Design decision rights so decisions land right

In many organisations, a surprising amount of leadership time goes into decisions that should never have reached the leadership team. A price exception for a mid-sized customer. A change to a supplier agreement. Whether to hire a temporary replacement for someone on leave. Each of these lands on the table because nobody was sure whether they were allowed to make the call themselves.

At the same time, some important decisions are made almost by accident. A team quietly changes how it handles complaints. A project grows in scope without anyone agreeing to it. A purchase is made that commits the business for years. Nobody intended to bypass anyone. There simply was no clear understanding of who decides what.

Both problems have the same root. Decision rights, the agreed answer to who decides, who must be consulted and who needs to know, are left unspoken. This article is about making them explicit, in a way that speeds the organisation up rather than adding another layer of process.

Why unclear decision rights cost so much

When decision rights are unclear, a few predictable things happen.

Decisions drift upwards. If people are unsure whether they can decide, the safe option is to ask someone more senior. That person may be equally unsure and ask further up. Over time, the leadership team becomes a bottleneck for choices that the people closest to the work are better placed to make.

Decisions slow down. Every unclear decision gathers a queue of meetings, emails and drafts. People wait for approvals that nobody is formally required to give. A choice that could have taken a day takes a month.

Decisions become political. When it is not clear who decides, influence matters more than responsibility. The person who speaks loudest, has the best relationship with the CEO or simply refuses to agree can hold a decision hostage. People learn that the way to get things done is to lobby, not to make a good case.

Decisions get reopened. A choice that was made without the right people feeling they had a say is fragile. It comes back at the next meeting, or it is quietly ignored in practice.

Most slow decisions are not hard decisions. They are decisions without a clear owner.

The good news is that this is fixable, and the fix is mostly a matter of agreeing and writing down what is already half understood.

Not all decisions are the same

Before deciding who should make which decisions, it helps to sort them. Two distinctions are particularly useful.

The first is between decisions that are easy to reverse and decisions that are not. Some decisions are like a door you can walk back through: if the choice turns out to be wrong, you can undo it at modest cost. Trying a new meeting format, adjusting a campaign, changing a work routine. Others are like a door that locks behind you: a major investment, a long-term contract, closing a location, a key hire at the top. These are expensive or impossible to undo.

Reversible decisions should be made quickly and as close to the work as possible. The cost of a slow decision is usually higher than the cost of a wrong one that can be corrected. Irreversible decisions deserve more care, more input and often a more senior owner.

The second distinction is between frequent and rare decisions. Frequent decisions, such as discounts, small purchases, scheduling or handling standard complaints, benefit from clear rules and wide delegation. If the same type of decision is made fifty times a year, it is worth defining once how it should be made. Rare decisions, such as entering a new market or changing the organisation structure, are handled individually, but it still helps to know in advance who owns them.

Put together, these give a simple rule of thumb:

• Frequent and reversible: delegate widely, with simple guidelines.
• Frequent and hard to reverse: delegate with clear limits and principles.
• Rare and reversible: let the responsible manager decide, and inform others.
• Rare and hard to reverse: decide at the top, with structured input.

Who decides, who is consulted, who is informed

There are many models for describing decision rights, and most of them share the same core. For each decision, you need to answer three questions:

1. Who decides? One person, or in some cases one named group, who makes the call and is accountable for it. If the answer is several people who must all agree, there is no clear owner.
2. Who must be consulted? The people whose knowledge or perspective is needed before the decision is made. They are entitled to be heard, and the decision-maker has to listen.
3. Who must be informed? The people who need to know the outcome so that they can act on it, but who do not need to be involved in making it.

This sounds simple, and it is. The value lies in writing it down and agreeing on it, so that people stop guessing.

A few practical guidelines make it work better:

• Keep the decision-maker to one role. Shared ownership usually means slow ownership.
• Keep the consulted group small. Ask who genuinely adds something, not who might feel left out.
• Put a time limit on consultation. Input that arrives after the decision has been made is not input.
• Make informing part of the decision. A decision is not finished until the people who need to know have been told.

Input is not the same as veto. The most common confusion in decision rights is between giving input and holding a veto. They are very different things, and mixing them up is a major cause of slow and political decisions.

Input means: I have a perspective you need to hear, and you have to take it seriously. But once you have heard me, you decide, even if you decide differently from what I recommended.

A veto means: this decision cannot go ahead without my agreement.

In many organisations, people who are meant to give input behave as if they hold a veto, and decision-makers treat them as if they do. Finance raises a concern, and the decision stops. A department head says they are not comfortable, and the proposal goes back for another round. Nobody has formally said that these people can block the decision, but in practice they can.

Some vetoes are legitimate. Legal and compliance should be able to stop a decision that breaks the law. Safety should be able to stop a decision that puts people at risk. The finance function may hold a veto on commitments above a certain size. These vetoes should be explicit, limited to a clear area and few in number.

Everyone else gives input. Making that distinction clear is often the single change that does the most to speed up decisions. It also tends to improve the quality of input, because people who know they cannot block a decision put more effort into making a convincing case.

A decision register for the twenty that matter

You do not need to map every decision in the organisation. In most businesses, a fairly small number of recurring decisions cause most of the delay, friction and debate. Start with those.

A decision register is a simple list of the twenty or so most important recurring decisions, with the decision rights for each. For every decision, write down:

• The decision, described concretely. Not pricing, but approving a discount above the standard range for an existing customer.
• Who decides.
• Who must be consulted.
• Who has a veto, if anyone, and on what grounds.
• Who must be informed.
• Any limits or guardrails, such as amounts, time frames or principles.
• When the decision should be escalated, and to whom.

To find the twenty, ask the leadership team and a handful of middle managers a few questions. Which decisions come back again and again? Which decisions do people regularly argue about? Where do things get stuck waiting for approval? Which decisions have caused trouble when they were made the wrong way? The answers will overlap quickly, and a list will emerge within a couple of conversations.

Typical candidates include price exceptions and discounts, hiring and replacement, purchases above a certain size, changes to customer contracts, prioritising between projects, budget reallocations during the year, and changes to products or services.

Escalation rules that work

Even with clear decision rights, some decisions will need to go higher. The question is when, and how.

Good escalation rules are specific. Instead of escalating when something is important, define the triggers:

• The decision exceeds a stated limit, for example an amount, a contract length or a number of people affected.
• The decision affects another unit's area and the two units cannot agree within a set time.
• The decision conflicts with an agreed principle or policy.
• The decision carries a risk that the owner is not authorised to accept, such as legal, safety or reputational risk.

Equally important is how escalation is done. A good escalation brings a clear question, the options considered, a recommendation and the reasoning behind it. It is not a way of handing over a problem. The person escalating should still own the preparation, and the person receiving it should respond quickly.

It also helps to agree that escalation is normal and not a sign of failure. If escalating feels like admitting defeat, people will either avoid it when they should not, or do it constantly to avoid risk. Neither is useful.

Delegating with guardrails

Delegation fails in two ways. Either leaders delegate in name but keep second-guessing every decision, or they delegate without any boundaries and are then surprised by the outcome. Guardrails solve both problems. They make clear what the person can decide on their own, so that the leader can step back with confidence.

There are three main types of guardrail:

1. Budgets. The person can spend within an agreed budget without further approval. Spending above it needs a decision from someone else.
2. Limits. Clear thresholds for amounts, discounts, contract terms, number of hires or other variables. Inside the limit, decide. Outside it, escalate.
3. Principles. Short statements that guide judgement where rules would be too rigid. For example: we do not offer discounts to win a first order unless the customer commits to a minimum volume. Or: we always prioritise delivery commitments to existing customers over new projects.

Principles are often the most powerful guardrail, because they allow people to handle situations nobody foresaw. They work best when they are few, concrete and explained with an example.

Say a sales manager can approve discounts up to 10 percent, needs to consult the commercial director between 10 and 15 percent, and must escalate anything above 15 percent. That one line removes dozens of emails a month, and the commercial director only sees the cases that actually need her attention.

Start with guardrails that are a little tight and loosen them as trust grows. It is much easier to widen a limit than to narrow it after things have gone wrong.

Introduce it lightly and keep it current

The risk with any work on decision rights is that it turns into a large exercise in charts and matrices that nobody uses. That is avoidable.

Keep it small. Start with the twenty decisions, not two hundred. Use plain language and a format people already know, such as a shared document or a simple page on the intranet.

Involve the people who make the decisions. Draft the register with the managers who own the decisions and the people who are consulted. This catches mistakes early and gives the result legitimacy. Decision rights that are imposed from the top tend to be ignored.

Test it in practice. Use the register for a couple of months and note where it is unclear or wrong. Adjust as you go, rather than trying to get everything right the first time.

Lead by example. The most important signal comes from the leadership team. If a leader overrules a decision that has been delegated, without a very good reason, everyone learns that the register means nothing. If leaders consistently respect it, and say so, people start to trust it.

Review it over time. Decision rights are not set once and left alone. As the business changes, so do the decisions that matter and the people best placed to make them.

A simple review, once or twice a year, is usually enough. Ask:

• Which decisions are still slow or contested?
• Which decisions are being escalated more often than expected?
• Where have guardrails been too tight, or too loose?
• Have roles changed in a way that should move ownership?
• Are there new recurring decisions that should be added to the register?

Where to start

You can begin within a week, without a project:

1. Ask the leadership team to list the ten decisions that most often get stuck, come back or cause argument.
2. For each, agree who decides, who must be consulted, who is informed, and whether anyone holds a veto.
3. Add simple guardrails and escalation triggers where they are needed.
4. Write it down on one or two pages and share it with everyone affected.
5. Use it for two months, then review what worked and extend the list towards twenty.

The goal is not a perfect map of every decision. It is an organisation where people know which calls are theirs to make, make them with confidence, and bring the right decisions to the right level at the right time.