There are a couple of topics that most Black Belt training includes that we do not. Here's why.
First, our training has been honed for over 20 years in Fortune 500 companies, so it includes those topics that we have seen our students need to know to extract the most benefit from their problem-solving activities. What you will find is that we cover more topics in more depth than any other Black Belt training, and that the topics covered are the ones you end up using. Secondly, there are a few topics that have worked their way into the cookie-cutter Black Belt curriculum that are not useful, or worse, invalid. What are these?
Financial calculations
We spend a lot of time in the course talking about the importance of showing a financial benefit, both to get a project started and at the end of the project. However, we do not spend a lot of time teaching the Black Belt how to do this. There are a couple of reasons for this. First is that no one is going to believe a Black Belt's assessment of their project's benefits without some other verification anyway. Second, and more insidiously, we have to watch out for the "careful what you ask for you might get it" mentality. For example, at a company of our acquaintance Black Belts calculated that they saved their company millions of dollars by efficiently recalling defective product from their customers. By focusing on the cost savings they neglected fixing the real problem, which was why the product was manufactured incorrectly in the first place!
We prefer to spend the time teaching Black Belts more tools that will be useful in their chosen area of expertise: advanced problem-solving. We do train the Black Belts how to use the Taguchi Loss Function, since that is a crucial technique for choosing the optimum solution. However, these calculations are not intended to generate actual dollar losses or savings, though they do generally predict correctly the order and magnitude of real losses.
Response Surface Methodology (RSM)
This is almost universally taught to Black Belts, and for no good reason. Imagine a simple experiment where you manipulate two continuous variables, Pressure and Temperature. If you picture the settings on an x- and y-axis, the effect on your response variable (say quality) can be plotted on the z-axis, resulting in a 3-D surface. The idea is that once you have a part of this surface, you can look for the steepest slope, then follow the surface to get settings in Temperature and Pressure and eventually home in on a local maximum.
While there is nothing wrong with the technique, we just don't find it useful in real life all that often. Frequently the variables that you are interested in manipulating are nominal, like vendor or production line, so there is no surface. Additionally, because of sampling error, it may take you a while to home in on that local maximum. Finally, many people find RSM to be confusing and complicated. We have seen that it much more efficient to encompass the entire range in a fractionated screening experiment and when needed set variables at three levels. Fractional orthogonal array designs are pretty straightforward to understand and can be easily customized to fit your research question. We also teach how to turn the output from such an experiment into predictions for settings you didn't even run to enable you to optimize your output.
Signal-to-Noise Ratio
Since we are talking about teaching highly fractionated orthogonal arrays, some people advocate the use of the signal-to-noise ratio in the analysis. This has been shown to be invalid, and there are better ways of achieving the same objective with the post-hoc analyses we teach.
Shainin Techniques - Multi-Vari
Shainin techniques, for example multi-vari analysis, are frequently taught in Black Belt classes. These too have been known for some time to be invalid, and again there are simple, rigorous ways of achieving the same objectives.