06-reference

commoncog making sense of deming

Tue Apr 14 2026 20:00:00 GMT-0400 (Eastern Daylight Time) ·reference ·source: Commoncog ·by Cedric Chin

“Making Sense of Deming” — @CedricChin

Why this is in the vault

This is Chin’s comprehensive introduction to W. Edwards Deming — the intellectual grandfather of Statistical Process Control, and therefore the upstream source of everything RDCO is building around the MAC framework. Where 2026-04-15-commoncog-becoming-data-driven-first-principles gives you the operator-level thinking, this piece roots that thinking in Deming’s full System of Profound Knowledge and shows why the discipline extends beyond manufacturing into any business system. Filed as background canon — the “why this matters” at the worldview level.

Paywall note

The public version of this article is a ~2,550-word introduction that reaches Points One, Two, and Three of the 14 Points before cutting off at “So, if you can detect when variation changes, you can …” followed by a membership CTA. Points 4-14, the Psychology pillar detail, and Chin’s concluding assessment of Deming are members-only. The vault captures the public intro — which is the Deming 101 + System of Profound Knowledge overview and the first three (easy) points. The controversial psychology-pillar material (Deming on OKRs, performance reviews, incentives) is gated at a higher tier and not captured here.

The core argument (paraphrased)

Deming is better understood as a business philosopher than a statistician — he took variation as a first principle and derived an entire philosophy of operational excellence from it.

Chin’s framing: most readers meet Deming as the statistician behind SPC, but the deeper contribution is a coherent worldview that makes variation, systems, knowledge, and psychology interlock. Chin quotes Charles Koch in Kochland: “You never get out of this hospital. You are going to be working at this forever.” Continuous improvement is a lifelong posture, not a project.

The System of Profound Knowledge has four pillars:

  1. Understanding of Variation — the SPC foundation. Distinguish routine from exceptional variation.
  2. Theory of Knowledge — epistemology applied to business. “Management is prediction.” “There is no such thing as truth in business, only knowledge.” Without good epistemology, operators blindly copy instantiations of Deming and fail.
  3. Appreciation of a System — systems thinking. You can only improve a business if you understand how the processes interlock, including how employee psychology affects system design.
  4. Psychology — how humans find meaning and motivation in complex organisational systems. Chin calls this the source of Deming’s “most challenging ideas” — and steers mostly clear of it in the public piece.

From the Profound Knowledge system fall the 14 Points. Chin presents the first three in the public portion:

Chin’s meta-point: Deming didn’t give prescriptive mechanisms, he gave principles — and the proof of the principles is the real-world instantiations (Toyota Production System, Koch Industries, post-war Japanese industrial reinvention). Internalise the principles and you can bootstrap your own business operating system.

Mapping against Ray Data Co

This article sits one level upstream of 2026-04-15-commoncog-becoming-data-driven-first-principles — it’s the worldview from which RDCO’s operating thesis descends. Four mappings:

1. MAC is Point 3 (“cease dependence on inspection”) applied to data quality. The industry default for data quality is end-of-pipeline inspection: dashboards break, someone notices, someone investigates. MAC inverts this exactly the way Deming inverts manufacturing QA — bring variation under control across the system (column × row × aggregate × source × production × reconciliation × temporal × human), and the defect rate drops permanently. See ../01-projects/data-quality-framework/testing-matrix-template. The MAC 3×6 matrix is a Deming-style characterisation of variation, not a downstream inspection regime.

2. “Theory of Knowledge” is the state-ownership thesis. Deming’s epistemology — “management is prediction”, no truth only knowledge — is structurally what ../04-tooling/rdco-state-ownership-architecture argues about vaults and skills. The client owns the knowledge; the model is a commodity prediction engine. Every operator action is a prediction that the vault lets you score and update. This is Deming’s epistemology expressed in harness-architecture terms.

3. “Appreciation of a System” is why the agent-deployer role exists. Per 2026-04-14-levie-agent-deployer-role-jd, the agent-deployer is the person who sees how prompts, tools, data pipes, evals, and humans interlock. Deming would recognise this immediately as Pillar 3. An enterprise that deploys AI without a systems-thinker gets local optimisations (one agent works great) and global failure (the outputs don’t compose with the rest of the business). RDCO’s consulting posture is: we are the systems-thinkers you don’t have yet.

4. The phData/MG comparable is the Japanese-imports story of 2026. Chin’s Point 2 historical arc — American manufacturing got sloppy, Japan adopted SPC, Japanese imports flooded America — has a direct 2026 parallel: most enterprises have sloppy data discipline, a handful of firms (phData, MG, and RDCO) are adopting SPC-for-data discipline, and the output quality gap is about to become a market flood. We want to be teaching the discipline before the crisis, not selling remediation after.

One thing this article implicitly warns us about: Deming didn’t give prescriptive mechanisms — he gave principles — and Chin flags that naive operators fail because they copy the instantiations (Toyota cards, WBR format) without internalising the epistemology underneath. RDCO’s consulting curriculum has to teach the principles before handing over the skills, or clients will cargo-cult the MAC matrix without understanding why each cell is there. Bake that into the coaching pipeline.