Lean Practitioners Beware: You May be Tampering

In a conversation with Daniel Sloan, Six Sigma Master Black Belt, CEO of Evidence-Based Decisions and author of Profit Signals, I learned about a Deming idea known as “tampering”. What I learned is that we, as Lean practitioners are at risk of making things worse, not better, if we make changes without applying the scientific method to prove a theory.
Deming (originally Deming’s teacher Shewhart) said “tampering” is the changing of a process in reaction to just one instance of its output. Tampering is depending on intuition and common sense. Deming demonstrated through many examples that this can often make things worse, not better.
An example of tampering often given is setting numerical goals or changing performance reviews and compensation systems. Rewarding what you measure often only improves that, and can be to the detriment of other factors, often for a net loss. See blog entry “Be Careful What You Measure, You Just Might Improve It”.
What does this mean for the Lean practitioner? It is often said that much of Lean is “common sense”. An experienced Lean facilitator can often “see” the Future State when walking through a Value Stream and will be able to “intuitively” tell you that you can improve productivity 30% or cut inventories 50%.
Are we, as Lean practitioners tampering? Yet we know through experience how effective Lean can be, even at this “common sense” level. We can back up our intuition with numbers. Is Deming wrong? Certainly we have all seen examples of Lean done poorly or even disastrously after a quick read through a book or after attending a seminar, when the implementation lacks a clear road map for implementing Lean throughout a Value Stream. The result is changing and measuring one factor without realizing the impact through the whole system.
As Daniel Sloan and I discussed this further, he enlightened me that Deming also said that the “tampering” or changing inputs based on one factor alone was okay as long as one had a theory or hypothesis one was testing. Essentially, Deming was calling for the scientific method. What then is the theory or hypothesis that Lean practitioners follow that prevents their work from becoming an example of “tampering”?
The most successful Lean company today is Toyota Motor Corporation. Most Lean practitioners would agree that Lean is the Toyota Production System (TPS). When we implement Lean, we are implementing as much of TPS as we can, as completely as we can. The theory behind Lean is that the majority of effort behind any process is “Muda” or waste, and that this must be reduced since Muda erodes profit.
There are many tools for reducing waste, and each tool or hypothesis must be tested and proven scientifically. Each time a Lean practitioner does a kaizen or implements a component of TPS, they should be following the scientific method to prove a Lean hypothesis. Examples of this include 1) the pull system requires less inventory than push to meet on time delivery at a low cost, 2) one-piece flow will cost less than batching, 3) high 5S scores correlate to high quality, 4) TPM activity will improve OEE (will reduce downtime).
If the Lean practitioner can not prove any hypothesis (disprove the null hypothesis), it may be better to study why this is the case rather than battle to prove that they “know” that Lean is right and suffer political damage. Very often there are real reasons particular to the time and place that one-piece flow is in fact not better than batching (e.g. long set up times, poor quality) and these problems must be addressed so the hypothesis can be proven.
When Lean practitioners focus on the tools only without understanding the theory or system behind it, tampering happens. One such example we have seen is where Quick Changeover was implemented successfully, but without understanding the Lean theory that smaller lot sizes and pull production must follow to achieve TPS. As a result the labor cost for changeovers was reduced, but velocity did not improve, inventories were not reduced, and throughput did not change.
Fortunately we need look nor further than the Toyota Production System and the scientific method to avoid being guilty of tampering with things in our efforts to implement Lean.

2 Comments

  1. Ron Jeffries

    March 19, 2004 - 12:07 pm

    An interesting thesis. But it leaves me uncertain as to what to do. If I suggest a change to a team’s process, I always have something in mind, and I pay attention to whether I get what I expect. If I don’t, I suggest something else.
    If I embraced the ideas here, what would I do differently? I look forward to subsequent articles elaborating on that.
    Thanks!

  2. Frank Patrick

    March 20, 2004 - 10:38 am

    What to do is to minimize the “tampering” up and down the full chain of the system and instead, focus efforts on exploitation and elevation of it’s weakest link — it’s constraint. IMHO, while we’re on the subject of Deming, his concept of “profound knowledge” is related to understanding the system in terms of it’s current and desired constraints, and acting appropriately. Lean success is accelerated and optimized when the global system is the subject, and not just the local components.