Eliminate Numerical Quotas

Transfers

Deming’s Points 11a and 11b call for the elimination of work standards (quotas) on the factory floor and the elimination of management by objective and management by numbers. The core insight: quotas substitute a number for understanding. They tell workers and managers what outcome to produce without addressing the system’s capability to produce it.

Key structural parallels:

• Quotas are ceilings disguised as floors — a production quota of 200 units per day is presented as a minimum standard, but in practice it functions as a ceiling. Workers who can produce 250 slow down to avoid raising the bar. Workers who struggle to reach 200 cut corners on quality. The quota converges everyone toward the average, eliminating both excellence and the information that variation would provide about the system. This dynamic appears wherever numerical targets are imposed: lines of code per day, tickets closed per sprint, papers published per year.
• The number becomes the goal — Deming’s critique anticipates what later became known as Goodhart’s Law: when a measure becomes a target, it ceases to be a good measure. Workers and managers who are evaluated on a number will optimize for the number rather than for the quality it was meant to represent. Call center agents measured on calls per hour will hang up on difficult cases. Software teams measured on velocity will inflate story points. Schools measured on test scores will teach to the test. The underlying process degrades while the metric improves.
• Measurement without theory is noise — Deming argued that numbers are meaningful only in the context of a theory about the process that generates them. A defect rate of 3% means nothing unless you understand the sources of variation in the process. Quotas strip away the theoretical context and present the number as self-explanatory. This is the epistemological core of the critique: management by numbers is management without understanding.
• Process capability versus arbitrary targets — Deming introduced the distinction between what a system is capable of producing (its statistical capability) and what management demands it produce. If the demand exceeds the capability, the only options are to falsify data, cut quality, or fail. Improving the capability requires understanding the system — which quotas actively discourage because they focus attention on the output rather than the process.

Limits

• Not all numerical targets are quotas — Deming’s critique targets arbitrary numerical standards imposed without process understanding. But some numbers represent genuine constraints: a delivery deadline, a regulatory threshold, a budget limit. These are not arbitrary quotas but real boundaries that must be respected. The model does not clearly distinguish imposed quotas from genuine constraints.
• Some domains need output targets for coordination — in complex organizations with many interdependent teams, quantitative targets serve as coordination mechanisms. If the logistics team does not know how many units manufacturing will produce, they cannot plan shipping. Deming’s manufacturing context often had simpler coordination needs than modern knowledge-work organizations.
• Removing quotas without replacing them leaves a vacuum — Deming prescribed replacing quotas with leadership and process understanding. But this requires statistical literacy and management skill that many organizations lack. In practice, eliminating quotas without building the capability to manage by process can produce aimlessness or even worse outcomes than the quotas did.
• The Goodhart dynamic is not inevitable — while measurement corruption is real and common, some organizations maintain healthy relationships with metrics by using them as diagnostics rather than evaluations. OKRs in well-run organizations, for example, are explicitly aspirational and divorced from compensation. The model can lead to an overcorrection where all measurement is viewed as toxic.
• Individual accountability still matters — the model’s emphasis on system-caused variation can be used to deflect accountability for genuinely individual failures. A developer who consistently introduces bugs is not solely a system problem. The model provides no framework for distinguishing system variation from individual performance issues in specific cases.

Expressions

• “You get what you measure” — the folk version of Goodhart’s Law, encoding both the power and the corruption of measurement
• “Making the numbers” — corporate shorthand for achieving quotas, revealing that the number itself has become the objective
• “Teaching to the test” — education’s version of quota corruption, where test scores replace learning as the goal
• “Velocity” in agile teams — a process diagnostic that frequently degrades into a quota, with teams gaming story points to hit targets
• “Stack ranking” — forced distribution of performance ratings, a numerical quota applied to human evaluation
• “Management by spreadsheet” — pejorative describing organizations that substitute numerical reporting for process understanding
• “Hitting quota” in sales — the paradigmatic example of numerical targets shaping behavior, for better and worse

Origin Story

Points 11a and 11b of Deming’s 14 Points (published in Out of the Crisis, 1986) target two distinct but related practices: work standards (quotas) on the factory floor (11a) and management by objective / management by numbers for management (11b). Deming argued that both substitute a number for understanding and guarantee that the system will never improve beyond its current capability.

Deming’s critique was grounded in Walter Shewhart’s statistical process control framework. Shewhart had shown that every process has a natural capability — a range of outcomes determined by the system’s design. Asking for outcomes outside this range without changing the system is asking for either miracles or fraud. Deming extended this insight from the factory floor to management practice, arguing that MBO (Management by Objectives, popularized by Peter Drucker) committed the same error at the executive level: setting numerical targets without understanding or improving the processes that generate the numbers.

The critique gained renewed relevance in software engineering with the adoption of velocity as an agile metric. Originally intended as a planning tool (how much work can this team typically complete in a sprint?), velocity was widely corrupted into a performance quota, producing exactly the dynamics Deming predicted: point inflation, quality sacrifice, and adversarial relationships between teams and management.

References

• Deming, W. Edwards. Out of the Crisis (1986), pp. 70-75
• Deming, W. Edwards. The New Economics for Industry, Government, Education (1993), Chapter 2
• Goodhart, Charles. “Problems of Monetary Management: The U.K. Experience” (1975) — the measurement corruption that Deming anticipated
• Drucker, Peter. The Practice of Management (1954) — MBO as Deming’s foil
• ASQ. “Deming’s 14 Points for Total Quality Management.” https://asq.org/quality-resources/tqm/deming-points