Deming's critique of institutional incentive structures is among his most radical and least understood contributions. Where conventional management theory assumes that the right incentives will produce the right behavior, Deming argued the reverse: that most incentive systems are not merely ineffective but actively destructive, because they treat system problems as individual problems and thereby guarantee that the system problems will never be addressed.
The foundation of this critique is what Deming called the 94/6 rule (sometimes stated as 96/4): approximately 94% of problems in any organization are caused by the system, and only 6% by individual workers. This follows from the distinction between common-cause-vs-special-cause-variation. Common-cause variation is built into the system — it is the product of how processes are designed, how materials are sourced, how departments interact, how information flows. Special-cause variation is attributable to specific, identifiable factors. When management applies incentives — rewards or punishments — to individuals working within a system dominated by common-cause variation, it is asking workers to overcome forces they did not create and cannot control. The result is frustration, gaming, and distortion, not improvement.
The the-red-bead-experiment is Deming's most vivid demonstration of this principle. In the experiment, workers use a paddle to scoop beads from a bowl containing both white beads (acceptable) and red beads (defective). No matter how motivated, skilled, or incentivized the workers are, the proportion of red beads they produce is determined by the system — the ratio of red to white beads in the bowl. Management in the experiment rewards the "best" workers and punishes the "worst," but the variation in their output is entirely random. The experiment makes visible what is invisible in most organizations: that performance differences attributed to individuals are often system artifacts. Incentives applied to such differences are worse than useless — they reward luck and punish luck, creating cynicism and undermining trust.
Several of the the-14-points-for-management address institutional incentives directly. Point 11 calls for eliminating work standards (quotas) on the factory floor and management by objectives and numerical goals. Point 12 calls for removing barriers that rob workers of their right to pride of workmanship, explicitly naming "the annual rating or merit system." Point 3 calls for ceasing dependence on inspection — which is itself an incentive structure that tells workers "produce volume; we'll catch the defects later." Together, these points constitute a systematic dismantling of the incentive architecture that most American organizations take for granted.
The seven-deadly-diseases extend this critique to the organizational level. "Evaluation by performance, merit rating, or annual review" is listed as a deadly disease — not a practice in need of reform but a pathology requiring elimination. Deming argued that performance reviews destroy teamwork (because they pit individuals against each other), encourage short-term thinking (because they evaluate performance over fixed periods), and attribute to individuals outcomes that are produced by the system. "Management by use only of visible figures" is another deadly disease — the assumption that what can be measured and incentivized captures what matters.
The connection to management-by-objectives-deming-s-critique is direct. MBO, popularized by Peter Drucker, assumes that setting targets and holding people accountable to them will improve performance. Deming saw this as a fundamental misunderstanding of systems. Targets set without a method for achieving them — without changing the system — produce only distortion. Workers learn to game the metrics, shift costs to unmeasured activities, or simply fabricate results. The target becomes the enemy of improvement, because meeting the target is treated as success even when the system has not changed. Deming's alternative is to improve the system, which will improve results — the reverse of the MBO logic.
Deming's critique connects to appreciation-for-a-system in a fundamental way. If most problems are system problems, then the appropriate response to poor performance is not to incentivize individuals differently but to study and improve the system. This requires management to take responsibility for the system — which is precisely what the incentive approach allows management to avoid. By holding individuals accountable for system outcomes, management can disclaim responsibility for the system itself. Deming saw this as a moral failure, not just a technical one: management creates the system, management alone has the authority to change the system, and therefore management bears responsibility for the system's results.
The implications reach far beyond manufacturing. In education, standardized testing and merit pay for teachers apply individual incentives to system outcomes. In healthcare, pay-for-performance schemes incentivize metric manipulation rather than patient care improvement. In software development, velocity targets and bug-count metrics create the same distortions Deming identified on the factory floor. In every case, the pattern is the same: individual incentives applied to system problems produce gaming, cynicism, and degraded performance — exactly the outcomes Deming predicted from out-of-the-crisis onward.