A practical framework for setting manager span of control
Three tests, applied in order, for deciding how many direct reports a manager can carry as a team grows. This is practitioner judgement, not published research.
Most growing companies set the number of direct reports by accident. A founder hires a sixth engineer, the sixth reports to the same person, and nobody notices until the manager sits in every meeting and no decision moves without her. Span of control is the count of people a manager leads directly. It is a simple number that hides the work behind it. This framework asks three questions in a fixed order and gives a test for each. Be clear about its status. The sources checked for this piece contain no research on management structure, so the framework carries no figures. It is a method for judgement. Treat the number it produces as a hypothesis to review, not a rule to defend.
Chapter one: Pillar one: role complexity
Start with the work each report does, not with the manager's diary. A team doing the same repeatable task can be supervised more widely than a team where each person solves a different problem every week. Ask how often a report needs a decision from the manager before the work can continue. Where the answer is rarely, the span can widen. Where the answer is most days, the span should narrow, because the manager's attention has become the bottleneck and no amount of process will free it.
A simple test works well. List every role on the team and mark each one as routine, varied or novel. A routine role is one where the written process answers most questions. A varied role follows known patterns with frequent exceptions. A novel role is one where the company has not yet written the answer down, often because the product, market or customer is new. If most roles are novel, the manager who leads them directly should carry fewer people than your first draft suggested. Judge this on the team as it is working now, not on the job descriptions.
What usually goes wrong is grading by seniority. Leaders assume that senior people need less supervision and junior people need more, and draw the span to match. That shortcut fails in both directions. A senior specialist in a long-established process may need almost no decisions from the manager. A junior hire sent into a new market may need a decision every hour. Grade the role, then grade the person, and keep the two judgements separate on the page.
Chapter two: Pillar two: manager experience
The same team can suit one manager and overwhelm another. Experience here means experience of leading this kind of work, not the number of years in the job. A manager who has already taken a similar team through a reorganisation can usually hold a wider span than someone promoted from within the team they now lead. The second person may still be doing the work of their old role alongside the new one, and the span count does not show that.
A useful test is to ask the manager to describe, on one page, how they would spend a normal week with the team as it stands. Ask them to list the hours by kind: delivery work, one-to-ones, coaching, hiring, reviews, and meetings with other teams. If delivery work fills most of the page, the span is probably already too wide, because there is no time left for the conversations a team of that size requires. If the page has steady time for one-to-ones and reviews, the current number may be workable. Use the page to start the conversation, not to grade the manager.
The common mistake is to treat a promotion as the moment to add reports. A new manager is learning the job while carrying the team, and the first quarter is when habits set. In our experience, the cheapest time to set a span low is at the start of a new appointment, and the most expensive time to set it high is the same moment. Raising a span later is easier than shrinking one that has already spread across people who expected regular contact. Say the review date out loud when the appointment is made.
Chapter three: Pillar three: coordination load
Coordination load is the work that exists because people on the team depend on each other. Every dependency adds conversations, handovers, shared decisions and waiting time. A team of six who all depend on one another is a different problem from six people who each deliver alone, even though the head count is the same. Span of control measures the first number and ignores the second, which is why two teams with the same size can feel completely different to the manager who runs them.
The test is a count of handovers. Take one typical piece of work, such as a feature, a client account or a monthly report, and write down each point where one person hands something to another, waits on someone in the team, or waits on someone outside it. Count them. Then repeat the exercise for a second piece of work. If the count rises faster than the team grows, the manager's calendar will fill with coordination before it fills with people, and the span should narrow to match.
The fix is often fewer handovers rather than fewer people. Move a shared task to one named owner. Split one team into two with a clear interface between them, so that the dependencies run inside each half. Write down who decides what at the boundary. Recheck the span after either change, because a smaller number of handovers can make a wider span workable for the first time. Do not assume the count will fall on its own when the team gets larger.
Span of control check
Answer each question for the team as it is now, then again for the team as it is expected to be in six months.
- Does most of the work in this team follow a written process?
- Can each report finish a typical task without a decision from the manager?
- Has this manager led a team of this kind before?
- Does the manager have at least one working day a week free of delivery work?
- Do fewer than half the handovers in a typical task cross a team boundary?
- Can a new hire find the answer to a common question without asking the manager?
- Has the span been reviewed in the past twelve months?
Outgrace checklist
Where it breaks
The framework fails in three places. The first is timing. It is applied once, at hiring or at a reorganisation, and never again, while the work changes underneath it. A team that was routine last year may be novel after a product launch, and the span should follow. The second is the manager's own work. A manager who also holds a client account, a product remit or a personal delivery target cannot give the team the attention the count assumes. Subtract that work before you read the number. The third is treating the output as a ceiling rather than a starting point. The number should be reviewed each quarter and after every change to the team's structure.
A further limit should be stated plainly. This framework does not rest on published research. We looked for organisational or management studies in the sources available for this piece and found none, so the framework contains no figures and no attributed findings. Any version with citations should wait until a specific published study has been read and checked. Until then, the three tests should be judged by whether they change the decisions a leader makes, not by whether an authority has endorsed them.
On Monday, pick one team where the manager has complained about workload, or where the manager's calendar is full of delivery work. Run the three pillars on that team with the manager in the room, and keep the notes on one page. Write the current span, the review date and the single change that would most reduce the manager's load. Ask the manager which of those changes they would make first, and make it before adding anything else to the plan. A framework earns its place only when it changes a diary, so judge it by the first calendar it alters.