I had a coaching session with a student who indicated that the primary project metric for the past 12 months had 500000 to 600000 transactions, including defect data. I agreed with him that this is too much to deal with in Minitab for a first project. His sponsor wanted him to focus on the top three product lines rather than look at all products as an effort to shrink the data set. This choice would cut the data set down to around 100000, which is probably manageable.
Is this a good plan forward?
Should the student do more?
I agree that this is a good start, but it is a pareto-like question. The project would only address the most frequent running products, which are about 20% of the entire production. There are other questions to answer. A belt should keep an eye on the project, but also on the enterprise. Even if his sponsor says to focus on only part of the products, the belt should also collect enough data to understand how generic his solutions end up. Your hope is that at the end of the project, the improvements will benefit all products. So do not avoid them right now.
My recommendation was to collect data across the year for the top products and perform a baseline analysis to ensure the process is predictable. Work your project on this data. But I asked him to also collect a month of data on all products. Use it to validate the ideas found in the year long data set to ensure they are generally true. This one month of data can also be used to validate the sponsor’s belief that we need to focus on only a few products to fix the business issues.
Today a Green Belt Student shared with me that his biggest takeaway in the class is not really the tools to run a DMAIC project (although they are good) but his new awareness of how all of the projects should be coordinated so that the overall result is a business improvement, not just local solutions for each problem. He experienced an epiphany that many of us reach, that every process is part of a system. Over-optimizing one process may make the system less successful. It made me feel good.