboration. It is hidden power.
Decision rights are power architecture
Decision rights define how authority moves through the company.
They determine who can approve, veto, recommend, escalate, sequence, allocate resources, accept risk, define standards, or make exceptions. They also determine who can block a decision without formally owning it.
Most companies write down org charts and ignore decision rights. That is a mistake. The org chart shows reporting lines. Decision rights show how work actually moves.
A product leader may own roadmap priorities, but security may have release veto power. A sales leader may own discounting within a band, but finance may control exceptions. A manager may own team priorities, but a founder may retain informal veto over product taste. A program lead may be accountable for a launch but have no authority over the teams whose work determines whether it ships.
That mismatch is where work gets stuck.
The five kinds of decision power
Operators should distinguish at least five decision powers.
Approval power is the right to say yes. It authorizes action, spending, hiring, launch, policy, or direction.
Veto power is the right to say no. It may be formal, as in legal or security review, or informal, as when a senior leader's disapproval effectively kills momentum.
Delay power is the ability to say not yet. This is often the most hidden form. People delay by requesting more analysis, adding stakeholders, reopening settled questions, withholding resources, or refusing to schedule the meeting where a decision can happen.
Recommendation power is the ability to shape the option set. The person who frames the choices often has more power than the person who selects among them.
Exception power is the right to override the standard rule. Exceptions reveal the real hierarchy. If the written policy says one thing but a certain executive can always waive it, the exception path is part of the true decision architecture.
A healthy organization names these powers. An unhealthy one lets them remain ambiguous and then calls the confusion alignment work.
Delay is often disguised veto
Delay deserves special attention because it is the polite form of power.
Few people want to be seen as blockers. So they ask for more input. They suggest another review. They reopen scope. They request a cleaner business case. They say the timing is not right. Each request sounds reasonable in isolation. Together, they prevent a decision from becoming action.
Sometimes delay is legitimate. Work may be under-specified, risky, badly sequenced, or disconnected from strategy. Saying not yet can protect the company.
But not yet should still be accountable. If someone delays work, they should be clear about what condition would change the answer, when the decision will be revisited, and what cost the delay creates.
Unaccountable delay is veto power without the courage to say no.
Accountability must follow decision rights
A decision right without accountability becomes privilege.
If legal can stop a launch, legal should help define the risk threshold and review path. If finance can reject a hire, finance should explain the constraint and tradeoff. If an executive can override the roadmap, that executive should own the downstream cost to customers, engineering focus, and morale. If a team can block a dependency, it should be visible in planning as a capacity and priority decision, not treated as random friction.
The rule is simple: the more power someone has to shape the decision, the more responsibility they should carry for the consequences.
This does not mean every stakeholder owns the outcome equally. It means vetoes, delays, and exceptions should not be free.
One useful standard is consequence symmetry: if you can impose a cost, you should help name it; if you can delay, you should define the condition for movement; if you can override, you should be visible as the override owner. Hidden asymmetry is where decision power turns into resentment.
Decision rights are not stakeholder management
A common mistake is solving decision-rights problems with more stakeholder management.
People add meetings, pre-reads, syncs, steering committees, alignment docs, and approval chains. Sometimes that helps. Often it just gives unclear power more surface area.
Stakeholder management asks, “Who needs to be informed, consulted, or brought along?”
Decision-rights design asks, “Who decides what, with what input, under what constraints, by when, and with what escalation path?”
Both matter. But only the second prevents endless ambiguity.
The decision-rights map
For any important initiative, write down:
- Decision to be made: what exactly is being decided?
- Decision owner: who is accountable for making the call?
- Approvers: who must say yes?
- Veto holders: who can say no, and on what grounds?
- Input providers: who must be consulted but does not decide?
- Recommendation owner: who frames the options and tradeoffs?
- Escalation path: where does disagreement go if unresolved?
- Deadline: by when must the decision be made?
- Exception path: who can override the normal rule?
- Consequence owner: who owns the result after the decision?
This does not have to be bureaucratic. A short table in a memo is often enough. The point is to make power legible before the work depends on it.
Ethical decision power
Decision rights should not be used to punish, humiliate, or hoard control. They should make action more legitimate.
Ethical decision power has a few habits:
- decisions are made at the right level, not always the highest level;
- people affected by a decision are heard early enough to matter;
- vetoes are tied to explicit risks or standards;
- delays include a condition and a date;
- exceptions are visible and explainable;
- credit and accountability are distributed honestly;
- authority is delegated with the context needed to use it well.
The goal is not to eliminate hierarchy. The goal is to prevent hidden hierarchy from masquerading as collaboration.
The hard truth
If nobody can name who can say yes, no, or not yet, the organization is not being collaborative. It is letting power operate in the dark.
Strong operators make decision rights visible so work can move with speed, legitimacy, and accountability.
