Business & Revenue OpsFree
Span of Control Analysis Template
Send the HR export, get every manager's span beside the number of distinct kinds of work under it, and a proposed chart that reconciles.
Search this and you get a ratio. One manager to seven people, one to ten, one to fifteen depending on which consultancy wrote the page, and a spreadsheet that divides headcount by managers and colours the outliers red. The ratio is doing all the work, and it is the one part of the analysis that cannot be right. It scores a manager of twenty-two cleaners on one shift at one site as the same job as a manager of six specialists across four counties.
The federal government's own position-classification guide settles this. It grades a supervisory job by scoring the work, and it awards credit for more than one kind of work, each requiring a distinctly different additional body of knowledge, and separately for shift operations, a fluctuating workforce and physical dispersion. Not one of those is a headcount. So every manager here carries a second number beside the span: how many distinct role families sit under them. On a real chart the two numbers flag different people.
Run it on a 226-person facilities contractor and the arithmetic is blunt. A twelve-report benchmark flags six managers, and all six run a single trade at a single site. The role-family count flags seven others with four to eight reports each, and the two sets do not overlap by one name. Then the output is a chart rather than a memo: the proposed structure with every one of the 226 people in it exactly once, so the ownership matrix and the process map still reconcile afterwards.
The two structures a ratio cannot see
A manager with one report is not a narrow span, it is a layer. The same classification guide will not grade supervisory duties at all unless they occupy at least twenty-five percent of the position's time, and it discounts the grade where the supervisory work does not sit a level above the work supervised. Both tests catch the same thing: a job that exists to relay. The contractor had one, a director with no individual contributors, one manager reporting in, and five people below.
The second structure is depth measured per function rather than averaged for the company. Marchmont runs two layers deep in Sales and six in Field Operations, a spread of four inside one 226-person business. A decision of the same weight therefore gets made at very different distances from the customer depending on which door it arrives through. The mean across the five delivery functions is 3.2 layers. That number describes no function here and would send the redesign at the wrong one.
A diagnosis nobody can act on is a diagnosis nobody acts on, so the proposal has to balance. Every person in the current chart appears exactly once in the proposed one, and the diff reads as a reconciliation: one manager role removed, two created, eleven people changing manager, no redundancies and no new hires. Eleven of 226 is the change-management bill, and it is also the list the procedure gap audit has to re-check once the moves land.
How it works
Send the roster
An HR export or a typed-out chart. Each person and their manager is the minimum.
Describe the work
Enough to tell roles apart by hiring pool rather than job title, because titles inflate.
Get both counts
Span and role families side by side, plus pass-through layers and depth per function.
Take the chart
The proposed structure drawn and balanced, with the number of people who change manager.
What you get
- Every manager in one sheet with raw span, manager reports, individual contributors and depth
- A role-family count beside every span, so two identical headcounts stop looking identical
- Sites and shifts per manager, both of which add supervisory load without adding a head
- Pass-through layers named: no individual contributors, under three manager reports, nothing added to scope
- Layer depth measured per function rather than averaged, because the average describes nobody
- The proposed structure as a diagram, with every person in the current chart placed exactly once
- A reconciliation on the proposal: roles removed, roles created, people changing manager, hires required
Common questions
What do I need to send?
An HR export or an org chart with each person's manager, plus enough about what each team does to tell the roles apart. Job titles usually carry it. Sites and shifts help, because both add supervisory load without adding headcount. Without the work descriptions you get span and layer counts and nothing else.
Is there a right number of direct reports?
Not one that transfers. A site lead running twenty-two cleaners on one trade at one address is a smaller job than a supervisor running six specialists over four counties and two shifts. The federal classification guide grades those differently and never counts heads to do it, which is the whole argument against the benchmark.
How do you decide what counts as a distinct role family?
By qualification, not by title. Two people are in the same family if you would recruit them from the same pool and a manager needs one body of knowledge to supervise both. That is close to how the federal series is defined, as positions similar as to specialized line of work and qualification requirements. Titles inflate; hiring pools do not.
What makes a layer a pass-through?
No individual contributors, fewer than three managers reporting in, and a scope that is the sum of theirs with nothing added. Marchmont had exactly one: a director with a single manager reporting to them and five people below. Deleting it shortened Field Operations from six layers to five and cost nobody a job.
Will this tell me to make people redundant?
It does not assume it. The proposed chart conserves headcount by default, so a removed manager role is a person who moves rather than a person who leaves, and the reconciliation says which. In the worked example one director role went, two were created, and both were filled internally. Ask for the other version and it gets costed.
What do I do with the diagram once I have it?
Take it to the conversation, because a table of outliers loses to a picture every time. Then fix what the move broke: the responsibility matrix has stale names in it, the handover notes need writing for eleven people, and the cost model needs the new manager roles in it.
Does layer depth matter as much as span?
More, in a company that grew by function. Marchmont ran two layers deep in Sales and six in Field Operations, so the average of 3.2 described nothing. Depth is also what makes escalation slow, which is why the escalation matrix gets shorter when a layer goes.
Span of Control Analysis Template
Fill in the form and your workspace opens with the work already underway.