Decision rights and delegation: who decides what

• Reversible and small. Pricing an individual quote within a band, scheduling a delivery, choosing a supplier under a limit. These can move furthest down the organisation, because a mistake is recoverable.
• Reversible and large. A new process, a system change, a campaign. Delegable with a reporting requirement rather than an approval requirement.
• Hard to reverse. Hiring a senior role, a contract with a long term, a platform migration. These usually stay with the owner, or move only with an explicit review step.
• Irreversible. Redundancies, exiting a market, sale of the business. These belong with the owner, and pretending otherwise creates confusion rather than delegation.

Sorting by reversibility first is more useful than sorting by size or by function, because it matches the decision to the cost of being wrong.

1. Attach a number where a number exists: contract value, discount percentage, spend limit, headcount.
2. Attach a category where the number is not the point: a new market, a regulated customer, a non-standard term.
3. State the route when the threshold is crossed: who is asked, in what form, and by when.
4. Say what must be reported afterwards, even when no approval was needed. Reporting is what makes delegated authority visible without slowing it down.

A useful test is to hand the written threshold to someone who was not in the conversation and ask them to classify three real cases. Any disagreement identifies a gap in the wording rather than a problem with the person.

Delegation does not remove the owner from the work; it changes what the work is. What remains is setting the limits, providing the information people need to decide well, and being available when a threshold is crossed. It also includes the less comfortable part: not reversing a decision that was within someone's authority, even where a different call would have been made.

Where an owner repeatedly reverses delegated decisions, the usual cause is that the limits were never genuinely agreed. Reopening the limits is more honest than continuing to overrule people inside them.

• Review at a fixed interval, quarterly for most businesses.
• Look at what was escalated: frequent escalation of the same decision means the limit is wrong.
• Look at what was decided and never reported, which usually means the reporting requirement is unclear.
• Move limits deliberately as confidence and capability grow, and say so, so the change is visible.

The pattern worth watching is escalation that is really a request for permission nobody can give. Where the same decision returns three times, the threshold or the role definition is the problem, not the person raising it.

The example below is hypothetical and describes a pattern rather than a client engagement.

• Written rights can outrun capability. Transferring authority to someone without the information or experience to use it produces confident mistakes. Limits should move with evidence.
• Thresholds invite edge cases. Someone will find a case the wording does not cover. The response should be to classify it once and update the document, not to rewrite the framework.
• Documentation does not create trust. Where the underlying relationship is weak, a decision-rights document is read as surveillance. It works better as a record of an agreement already made.
• Some owners will not let go. Where the founder or chief executive retains every meaningful decision, delegation produces task transfer only, and the constraint stays where it was.

Decision rights sit inside the Leadership & People capability, alongside the role definitions that make authority meaningful. Where ambiguity is affecting work that changes hands, the notes on role clarity and accountability cover the adjacent question.