Who Decides What in a Leadership Team Is a Short List

The pattern is easiest to see in a single quarter of one team. A digital bank of a few hundred people runs an eight-person leadership team that meets every Monday. Over three months the room approves a senior engineering hire that was already inside the technology budget, debates the wording of a customer email, and signs off a vendor renewal the chief operating officer had authority to sign alone. In the same quarter the chief technology officer moves half of the engineering capacity from the lending app to a compliance remediation. Nobody hides it. It sits inside the technology remit as written. The lending launch slips by a quarter, and the commercial lead learns about it from a release note.

A full agenda and an absent decision

Nothing in that quarter was negligent. Each person acted inside their role, and the meetings ran on time. The team spent its shared hours on decisions any one member could have made, and left the one decision only the team could make to whoever happened to hold the resource. That inversion is the most common answer to the question of who decides what in a leadership team, and almost nobody chose it.

It forms for two ordinary reasons. Members bring their own items to the room because an approved decision is a shared one, and shared decisions are safer to own. Meanwhile the cross-functional decision has no natural proposer, because it lives inside one member's budget and that member has no reason to ask permission to spend it.

Most teams answer the question with the organisation chart: each executive decides inside their function, and the CEO decides whatever is left. That covers most decisions. It misses the class that does the damage, which is a decision made in one function whose consequences are carried in another.

The test of where the cost lands

The test that sorts this class has two parts. The first is whether the decision moves a number that a different member answers for. The second is whether any single member can see the whole trade-off from their own seat. When the cost lands on more than one member and no single member can see all of it, the decision belongs to the team. When the cost lands on one member, that member decides, however large the sum. Size is not the criterion. The same logic, applied to where information and consequence meet inside a growing company, is set out in how to assign decision rights in a scaling company.

Applied to most leadership teams, the test produces a short list. The decisions that reliably pass it are these:

• the allocation of shared capacity, such as engineering, data or a shared sales team, across functions
• changes to a number another member is measured on, such as a discount band that moves gross margin
• commitments that bind the whole company externally: to a regulator, a large partner, or a public date
• the order of two priorities once they collide on the same people or the same money

What the team does not decide

The senior hire inside budget, the vendor renewal and the email all fail the test. Each of them lands on one member. Bringing them to the room looks like collaboration and works as a request for cover, and it spends the hour the capacity decision needed. Those items still reach the team, as a written note that the member has decided, not as a question.

There is a third category, and in most teams it is left unwritten. Some decisions belong to the CEO alone: appointing and removing members of the team, the final call when the team cannot agree on one of its own decisions by the date it set, and the few matters the board holds the CEO personally to.

Reserved rather than residual

When that list is not written down, the CEO's authority becomes residual. Anything nobody else clearly owns drifts upward, every ambiguous case waits for one person, and the weekly meeting turns into the briefing described in a CEO's weekly meeting as eight public one-to-ones. A written reserved list keeps the CEO out of the team's decisions as firmly as it keeps the team out of the CEO's. On a team decision the CEO's job is to see that it is made by its date, and to break the tie if it is not.

A team decision is not a consensus decision

Calling something a team decision does not mean everyone has to agree to it. Each decision type on the team's list carries a named recommender, usually the member closest to the trade-off, who brings a proposal with the options and their costs. The team discusses, the recommender revises, and if the room is still split on the date, the CEO decides. Why unanimity fails on exactly these allocation decisions is covered in when consensus is the wrong decision rule.

One page, three lists

The artefact is a single page. It holds the team's list, each entry with its recommender, and the CEO's reserved list. It closes with one default sentence: every other decision belongs to the member whose function carries its cost, who tells the team in writing once it is made. The page fits on a screen, and it is meant to.

In the bank above, the first month after the page was agreed looked different. The hire and the vendor renewal left the agenda. Engineering capacity became a standing monthly item, with the technology lead recommending and the commercial lead's objection heard before the plan was set rather than after the release. The hour got shorter and harder, which is the usual sign that the right decisions have reached it.

The hour got shorter and harder, which is the usual sign that the right decisions have reached it.

The drift to watch for

The page decays in two predictable ways. Member decisions creep back onto the agenda for cover, one reasonable exception at a time. And new cross-functional decisions appear that the list does not name, because the company has grown into them. The signal that the list needs revising is a member being surprised by a decision that moved their numbers. A surprise means a cost landed on someone who had no seat in the decision, which is the condition the list exists to prevent. Once a decision type is on the team's list, the roles inside it can be set out more finely; the RAPID framework is one way of doing that.

We usually start with three months of the team's minutes and sort every item into one of three piles: decided by the team, decided by a member, decided by the CEO. Then we build a fourth pile, the decisions that never reached the room but changed another member's numbers. The fourth pile is usually the more revealing one, because it shows where the team's real decisions are being made without it.

From there our consulting work inside the decision architecture pillar writes the three lists with the team and sets the revision trigger, and our training programmes build the recommender role so that team decisions arrive ready to be made. Who decides what in a leadership team is settled by a short page agreed in advance, not by whoever moves first.