Six Companies After Near Zero

Every operator hears the argument and asks the same question: fine, but what does that mean for a company like mine? These are six answers — one per industry, each a 150-person business, each modeled the same way: what the coordination layer costs today, what happens to profit when it goes, which work the machine actually takes, and what the sphere looks like on the other side.

What the coordination layer actually costs you

Start with money, because headcount charts don't move anybody and this does.

Take the coordination people out of each of these six companies — the CSRs, the dispatchers, the schedulers, the order-entry clerks, the billing staff, the people whose job is moving information between other people and systems — and price them at what they actually cost: salary plus payroll tax, benefits, workspace and equipment. Then set that number beside what the company earns in a year.

Industry (150 people)	Annual profit	Cost of the coordination layer	As % of profit
Freight brokerage / 3PL	$3.0M	$6.88M	229%
Staffing & recruiting	$2.6M	$4.70M	184%
Wholesale distribution	$3.0M	$4.93M	164%
Residential home services	$2.1M	$2.67M	127%
Multi-site healthcare group	$4.0M	$3.74M	92%
Precision manufacturing	$3.4M	$2.34M	70%

Modeled. Profit uses typical margins for each industry at this size; coordination cost is headcount × blended salary × 1.30 for payroll tax, benefits, space and equipment.

That is not an efficiency observation. It's a statement about where the earnings of most mid-sized American companies actually go — and it requires no exotic structure to see. You don't need a franchise model, an agent network, or a Silicon Valley org chart. You need your own payroll register and your own P&L.

What happens to profit

Now run the models forward. Remove roughly three-quarters of the coordination layer — the amount that consistently disappears across every industry here — and subtract a generous annual budget for the software and agents that replace it.

Industry	Profit today	Net annual gain	Profit after	Change	Enterprise value created
Freight brokerage / 3PL	$3.0M	+$5.1M	$8.1M	+171%	$36M
Staffing & recruiting	$2.6M	+$3.6M	$6.2M	+142%	$22M
Wholesale distribution	$3.0M	+$3.3M	$6.3M	+109%	$20M
Residential home services	$2.1M	+$1.9M	$4.0M	+90%	$11M
Multi-site healthcare group	$4.0M	+$2.3M	$6.4M	+57%	$18M
Precision manufacturing	$3.4M	+$1.4M	$4.8M	+41%	$8M

Modeled. Net gain is coordination payroll removed minus an annual software/agent budget of $250K–$600K depending on industry. Enterprise value uses conservative current multiples for each sector (6–8× EBITDA), applied only to the incremental earnings.

The home-services number is the one worth sitting with, because it's the least dramatic on the page and still ends the argument. A $30M contractor earning $2.1M removes 33 office roles, spends a quarter-million a year on software to replace them, and roughly doubles its net income — without touching a single technician, without selling more, without raising a price. At a normal multiple, that owner just created about $11 million of enterprise value out of work nobody enjoyed doing.

And it compounds twice. The earnings improve, and the multiple those earnings trade at improves too — because a company that runs on structure rather than on twenty people's institutional memory is worth more to an acquirer than one that doesn't. Nobody pays a premium for a business that only works because Debbie knows where everything is.

The one number that predicts everything

Running six industries through the same exercise surfaced a pattern I did not expect to be this clean.

The size of a company's transformation is almost entirely predicted by one ratio: what share of its headcount is coordination work. Not how technical the industry is. Not how sophisticated its software is. Not whether the owner is an early adopter. Just: how many of these people exist to move information between other people and systems?

Across all six models, roughly three-quarters to four-fifths of pure coordination headcount disappears — and that ratio holds whether the company is a freight brokerage or a machine shop. What differs wildly is how much coordination each industry had to begin with.

Industry	Coordination share today	Modeled headcount change	Office/admin change
Freight brokerage / 3PL	77%	150 → 52	−72%
Staffing & recruiting	57%	150 → 66	−73%
Wholesale distribution	53%	150 → 88	−75%
Multi-site healthcare group	51%	150 → 96	−70%
Residential home services	27%	150 → 117	−80%
Precision manufacturing	24%	150 → 119	−72%

Read the last two columns together, because that's the whole finding. The machine shop and the freight broker lose the same proportion of their back office — around three-quarters. They just have wildly different amounts of back office to lose.

The benchmark that proves it's structural, not technological

Before any of the models, look at two companies in the same industry, doing the same thing, today:

Company	Industry	Employees	Revenue	Per employee
Landstar System	Freight & logistics	1,378	$5.00B	$3.63M
C.H. Robinson	Freight & logistics	11,855	$17.0B	$1.43M
Hub Group	Freight & logistics	6,604	$3.73B	$565K
GEE Group	Staffing	173	$84M	$487K
BGSF	Staffing	189	$92M	$487K
Comfort Systems USA	Mechanical contracting	22,700	$11.2B	$495K
Limbach Holdings	Mechanical contracting	1,600	$684M	$427K
DXP Enterprises	Industrial distribution	3,286	$2.14B	$651K
Fastenal	Industrial distribution	24,489	$8.75B	$357K
Espey Mfg.	Electronics manufacturing	152	$42M	$278K
The Joint Corp	Multi-site clinics	330	$59M	$177K

Source: current public filings via stockanalysis.com, retrieved August 2026. Espey (152 employees) and GEE Group (173) sit inside the target size band and anchor the manufacturing and staffing models directly.

Landstar produces 2.5 times the revenue per employee of C.H. Robinson and 6.4 times Hub Group's — in the same industry, before any of this. The difference isn't better people or better AI. Landstar pushes coordination out to an agent network and owner-operators instead of carrying it on payroll. The spread between these companies is a structural spread, and it existed before the machines showed up.

Which is the argument in miniature: near zero doesn't create the possibility of extreme operating leverage. It makes that leverage available to companies that never had a network to push the work onto — because now they can push it onto software instead.

Five things true in all six

• Nobody who touches a physical thing loses their job. Not one technician, picker, driver, machine operator, hygienist or install crew member. Every single reduction in these six models comes from the coordination layer. That is worth saying out loud in every room where this argument gets made, because it is the opposite of what people assume.
• The customer-facing function grows. In four of six models, the revenue-producing relational roles go up — more account managers, more field sales, more comfort advisors. When execution is free, the constraint moves to relationships.
• Every industry converges on the same six or seven struts. Demand, customer, capacity, supply, money, people, systems. The names differ; the geometry doesn't. That is the size-invariance claim showing up in real org charts.
• Roughly three-quarters of coordination headcount goes, everywhere. 80% in home services, 83% in brokerage, 86% in staffing, 75% in distribution, 72% in manufacturing, 71% in healthcare. The consistency across wildly different industries is the strongest evidence that this is structural rather than sector-specific.
• The prize isn't always payroll. In manufacturing it's quote volume and machine utilization. In home services it's the ability to add trucks without adding desks. In healthcare it's an hour a day back per provider. Companies that measure this transformation only in headcount will underestimate it — and mistreat their people getting there.