Who controls what, and what they are judged on

Why decentralise

Boreal Bikes Ltd. started as one factory and one founder who priced every bike. It now has two divisions, a dozen stores, an online team and a plant, and the founder cannot make every decision. Decentralisation pushes decision rights down to the managers closest to the information. Its benefits: decisions are made faster; they use local knowledge head office does not have; divisional managers are trained as future executives by running something real; and head office is freed to think about strategy. Its costs: duplication (two divisions, two marketing teams); goal incongruence, where a manager’s best choice for her unit is not the company’s best choice; and dysfunctional competition between units that should cooperate.

The device that makes decentralisation manageable is the responsibility centre: a unit whose manager is accountable for a defined set of results — and, crucially, only for what she controls. Five types, defined by what the manager controls.

Boreal Bikes’ organisation, classified

Investment centres — the Road and Trail divisions. Each divisional manager sets prices, chooses the product range, hires, and decides on capital spending: new tooling, a new store. She controls revenue, costs and the assets employed, so she is judged on the return those assets earn (ROI, residual income — Lesson 11.6).

Profit centres — the retail stores. A store manager controls what the store sells and at what local promotions, and controls store wages and supplies. She does not choose the building or the lease; head office does. She controls revenue and costs but not the investment, so she is judged on controllable profit.

Revenue centre — the online sales team. The team is judged on revenue against target, with a budgeted cost allowance for the platform and the staff it is given. It has costs — every unit does — but its decisions move revenue, not cost, so revenue is the measure.

Cost centre — the frame-assembly plant. The plant does not sell anything; the stores order frames and the plant builds them. Its manager controls labour, materials and overhead, and is judged on cost per frame at the output the stores ordered — never on a profit she cannot influence.

Discretionary cost centres — R&D and HR. These also spend without selling, but with a difference: the assembly plant’s output can be counted and costed per frame, while R&D’s output (next year’s designs) and HR’s (a workforce that stays) cannot be tied to the spending by any formula. A discretionary cost centre is one whose output cannot be measured against its input, so it is judged on staying within budget and on qualitative results — a design shipped, a turnover rate — not on cost per unit of anything.

Type                     Manager controls                 Judged on
Cost centre              costs, at a given output         cost vs flexible budget at actual output
Discretionary cost       spending, output unmeasurable    budget adherence + qualitative results
Revenue centre           revenue (cost allowance given)   revenue vs budget, sales mix
Profit centre            revenue and costs                controllable profit (segment margin)
Investment centre        revenue, costs and assets        ROI, residual income

Goal congruence

The purpose of the whole structure is goal congruence: designing each centre’s measure so that the choice that is best for the manager is also best for Boreal. It does not mean every manager has the same goal; it means their different goals add up. When the measure is wrong — a plant judged on profit, a division judged on a ratio it can improve by shrinking — the structure produces the incongruence it was built to prevent. Lessons 11.6 and 11.7 are about those failures.