The Architecture of Clarity: Who Decides What

Businesses slow down when decision rights are undefined. Work stops moving not because the choice is difficult but because it is unclear who is entitled to make it, so everything routes to whoever has the most authority — usually the founder, who becomes the bottleneck for the entire organisation.

Key takeaways

• Ambiguity about who decides costs more than making occasional wrong decisions.
• Most decisions are reversible and should be delegated on that basis.
• Decision rights need writing down; assuming they are understood is how they stay unclear.
• The founder's calendar is the clearest diagnostic available.

The diagnostic

Look at where time goes and where work waits:

• What is currently waiting for the founder's input?
• How long do things sit in that queue?
• Of those items, how many genuinely required that level of authority?
• How often is the eventual decision the same one the person asking would have made?

If most decisions routed upward would have been decided identically at a lower level, you do not have a judgement problem. You have an authority problem, and it is costing you the waiting time on every one of them.

Sort decisions by reversibility

The most useful distinction is not size but reversibility:

Reversible Can be undone at modest cost. Most operational decisions belong here — pricing an individual job within a range, sequencing work, choosing a supplier for a one-off order. These should be delegated with a stated boundary, and speed matters more than precision because a wrong answer can be corrected.

Hard to reverse Long commitments, significant capital, anything affecting reputation or legal position. These deserve deliberation and senior involvement, and there are far fewer of them than most organisations behave as if there are.

Treating every decision as irreversible produces an organisation where nothing moves without permission, and permission is always scarce.

Writing decision rights down

For each recurring decision type, state four things:

1. 1.Who decides — a named role, not a committee.
2. 2.The boundary — the range within which they decide alone.
3. 3.Who must be informed — after the fact, not before.
4. 4.What escalates — the specific conditions requiring a higher level.

The boundary is what makes this real. "Use your judgement" is not delegation, because the person still cannot tell whether they have overstepped until after they have.

The founder's transition

Moving from deciding to defining is the hardest transition in a growing business, and it fails predictably when founders delegate tasks while retaining judgement. Transferring judgement means writing down the criteria you actually apply — usually for the first time, since experienced judgement is mostly tacit.

Expect the first months to produce decisions you would have made differently. The relevant comparison is not against your decision; it is against the cost of every decision waiting for you.

Signals it is working

• Work moves without waiting for a specific person's availability.
• Escalations concern genuinely unusual cases, not routine ones.
• People can state their own boundaries without checking.
• The founder's time shifts from approving to deciding what matters.

Where to start

Log everything that waits for you for two weeks. Sort it by reversibility. Everything reversible gets a named owner and a stated boundary this month. That single exercise typically removes most of the queue.