Loading field map
What does it mean to improve a process? The answer has shifted dramatically over the past century. For some, improvement means standardizing the one best way to work. For others, it means reducing variation, eliminating waste, focusing on a bottleneck, or redesigning the process entirely. Each era of process improvement has produced a framework that redefined the very concept of improvement—sometimes incrementally, sometimes by radical departure.
The first systematic approach to process improvement emerged with Scientific Management (1900–1930). Frederick Taylor and his followers used time-and-motion studies to break down work into its smallest elements, identify the “one best way” to perform each task, and enforce that standard through strict supervision. The central problem was inefficiency caused by informal, rule-of-thumb methods. Taylor’s solution was top-down standardization, separating planning from execution. Improvement, in this view, was a one-time event: set the standard and enforce it.
But even as Scientific Management spread, a different logic was taking shape in industrial laboratories. Statistical Quality Control (SQC, 1920–1950), pioneered by Walter Shewhart at Bell Labs, introduced control charts and the distinction between common-cause and special-cause variation. Where Taylor assumed that a standard could be set once and left alone, SQC recognized that all processes vary and that improvement requires ongoing monitoring and adjustment. SQC did not replace Scientific Management; it added a statistical lens that revealed when a process was stable (and thus ready for standardization) and when it was out of control. The two frameworks coexisted in practice: Scientific Management provided the standard work methods, and SQC provided the feedback loop to detect when those methods drifted.
The mid-20th century saw a cultural expansion of improvement thinking. Total Quality Management (TQM, 1950–1990)—advanced by W. Edwards Deming, Joseph Juran, and others—broadened the improvement mandate from the factory floor to every department. TQM absorbed SQC’s statistical tools but placed them within a philosophy of customer focus, employee empowerment, and continuous improvement driven by top management. The unit of analysis shifted from the individual task or control chart to the entire organization’s culture and systems. TQM’s limitation was its very breadth: it offered principles but few concrete techniques for diagnosing operational problems or prioritizing action.
Two responses to TQM’s breadth emerged in the 1970s and 1980s, each narrowing the improvement agenda in a different direction. Lean Production (1970–present), developed at Toyota and systematized by Womack, Jones, and Roos, focused on the elimination of waste (muda) in all forms—overproduction, waiting, transport, excess inventory, motion, defects, and overprocessing. Lean preserved TQM’s emphasis on employee involvement and continuous improvement but operationalized it through value-stream mapping, just-in-time flow, and the 5S system. Improvement meant creating flow by removing obstacles, not just reducing defects.
At nearly the same time, Theory of Constraints (TOC, 1980–present), introduced by Eliyahu Goldratt in The Goal, offered a radically different diagnosis: improvement efforts must target the system’s bottleneck. Goldratt argued that improving non-constraints yields no systemic gain and that Lean’s waste-elimination drive could be misdirected if it ignored the constraint. TOC’s drum-buffer-rope method and five focusing steps provided a prioritization logic that Lean lacked. In practice, Lean and TOC became complementary rather than competitive: Lean tools (e.g., 5S, kanban) are often applied at the constraint, while TOC prevents overinvestment in non-bottleneck improvements. Both frameworks remain active, and many organizations blend them.
The early 1990s produced a direct assault on incremental improvement. Business Process Re-engineering (BPR, 1990–2005), championed by Michael Hammer and James Champy, argued that organizations should not improve existing processes but reinvent them from scratch. BPR rejected the gradualist assumptions of TQM and Lean, calling for dramatic breakthroughs through radical redesign and the use of information technology. The method involved ignoring the current “as-is” process and envisioning a “to-be” process that would deliver order-of-magnitude gains. However, BPR’s neglect of human factors—employees were often treated as interchangeable resources—led to widespread downsizing and disillusionment. The framework’s decline was swift, but its legacy was lasting: it cemented the notion of the “business process” as the fundamental unit of analysis and paved the way for process-oriented management and enterprise systems.
In the same period, another framework took a more structured path to improvement. Capability Maturity Model Integration (CMMI, 1990–present) emerged from the Software Engineering Institute as a staged model for assessing and improving process capability. Instead of focusing on waste (Lean) or constraints (TOC), CMMI defines five maturity levels—Initial, Managed, Defined, Quantitatively Managed, and Optimizing—each representing a level of process discipline and predictability. Organizations are assessed against key process areas that accumulate as they climb the ladder. CMMI differs from earlier frameworks in that it treats improvement as a capability-building journey with measurable milestones. It absorbs incremental improvement ideas (Level 5 explicitly calls for continuous process improvement) but structures them within a formal assessment architecture. CMMI remains dominant in regulated industries such as defense, aerospace, and software, where process auditability is critical.
Today, no single framework dominates. Lean (often combined with Six Sigma as Lean Six Sigma) is widespread in manufacturing, healthcare, and services. TOC continues to be used in production scheduling and supply chain management. CMMI is the standard in government and software development. Practitioners often mix them: Lean for everyday waste reduction, TOC for strategic focus, CMMI for process maturity assessment. The frameworks agree that improvement must be systematic, data-driven, and involve front-line workers. They disagree on the proper unit of analysis (the individual process vs. the system constraint vs. the organization’s maturity), the pace of change (incremental vs. breakthrough vs. staged), and the role of measurement (defect rates, throughput, or capability level).
New pressures are reshaping the subfield. Digital technologies—automation, IoT, AI—challenge the traditional assumption that process steps must be performed by humans in a fixed sequence. Agile methods, born in software, propose lightweight, iterative improvement cycles that contrast with the heavy documentation of CMMI and the disciplined standardization of Lean. Some practitioners argue that the next framework will need to integrate digital tools, human-centered design, and adaptive process orchestration. Whatever form it takes, it will build on a century of competing visions of what it means to make an operation better.