Management Model Coherence

Definition

Budgeting is not one isolated process but a coherent, self-reinforcing set of practices — target-setting, rewards, planning, resourcing, coordination, and control — all built on the same underlying assumption of central control and low trust. Introducing a single new practice, like a stretch target or a rolling forecast, into that set without changing the others doesn't improve the system; it "ruptures the coherence" and produces worse outcomes than leaving the old model alone, because the surviving old practices pull against the new one.

In the Book

Chapter 2 states the principle directly: "The budgeting model, for all its flaws, is coherent. All its principles and practices assume central control. Therefore, introducing foreign bodies such as relative improvement contracts that assume self-regulation will be rejected." Chapter 5 grounds this in two failed change efforts. A large consumer products group told managers to set "stretch forecasts" without being held accountable for them, while leaving its plan-make-and-sell business model and volume-based sales incentives untouched — managers duly inflated their forecasts to please the chairman, production and inventory scaled up to match, and the company took large write-offs when the numbers proved fantasy: "empowerment without accountability... guaranteed to fail." A large telecommunications company tried to modernize only its finance department's planning process, in isolation from the fixed performance contracts and incentive structure still governing every operating manager, and made no lasting progress until the authors' core lesson landed: rewards, targets, planning, and control must be realigned together, or not at all.

Why It Matters

This is a general warning against piecemeal reform of any system whose parts were designed to reinforce one another: swapping in one modern-sounding practice — an OKR framework, a new performance review cycle, an agile ceremony — onto an incentive structure and culture still built for the old regime doesn't average the two into something better, it can produce a worse hybrid than either pure model, because the old parts actively punish the behavior the new part is trying to reward.