Organization Design

Span of Control and the Arithmetic of Layers

Work out how many people one manager can genuinely hold, how many layers your headcount actually requires, and what each extra layer costs you in cash, delay and distortion — all from arithmetic you can do on one page.

  • Intermediate
  • 12 min total
  • 13 chapters

What decision this helps you make: Whether to add a manager, widen a span, remove a layer, or leave the structure alone and fix the thing that is really making the manager the bottleneck.

What this topic is

Span of control is the number of people reporting directly to one manager. Layers are the number of management levels between the front line and the top. The two are not separate design choices: for a given headcount they are locked together by arithmetic, because a company of a thousand people can be four layers deep with wide spans or six layers deep with narrow ones, and it cannot be both.

Why it matters

Layers are the most expensive thing in an organization that nobody prices. Each one adds salary, adds a relay that delays every decision in both directions, and adds a point at which information is summarised by someone with a view about how it should sound. Companies add layers one manager at a time, each addition individually defensible, and discover four years later that a decision takes six weeks and nobody can say why.

Who should learn it

Any manager deciding whether to promote someone into a lead role, any founder crossing the point where they cannot hold everyone directly, and anyone who has been asked to take cost out of an organization and does not want to do it by cutting the people who make the product.

What you will understand

  • The arithmetic tying headcount, span and layers together, and how to run it for your own business
  • The five factors that genuinely set the right span, and the two that people use instead
  • What a layer costs per year in salary, in delay, and in the fidelity of what reaches the top
  • When a narrow span is correct, and how to tell that from a manager who simply likes a small team

Prerequisites

Common misconception

"The right span is seven, plus or minus two." That number comes from a 1956 article summarising military practice and a 1930s combinatorial argument about relationships, and it has been repeated for seventy years as though it were a measurement. There is no universal right span. A team doing similar, well-standardised work with experienced people can be run at fifteen or twenty; a team of new joiners doing bespoke work with high interdependence may be genuinely hard at five. The second misconception is that flattening is a cost exercise. It is mostly a speed exercise — the salary saving is real but usually smaller than the decision-latency saving, which nobody measures.