Cost per unit falls as volume rises. Three mechanisms produce it, and they behave differently enough to be worth separating.

Fixed costs spread. The accounting system, the brand, the premises and the software cost about the same whether you serve ten clients or a hundred. Per client, they fall.

Specialization becomes possible. At small scale everyone does everything. At larger scale people do one thing and get better and faster at it, which is Adam Smith's pin factory and still the strongest of the three.

Purchasing power. Buying more of something usually lowers the price of each. Real, and generally the smallest effect for a service business.

Whether you have any

The important question, and one most owners answer optimistically.

The test is what happens to cost per unit when volume doubles. If the main input is skilled human time, the honest answer is often almost nothing. Two consultants cost twice one consultant. Serving twice the clients requires roughly twice the delivery hours, and the fixed costs being spread are a small share of the total.

This is why professional services businesses grow profit mostly by raising price rather than by adding volume, and why an agency that doubles headcount frequently finds margin unchanged or worse. The scale economies were assumed and were not there.

Where they genuinely exist in a service business, they come from turning time into something that is not time:

Productization. A defined offering delivered the same way each time has real fixed costs — the design of it — spread over every delivery.

Documented process. The training cost of the fifth person is lower than the second, because the process was written down in between.

Tooling. Anything built once and used repeatedly converts a variable cost into a fixed one, which is the only reliable route to scale economies in a people business.

The reversal

Past some point cost per unit starts rising again, and this is not a rare pathology. It is normal and it is why large organizations behave as they do.

Coordination. Communication paths grow faster than headcount. The tenth person adds more coordination load than the fourth, and eventually somebody is employed entirely to manage the fact that there are many people.

Agency costs. More distance between decision and consequence, which is the principal-agent problem scaling with the organization and being paid for in monitoring.

Loss of specificity. A large supplier serves the average customer and cannot afford to adapt to each one, which is the gap a small competitor lives in.

Taken together with diminishing returns, this is why the U-shaped cost curve exists and why "get bigger" is a strategy with a natural stopping point. The size at which the curve turns is different for every business, and knowing roughly where yours is worth more than any general view about growth.