I overheard a fellow say “I’ve got Deming’s principles down pat, now all I have to do is understand this variation thing.” Hmmm, Dr. Deming was a statistician and his philosophy did come from his understanding of variation as taught to him at the Western Electric plant (Chicago, IL) in the late 1920s from Dr. Walter Shewhart. What W. Edwards Deming learned was how to evaluate data using a statistical process control (SPC) chart. To me, the difference between knowledge and tampering or guessing.
Early in my career I was a corporate director of operations where I learned to evaluate income statements and compare last months revenue, expenses, etc. to this months and all types of dictates and commands came from this naive view of data.
After attending Dr. Deming’s 4-day seminar and learning from the likes of Dr. Don Wheeler and Dr. “Frony” Ward, I learned a better way to manage with data. In statistical terms understanding the differences between common and special causes of variation. Let’s pretend we have sales of 15, 19, 14, 16, 12, 17, 15, 17 and 11 (in thousands). A manager might conclude that the month with 19,000 in sales is a celebratory moment best month on record and the last month with 11,000 is reason to “bark” at the salespeople for poor sales.
By plotting data using the SPC chart (below), we can tell that we can expect anywhere from 5.1 (LCL-Lower Control Limit) to 25.1 (UPL-Upper Control Limit) with an average of 15.1. A manager celebrating 19,000 or getting upset over 11,000 is foolishness. In a matter of fact, we can expect between 5,100 and 25,100 (the control limits) in sales and it wouldn’t be unusual. This is called common cause variation.
Conversely, if the next month showed 28,000 in sales (see chart below) this would be outside the UCL (Upper Control Limit). The $28,000 month is unusual (outside the limits) meaning we have a special cause. Something unusual has happened. Now is the time to investigate the reason there is overwhelming evidence that we should investigate the “special cause.” There are other indicators of special causes (run of 8 and others) that need to be accounted for, but this is a blog.
Not understanding the differences between common and special causes leads a manager to tamper with the system. Dr. Deming outlined two types of mistakes:
- Reacting to an outcome as if it came from a special cause, when it came from common causes of variation.
- To treat an outcome as if it came from common causes of variation, when it was from a special cause.
A systems thinking organization (or any other organization) must understand the differences between special and common causes of variation in order to manage effectively. Leadership development, organization change management programs and even technology implemented devoid of these basics are keeping service organizations from making better decisions. This isn’t just for Lean Six Sigma Black Belts and Master Black Belts, we all use data. We must know how to use this data to make better decisions and avoiding the mistakes Dr. Deming warned us about.
Tripp Babbitt is a speaker, blogger and consultant to service industry (private and public). He is focused on exposing the problems of command and control management and the termination of bad service through application of new thinking . . . systems thinking. Download free “Understanding Your Organization as a System” and gain knowledge of systems thinking or contact us about our intervention services at [email protected]. Reach him on Twitter at www.twitter.com/TriBabbitt.Share This:
[…] reality is without understanding both variation and context we are learning to make the system worse through tampering. In terms of what W. […]
[…] my most read post Service Metrics: What You Need to Understand I outline how to analyze data. You can do the same thing with employees performance when data is […]
[…] must understand variation. If you do not understand variation you may want to read this post (Service Metrics: What You Need to Understand) to give you some clue before reading […]
[…] using SPC (statistical process control) and have a post that is one of my most read posts – Service Metrics: What You Need to Understand. I won’t belabor the point except to say if you are not charting data, you stand to tamper […]
[…] to many of us describes a statistical term. Variation has been descibed in a previous post titled Service Metrics – What You Need to Understand (take a minute to read). Data taken from organizations will display common cause variation or […]
[…] So, take the pledge by first understanding variation or start by reading Service Metrics: What You Need to Understand. […]
[…] Most important, is that the Funnel Experiment describes variation. Tampering is to treat a common cause of variation as a special cause and vice versa. Without a fundamental understanding of variation, you stand to make mistakes or “tamper.” If you are not familiar with the concepts of variation, I wrote a post titled, Service Metrics: What You Need to Understand. […]
I agree to “… avoiding the mistakes Dr. Deming warned us about”, such as the vast nonsense promoted by Six Sigma … counting defects; normalizing data; confusing analytical and enumerative tools; creating barriers with the belt system …etc etc …
If you are not already, I suggest becoming familiar with the work of Nassim Taleb on randomness. As with Mr. Porter, Mr. Taleb offers a unique perspective by which to interpret variation that again in my interpretation is compatible with Dr. Demming’s, and when considered together provides additional benefits.
1. “Fooled by Randomness – The Hidden Role of Chance in Life and in the Markets,” by Nassim Nicholas Taleb, Random House, 2005.