Companies struggle with “pull” systems in attempts to improve performance. Pundits exhort organizations to “pull to demand” or “pull, don’t push.” The general view is that pull is good, and push is bad. A radical idea will be proposed here to resolve decades of ongoing confusion on this topic, unite employee efforts, and improve organization performance. First, some history about the origin of the term.
For those who are not familiar with the term “pull,” go straight to operations science and save yourself the confusion of the last three decades. The origin of the concept of pull was in Lean Manufacturing which sprang from the Toyota Production System. Nowadays labeled just “Lean” for organizations across many different industries, the practices have drifted away from their origins in the Toyota Production System and those connections are fading into the mists of time. This is a good thing because
- most companies are not automobile production companies like Toyota
- the Toyota Production System (TPS) was developed through trial and error and its creators were not aware of the science behind all the changes they had made (see “There is a Science of Lean”)
- operations science provides a practical, objective approach to evaluate all options for improving performance
This is not to denigrate Toyota or the TPS, that would be foolhardy given the TPS’ and Toyota’s demonstrated success.
However, one size does not fit all. Promoters of Lean, especially in the early 90s as Lean Manufacturing was hitting its peak of inflated expectations per the Gartner Hype Cycle, would insist on mimicking TPS practices such as “pull,” heijunka (production leveling), or jidoka (error-proofing) to truly become enlightened and as productive as possible. The concept of “pull” still thrives most vigorously. It is dogma for nearly everyone promoting Lean, for many in operations control, and for many in project management, i.e., “Agile.” There is a more productive way.
The term “pull” is nebulous and there are many variations which leads to ongoing confusion. Also, there is not a universally accepted science of pull. Again, this is not a surprise as the concept developed through trial and error. Spearman and Hopp, in Factory Physics, which is the origin of operations science, defined the difference between push and pull as “A pull system establishes an a priori limit on work in process while a push system does not.” Spearman and Hopp then quantitatively defined the behavior of these systems and their characteristics. That’s the great thing about defining systems, the truth is self-contained. Especially when the definitions can be supported with science-based, quantitative description of system behavior. While the Factory Physics approach to defining push systems and pull systems is of great benefit, there is an even less confusing approach—don’t use the terms at all.
This may be a radical step for many, but no one need be forbidden from talking about pull. The urge will pass on its own once operations science is understood. The following is a brief foray into the operations science on this topic. (For more on operations science basics, see What is Operations Science).
Given resources (equipment, people, space), demand, and variability, the amount of WIP in a system determines throughput and cycle time. Cycle time here means the time to complete operations at various resources in a routing, not the time at any single resource. These dynamics translate into known, quantitatively described operations science behavior (see the curves below) and the best part is that understanding these concepts requires only basic math—addition, subtraction, multiplication, and division. For those appropriately concerned about the vital need to focus on people when making changes, operations science concepts are easily adaptable to any organization or process and at any level in an organization.
Operations science provides a quantitative and qualitative description of these curves , and how to apply the concepts in practice. With training in operations science, it becomes obvious that the amount of WIP in a process determines the throughput and cycle time of the system. Operations science also provides powerful insight into the optimal levels of WIP for a process. It doesn’t really matter how one accomplishes the WIP level or what that WIP control is called. Though, certainly, some approaches to WIP control are easier to manage than others.
The operations science approach is clear, easily understandable, and has been successfully taught to the entire spectrum of employees in organizations across many industries, from line workers, construction workers, customer-facing employees such as nurses, all the way up to the C-suite. It can be easily shown that there is great confusion within organizations about the operations behavior described by the curves. More importantly, it is fairly easy to clear up the confusion. Using unscientific terms like “pull” does not help. One could call WIP control “jump,” “shout,” or “kick,” the label doesn’t matter. Operations science shows that the amount of WIP in a system is a design parameter for achieving desired system performance.
Be wary of the siren song of slogans, expensive new technology, software, and expensive, extensive consulting engagements. Even AI is not immune to confusion. If an AI chatbot is trained through access to the worldwide web, it will pick up the rampant confusion about pull and feed that right back. To achieve results quickly and with the least stress possible, apply operations science concepts and use the science in a practical way to evaluate any proposed “best practices” such as pull.
Determine the best way forward with your existing people and technology. At the Operations Science Institute, we have often seen organizations rapidly improve their performance with only training and application of operations science concepts. At the least, it is a much less complex and risky endeavor than implementing a new software system, hiring expensive consultants, or kicking off a company-wide initiative.
Visit us at www.opscience.org for more information or contact us at email@example.com to find out how your organization can rapidly achieve successful results with your existing people and technology.
 Hopp and Spearman, Factory Physics, 2008, Waveland Press, Long Grove, p. 358
 Ibid, Chapter 7
 Pound, Spearman and Bell, Factory Physics for Managers, 2014, McGrawHill, New York, pp. 82-92
Edward S. Pound is Managing Director of the Operations Science Institute.
Banner illustration courtesy of DALL-E, and lots of editing, at openai.com