Conflating Systems, Strategies, and Tools
As I said in my premier Heretic column, “Dogma and Definition,” I’m interested in examining our assumptions and premises about Six Sigma so we can discard the dross and refine the potential benefits in implementing it.
To that end, I have decided to co-opt an omnipresent element of pop culture. No, not “reality” TV (though a Six Sigma reality show strikes me as very funny—you heard it here first).
And so I humbly present to you the Top Ten Stupid Six Sigma Tricks, beginning with no. 10: Conflating systems, strategies and tools.
First off, the word “stupid” here isn’t intended to insult anyone, just to get your attention and make you want to forward this article to all your buddies. I could have called it the “Top 10 Ill-Advised Six Sigma Activities,” but that just doesn’t trip off the tongue, nor does it capitalize on the popularity of certain other cultural memes. Besides, I’m obviously not talking about you, so lighten up. It’s all those other people about whom I am speaking.
Mental models are approximations that we use to understand and order our thoughts in relationship to reality. If we have a mental model that convinces us of one thing, and reality is different, reality wins every time. This stupid Six Sigma trick (SSST) is due to misunderstanding the different functions a particular mental model can perform. In the business of running a business, I divide these models up into a hierarchy of systems, strategies and tools. To illustrate these in real terms, let’scompare running a business to building a house.
In a business context, I define a system as the interlocking of elements that a business uses to manage and coordinate its processes. The system consists of rules of how strategies and tools are used and interact. In construction, the system is how the buyer, suppliers, contractors, workers, architect and inspectors interact to build the house. In my experience companies rarely consider the system they are intending to use to accomplish their objectives. Instead, businesses seem to rely on individuals forging their own way of getting things done, and of course that results in variation, miscommunication and wasted effort at all levels.
I think that the next big thing (NBT) will be an integration of strategies and tools into a cohesive system. Some people claim that Six Sigma is such a system. I tend to think not, for two reasons.
First, in “Dogma and Definition” I presented a case for narrowing the definition of Six Sigma to its original roots as a problem-solving strategy to reduce confusion. Second, as the rest of the top 10 list will show, there are a lot of holes in many Six Sigma implementations that suggest it’s not a holistic system. There are models out there that do integrate the various strategies and tools, including the Malcolm Baldrige systems perspective and Business Performance Excellence, so this isn’t stuff you need to invent. But you do need to think about it, communicate it and implement it at your business, if you haven’t already.
A strategy is an organized set of steps or actions that you take to address some type of issue. In our building example, the blueprint is the strategy. It is the thing that prevents us from making a house with no doors, or putting sewer pipes in the middle of the dining room. In Six Sigma, DMAIC (define, measure, analyze, improve, control) is a strategy for solving problems and closing gaps. Others include:
- Supplier quality assurance—steps you follow to make sure you get what you need from your suppliers
- Customer quality assurance—steps you follow to query your customers to make sure they are getting what they need from you
- Daily management—steps you follow to manage and improve your process on an ongoing basis
- Quality improvement—steps you follow to validate your process targets and reduce variability
- Cross-functional management—steps to eliminate “vertical” management silos and coordinate activities across your business
- New process start-ups—steps you follow to ensure that new products and processes start off in high conformance with internal and external customer requirements (e.g. design for Six Sigma)
As you can see, different strategies apply to different situations, and multiple strategies are called for in an organized business management system. I find that most teams working on addressing these issues have great creativity and drive to accomplish their task, and they know where they want to go; what they’re missing is a strategy or roadmap for getting there. So teams try to invent how to do something, or they incorrectly apply a strategy that’s useful in some other context, instead of following a structure that has survived the crucible of years of use.
I think of a tool as a general method for generating knowledge or accomplishing a focused task. When building a house, we use many tools, including hammers, saws, and tape measures to connect, disconnect, and measure many different things. Tools can be used in many different strategies. A hammer can be used to create the frame, attach the dry wall, or hang a picture, but it always involves whacking something soundly, and is used whenever a good whacking is needed. A hammer is not an efficient tool for tightening a pipe, though.
In my experience, people frequently confuse tools with strategies. I often hear people say that they are using statistical process control (SPC) or experimental design (EXD) as their strategy for doing something. These are tools, and you could use them during the process of any of the strategies mentioned earlier. This is the defining characteristic of tools—they are used to accomplish the same task whenever that task is called for by the strategy you are using. Using tools in the absence of a strategy is like building a house without a blueprint. A lot of work is done, but it is not likely to result in something that you can use, except by accident.
Tools that may be used to accomplish your tasks include:
- Process control technologies—tools that monitor and maintain process settings, including, among others:
- Statistical process control—tools that assess processes for the presence of unexpected sources of variability
- Standard operating procedures (SOPs)—a state of consensus amongst the operators of a process on how that process will be run
- The common quality tools—low-tech and very useful aids in defining processes, organizing ideas, organizing data, understanding variation and managing projects
- Probability and statistics—the basics for understanding variability in your processes
- Experimental design—techniques for efficient and economical interventions in a process to determine how to affect process mean, distribution shape and variation
- Reliability tools—tools to make predictions about the ability of a process to perform its intended function for a given amount of time
- Housekeeping—tools for keeping workplaces organized, neat and clean
- Asset utilization metrics—measures to track how effectively you are using your assets to produce your product or service
- Failure mode and effects analysis (FMEA)—a tool for identifying the ways a process can fail, and prioritizing changes to minimize the severity of the failures, the frequency of the failures, and/or the probability of missing the failures if they occur
- Root cause analysis—a tool to identify what events and states led to a particular occurrence
The effect of SSST no. 10
A prime example of this error is what happened to SPC in the 1990s. People were slapping control charts on anything that moved (and some things that didn’t) and expecting the process to get better. But frequently there was no system to take that information and use it to identify and prioritize what to fix in order to prevent out-of-control causes from happening. There was no strategy to figure out what the most important aspects of the process were, so that an intelligent decision could be made to allocate resources to monitor just the critical process outputs and product characteristics. Instead, we control charted everything and reacted to nothing.
My favorite was the control chart on the number of out-of-control points on all the control charts in the department. “See, our processes are in control in how out of control they are!”
You would run into business areas that had a macro that would create and print more than 200 control charts with the touch of a button. They did this weekly. They had 20 employees, all “willing workers” doing their best to figure out how all this paper was supposed to make their jobs better, and mostly wondering why they were spending time making the ugliest wallpaper known to mankind.
When it’s time to understand the variability of a process, and you want to know if the process is subject to special causes, SPC is unparalleled. You might do this to track incoming materiel from your supplier as part of supplier quality assurance, you might be doing this to track down a potential problem source as part of a DMAIC project, or you might be doing this to track the effectiveness of a management intervention in your safety performance. However, SPC by itself doesn’t improve anything; it only identifies unusual sources of variation, and even eliminating those only leads you to the process and doesn’t improve the underlying variability.
When it’s time to understand how to affect a process’s mean or variability, then experimental design is a great tool. You might do this to see if different product stocking affects sales, or to optimize the 10 process variables of your existing line, or to find what the ideal settings for a new piece of test equipment are. Again, experimental design does not guarantee that what you’re working on has a positive effect on the bottom line unless it is in the context of a strategy (maybe DMAIC) and system (maybe strategic planning).
Sure, in some cases the exceptional business leaders might build a system or strategy on their own, but these things should have been part of the deployment of SPC or EXD or whatever, not something everyone had to invent. Why wasn’t it? Because the leaders performed Stupid Six Sigma Trick no.10 and mistook a tool for a strategy, training their people in how to use a tool but not giving them a strategy within which it could be used.
And so, you run into people who say, “Yeah, we tried quality circles (or SPC or TQM or EXD or whatever), and it was a big waste of time and money.” A perfectly good technique was taken out of context and mistaken for something that it is not, it makes a stink, and the business loses a small, critical element each time it holds its nose and runs away.
Only a fool would think that Six Sigma is immune to this. Hence these ravings.
The aficionado might add a category of meta-strategies—strategies on how to use other strategies—such as strategic planning (to be covered in a future column), disciplines, like plan-do-check-act (PDCA), that are habits that should permeate every part of the system. But this is quite enough for now.
To avoid Stupid Six Sigma Trick no.10, businesses must understand the differences between systems, strategies and tools. Failure to keep these differences in mind leads to a misunderstanding of when to use a technique and to a propensity for being overwhelmed by terms that seem to have no relationship to each other. This, in turn, leads a business to dismiss a perfectly good tool (say, SPC) as having been superseded by a different tool (say, experimental design) when they are actually complimentary tools for different objectives.
Then again, I could be wrong.