The System of Profound Knowledge Lens
Deming's System of Profound Knowledge has four interdependent parts: appreciation for a system, knowledge about variation, theory of knowledge, and psychology. He called these the lens through which managers should view their work.
The four parts are also a useful lens for diagnosing OKR practice. Each one names something about OKRs that most teaching ignores.
This section is a hypothesis. The mapping below has worked diagnostically in engagements but has not been formally tested.
Appreciation for a system
A system is a set of interdependent components working toward a common aim. Deming's point: components optimised in isolation degrade the system's overall performance, because the system's behaviour emerges from interactions, not from the individual parts.
OKRs operate at the team level. The system operates above the team level. Most OKR adoption ignores this gap. Teams are asked to write OKRs that improve their local performance, with no structured account of whether better local performance produces better system-level outcomes - or whether it actively degrades them.
The diagnostic question this lens asks: do your team OKRs add up to a coherent system, or are they a portfolio of local optimisations that no one is reconciling at the system level?
Three things to look for:
- Whether anyone at the system level is reading across team OKRs to see what is missing or duplicated
- Whether the system-level Objective (if there is one) decomposes meaningfully into team Objectives, or whether the relationship is decorative
- Whether teams are rewarded for hitting their own KRs even when doing so undermines another team's work
Where all three answers are bad, the OKR system is producing local-optimisation behaviour reliably. The artefact is not the cause. The absence of system-level appreciation is.
Knowledge about variation
Variation is the property of a process. Most measured outcomes vary from one period to the next. Deming's contribution was distinguishing two kinds of variation: common-cause (the inherent noise of the system) and special-cause (something specific happened that produced an unusual result).
The crucial implication: treating common-cause variation as if it were special-cause produces tampering. The system gets worse because someone has reacted to noise as if it were signal.
OKR practice routinely tampers. A KR misses its target by 15%. The team is asked to explain why. The team produces a narrative - usually a plausible-sounding one - that attributes the miss to a specific cause. Action is taken against that specific cause. The KR continues to vary in the next quarter, and the cycle repeats.
The diagnostic question: can your OKR system distinguish a meaningful miss from normal variation?
Most cannot. The reporting infrastructure does not support it. Targets are usually single numbers without confidence intervals. Comparison is to last quarter or to plan rather than to a control range. Every miss is a story, every story is a cause, every cause produces an intervention, and the system tampers itself sideways.
Where this is happening, the right response is not better OKR writing. It is statistical literacy at the leadership level - at minimum, recognising that not every miss is a signal, and that asking teams to explain noise produces fiction.
Theory of knowledge
A theory of knowledge is an account of how we come to know things. Deming insisted that managers operate with explicit theories, tested against evidence, rather than implicit assumptions never named.
OKRs encode a theory of how change happens. The implicit theory in most OKR rollouts is something like: if a team commits publicly to a measurable outcome and is held to account for it, the team will produce that outcome. This is a theory. It can be tested. The evidence is mostly against it - public commitment to measurable outcomes produces the outcomes only when several other conditions are also present, and those conditions are usually the constraints discussed in the previous section.
The diagnostic question: what theory of change is your OKR system implicitly running, and is it the one you would endorse if it were named?
This is one of the more uncomfortable questions in the playbook because the answer is often: we are running a theory of change we would not defend if it were named. The reasoning sounds like a parody of itself once written down - but the implicit theory survives precisely because it is never written down.
Naming the theory exposes it to scrutiny. Most organisations, asked to defend the theory of change underneath their OKR practice, would not defend it. The exposure is the first move toward improvement.
Psychology
Psychology, in Deming's usage, is the study of why humans behave as they do. He emphasised intrinsic motivation, dignity, fear, and the system-level conditions that determine whether humans do good work.
The dominant theory of motivation embedded in most OKR practice is extrinsic and consequentialist: people will do what they are measured against, especially if there are rewards and punishments attached to the measure. This is the theory underneath signal one - used as a stick.
The contemporary evidence on motivation does not support this theory for knowledge work. Daniel Pink's synthesis, drawing on decades of research, points to autonomy, mastery, and purpose as the primary motivators of complex performance. Extrinsic incentives narrow attention and reduce creative performance, particularly for tasks that require judgement.
The diagnostic question: what theory of motivation does your OKR practice assume, and is it the one the evidence supports?
Where the answer is mismatched, the OKR system is producing extrinsic-motivation effects in a context that calls for intrinsic motivation. People narrow their focus to the measurable, hide what they cannot measure cleanly, and avoid ambitious work because the cost of missing is higher than the reward of stretching. This is not a failure of the people. It is the system working as designed, under a theory of motivation that is wrong for the work.
Why these four
Deming's argument was that the four parts of profound knowledge are interdependent. You cannot fix a system by attending to one of them in isolation. A statistician who does not understand psychology will produce reports that demoralise teams. A leader who understands psychology but not variation will tamper with their best work. An expert in systems thinking without a theory of knowledge will propose interventions that cannot be tested.
The same applies here. OKR practice that addresses one of the four - perhaps better statistical literacy, or better motivation theory - without addressing the others will produce partial improvement at best. The four parts of the lens have to be applied together.
This is the home-base argument for why OKRs belong in a Profound treatment. They are not a tool, and they are not a framework. They are an instrument of management - and management, in Deming's account, is the act of designing systems within which humans can do their best work. OKR practice is one place that design becomes visible. Reading it carefully is reading the management.
Last updated 15 May 2026
Licensed under CC BY-SA 4.0.