It started with a problem.
Walter Shewhart, an engineer at Bell Telephone Laboratories, published Economic Control of Quality of Manufactured Product in 1931, introducing statistical process control (SPC) and the control chart. Shewhart argued that all manufacturing processes had natural variation, and that identifying and reducing variation — especially variation from assignable causes — was the key to consistent quality. W. Edwards Deming, a statistician who had worked with Shewhart, carried these methods to Japan after World War II through a series of lectures in 1950, where they were adopted by Japanese manufacturers. Deming's work contributed to what became known as Total Quality Management (TQM) and influenced Toyota's production system.
Historical sources below ↓Then something changed.
Quality management shifted the responsibility for quality from inspection after production to control during production. The old model found defects after they were made; the new model prevented them. This changed the economics: prevention was typically cheaper than correction, and consistent processes produced fewer defects than variable ones. Deming's Fourteen Points and his critique of management by fear became influential frameworks for continuous improvement in manufacturing and later in services. The quality movement produced ISO 9000, Six Sigma, and a range of certification frameworks that codified quality management practice.
What a builder can learn.
Our interpretation: quality management is the discipline of taking variation seriously. Any repeated process has natural variation in outcome — some outputs better than others. Statistical quality control provides tools to distinguish variation from random causes (acceptable) from variation from assignable causes (correctable). For a business builder, this is the underlying logic of process improvement: measure the output, identify where the variation is largest, find its source, correct it, and measure again.
A door still open.
A hypothesis worth testing: if a small business chose one important and repeated process — a product delivery, a service call, a fulfillment step — and measured the outcome quality across twenty consecutive instances, would it find that the variation was larger than expected, and that some instances of the cause were identifiable and correctable?
Which repeated process in your business produces the most inconsistent outcomes — and have you measured the variation or only noticed it anecdotally?
Follow the evidence.
History is sourced below. Business lessons are editorial interpretations; future possibilities are questions to test, not predictions.
Prepared from primary-source histories. First edition: 25 Sept 2026.