Pull the numbers and your company-wide span of control comes back at 6.2 direct reports per manager. Industry average for a professional services firm your size is 6. On paper, you're fine. Decisions still take three weeks, two reorgs in the last two years haven't fixed it, and nobody can explain why a $40K vendor contract needs four signatures.
The average isn't lying to you. It's just answering a different question than the one you're actually asking. A company-wide span of control analysis tells you whether your org is over-managed in aggregate. It says nothing about whether it's over-managed in the specific three departments where a VP is protecting headcount, a director inherited a team from an acquisition and never rationalized it, or a manager has two direct reports because that's what the role paid in 2019 and nobody's revisited it since. Average out a function with 11 reports per manager against one with 3, and the 6.2 that comes out the other side is true and useless at the same time.
What a Real Span of Control Analysis Finds That the Average Hides
The number that matters is the distribution, not the mean. A thin-span manager -- fewer than 5 direct reports -- is a specific, findable person: a fixed cost of a full salary, benefits load, and a layer of approval latency, in exchange for supervising a team that doesn't need full-time supervision. The output of a spans-and-layers analysis worth paying for is a list of names and functions. If what comes back is a ratio, you bought a thermometer, not a diagnosis.
Performis's benchmark data across service and industrial verticals puts the pattern in relief: professional services firms average 6 direct reports per manager, with top-quartile operators at 9 and bottom-quartile at 4; manufacturing averages 7, with top quartile at 11; SaaS organizations average 6, with a bottom quartile of 4; distribution averages 8, with top quartile at 12. Notice what that does to the 6.2 we started with. In professional services it's dead average. In a distribution business, where the average is 8 and the bottom quartile is 5, the same 6.2 is closer to bottom-quartile than to par. "We benchmarked our span" is not a finding until someone says benchmarked against what.
Then there's the error I'd argue is more expensive than the averaging one: applying a single target span to the entire org chart. Operational and frontline management should be running 6 to 10 direct reports. Executive spans of 5 to 7 are structurally normal at the C-suite, because the work at that level is judgment, not supervision. A CFO with five direct reports is not a finding. A warehouse supervisor with five is. Decks that run one threshold top to bottom generate a list of false positives at the top of the chart -- where consolidation is politically hardest, most disruptive, and least valuable -- while systematically understating the cost sitting two and three levels down, where the roles are cheaper individually and far more numerous.
Layers compound the problem in a way spans alone don't show, and they benchmark on a different axis than most people assume. Management layers track company size far more than industry: under 150 employees, the benchmark is 4 layers from CEO to frontline with top-quartile operators at 3 and bottom quartile at 6; at 150-500 employees it's 5, with a bottom quartile of 7; at 500+ it's 6, with a bottom quartile of 9. Benchmarking your layer count against industry peers instead of size peers is a category error that will tell a 120-person firm it's fine when it's carrying two layers it can't justify. And the cost isn't linear -- a company running 7 layers where 5 is the benchmark carries roughly twice the management overhead of a benchmark-equivalent org, because every extra layer brings its own thin-span managers, its own approval step, and its own translation loss between what leadership decided and what the frontline actually does. McKinsey's research on this (industry-sourced, not a Performis-proprietary number) puts the reduction in managerial costs at 10-15% for companies that go through and actually correct span of control -- and that's before counting the decision-velocity gains that don't show up on a payroll report at all.
What This Actually Costs, in Dollars
Here's the arithmetic that should be driving the priority list, not the average. Take a $25M revenue services business with payroll running at 65% of revenue -- above benchmark, which is usually the first sign something structural is off. Identify the three managers whose span is thin enough that consolidating their teams under one director is a real option, not a morale problem waiting to happen, and eliminating those three roles alone saves $450,000 to $600,000 in annual payroll. Scale that pattern across a 100-300 person organization and the realistic range is $300,000 to $1,500,000 in annual savings, with span improving from a thin 4-5 to a healthy 7-8 and total EBITDA margin gaining 2 to 5 points -- without cutting a single frontline role or touching the work that actually generates revenue.
Two caveats a good operator will raise before signing off on that, and both are worth raising yourself. First, three thin-span managers is not automatically three people out the door. Often the correct move is reclassifying the manager as a senior individual contributor: you keep the expertise, you keep the person, and you still remove the layer and the approval step. The layer is the cost; the human is frequently not. Second, structural consolidation is only one of the levers. Correcting compensation that overpays for average performance while top performers go underpaid is worth 1 to 2 points of EBITDA margin on its own with no headcount change at all -- which means if a spans analysis is the only thing on the table, the cheaper, less disruptive money may be getting left behind.
Timing matters for how you frame the ask. A sized diagnostic and action plan is a 30-to-45-day exercise; the structural changes themselves are implementable inside 60 to 90 days; the EBITDA impact is visible at the next quarterly close. That's the number a board or a PE operating partner should be asking for before a spans-and-layers effort gets funded: not "what's our average span," but "which specific managers, worth how much, by which quarter." A diagnostic that can't answer that in dollars and dates hasn't actually found the problem yet.
Why the One-Time Readout Doesn't Hold
A consulting engagement that produces a spans-and-layers deck once, at the start of an engagement, is diagnosing a moving target with a single photograph. The org chart the deck describes starts drifting the moment the deck is presented.
And here is the part that makes this metric genuinely treacherous, which I have never seen put on a slide: span of control is calculated as non-manager headcount divided by people-manager headcount. Create a "coordinator" or "team lead" role that isn't coded as a manager in the HRIS but functions as one in practice, and the arithmetic moves twice in your favor -- that person counts in the numerator as an individual contributor, and never enters the denominator as a manager. Your reported span goes up. Your reported layer count doesn't move. Meanwhile you have added a real supervisory relationship and a real approval step. The metric improves while the organization gets slower. Nobody has to be acting in bad faith for this to happen; it is simply the path of least resistance for a leader who has been told to raise their span and still needs someone to run the day-to-day.
Consider a composite pattern, illustrative of what shows up repeatedly in this kind of engagement rather than any single client: an external consultant's spans-and-layers review correctly flags two thin-span managers in Operations for consolidation. Leadership agrees, the org chart is redrawn, and the deck goes in a drawer. Fourteen months later, headcount in that same function has crept back to nearly its original size -- not through a formal reorg, but through a series of individually reasonable-looking hires, plus two new "coordinator" titles that don't register as management anywhere in the system. Nobody checked any of it against the original span target, because nobody was still tracking it. The finding wasn't wrong. It just wasn't being watched, and the reported ratio spent those fourteen months getting better while the actual structure got worse.
This is exactly the gap that real analytics software in this category -- tools like Nakisa, Crunchr, and ingentis -- already closes better than a static PowerPoint does: they track spans and layers continuously, not once. Where the gap still sits, for a mid-market operator without a dedicated people-analytics function running one of those platforms, is tying that ongoing structural read to the company's own EBITDA target and value-creation thesis, the way a value creation plan works backward from the required exit number rather than from a generic checklist. Continuous span tracking that reports a ratio is useful. Continuous span tracking that reports "these three roles hold $450K to $600K of annual payroll, consolidation lands by Q3, and here is who owns it" is the version that actually gets acted on.
What Good Looks Like
A structural design engagement done right doesn't stop at the average, and it doesn't stop at the handoff meeting. It names the specific thin-span managers by function, uses the right threshold for their level rather than one number for everyone, states the EBITDA dollar impact of consolidating each one, checks the layer count against the benchmark for your headcount band rather than your industry, and -- because org charts drift the same way pricing and cost structures do -- keeps checking it quarterly rather than filing the finding away until the next engagement gets funded.
One more thing decides whether any of this survives contact with the next hiring cycle: who owns the number. When span of control lives with HR, it is a reporting metric and it decays. When it sits on the same quarterly review as the EBITDA bridge, owned by whoever owns the margin, it behaves like a financial control -- because that is what it is. Align the review cadence to the close, not to the performance-review calendar. That's the same discipline an agentic approach applies elsewhere in the business: not a report a leadership team reads once and reacts to, but a standing read of a number that moves on its own schedule whether anyone's watching or not.
Performis's Org Design work builds the structural design finding this way -- span of control and management layer count benchmarked against your industry and size, thin-span managers identified by function with the EBITDA impact attached to each, and the finding tracked quarterly rather than diagnosed once and left to drift.
If your span of control looks fine on the company-wide average and you still can't explain why decisions take three weeks, start an Org Design engagement and find out what the average is hiding.