That’s Not What I Ordered

Remember the first time you went to that new restaurant down the block and it was SO GOOD! It was so good, you just had to take your friend there, but then… it wasn’t as good. Something was different. While this might make your dinner experience less than it could be, it (usually) won’t hurt you. In some industries, however, this amount of variation is unacceptable, because it could be deadly.

Consider an airplane manufacturer. The processes they use to manufacture must be designed with the highest standards in mind and the tightest tolerances. A mistake on your plate at dinner is annoying. A mistake on your engine 32,000 feet in the air is lethal. So, we need a way to anticipate and test for issues in our processes and products to a specific, quantifiable standard.


Processes in Review

When discussing process design a few weeks ago, we examined the IPO Method. This is the cradle of processes, where we discover their necessary inputs and desired outputs and then find what it will take in the middle to create that. These baby steps are where you start.

But, we made the important distinction that processes are never complete. They always require improvement, and this is where the PDCA Model comes into play. Improvement is a continuous process, but it is even more than that. It is also a culture, a mindset. Everyone in the organization must be empowered to pursue perfection.

After that we addressed how to deal with problems as they arise. The DMAIC framework is your go to bug stomper in any organization. By providing a way to not only create the necessary change but also institutionalize it, DMAIC is needed in any tool box.

What we are lacking is a design-oriented approach to controlling variation in our outputs. This is the problem that Genichi Taguchi set out to solve in the 50’s at Nippon Telegraph and Telephone Corporation. The end product; the Taguchi Method.


Introduction to Robust Design

The Taguchi Method classifies variation into three categories; internal variation, external variation, and unit-to-unit variation.

Internal variation deals with the natural degradation of inputs. Thinking of your meal, perhaps the chef had gotten a fresh delivery the first day you went, but decided to save money by using yesterday’s ingredients the day you went back.

External variation deals with things… external to the process. Perhaps the restaurant had a sewage line bust while you were enjoying your meal, or the restaurant was cold, or something else made the product – your experience of the meal – unenjoyable.

Finally, unit-to-unit variation deals with the variation inherent in two of the same part. Every beef patty and bun is not the same, and some are just better than others.

Taguchi provides a statistical method for anticipating the amount of each type of variation in any given process and then controlling it. If your restaurant had used his findings, they might have been able to deliver the sort of experience you expect each time,


The Process of Design

Designing the sorts of processes that that can reliably operate and minimize this variation is challenging. Fortunately, there are tools designed for this as well. While the IPO Model is useful for creating processes in the first place, other methodologies allow us to specifically tailor our processes for their purpose.

Consider DMADV. This process is used to create new processes, so while it shares the first three steps with DMAIC (Define, Measure, Analyze), the actual focus is different. The last two steps – Design and Verify – are meant to make sure the process will produce the product or service the end user needs. There are lots of great resources out there on DMADV. If you want to learn more in detail about this methodology specifically and how it differs from DMAIC we suggest starting with an introduction here.

The Taguchi Method may be embedded within the DMADV, beginning in Analyze and continuing throughout Design and Verify. Another key tool to address variation issues include Quality Function Development, also known as the House of Quality. While it is important to pick the right tool for the job, this should never prevent you from starting the job in the first place. There are always insights to gain and ideas to further develop. The best advice is to go for it!


In Closing…

As we conclude our discussion on the process of process, we should remember the power of all this. According to research published by isixsigma, Six Sigma methodology saved Fortune 500 companies approximately $427 billion between 1987 and 2007. Continuous improvement sometimes gets a bad rap from people claiming that it stifles employee creativity and is too rigid in its approach to things. While that might be true of companies implementing it in a poor manner, true Six Sigma ought to be empowering to employees and customers alike.

While our series has been focused on the employee side of the issue, there is no reason customer input, insights, and feedback can’t be involved at any step of any of these processes, or all of them for that matter. While the pushback against overly bureaucratic and corporate approaches to Six Sigma is understandable, done properly it ought to be one of the greatest tools for employee and customer empowerment. Numbers and process don’t lie, and Six Sigma is all about finding out what people’s actual talents and abilities are and using them to their fullest extent to achieve a common goal. That is what process is all about.


Sharing Is Caring!