One of the things I’ve learned about process improvement is that to change a process dramatically you must understand the process intimately. You must understand the ins and outs; the ups and downs; what is good and what is bad. You must not only know what things you are doing right, but you must also know what things you are doing wrong.
In the lean manufacturing world, this is accomplished through the use of Value Stream Maps (VSM). The Current State Map (CSM) provides an amazing amount of data, in a very visual manner, and points out where your problems are. Through this methodology, it is then possible to create a vision of what you want with a Future State Map (FSM). And from the Future State Map you can create an action plan for improvement.
But for many organizations, you can’t just go out and create a new CSM on a weekly or monthly basis for every process in the company. So how do you learn what your processes are really producing and how they are acting? A very easy way is through the use of Process Behavior Charts (PBC); often times called control charts. This technique can provide us with a tremendous amount of knowledge and insight if we will take the time to study the chart and learn from it. Additionally, these charts can be established as real-time, ongoing charts through the use of off-the-shelf software.
If you are not familiar with PBCs (a.k.a. control charts), then a quick explanation is in order. These charts show how a process functions over a period of time. The results are posted either as an individuals and moving range chart (where there is only one value per data point) or as an average and moving range chart (where multiple values are collected and averaged for each data point).
A center line (average line) is drawn, as well as an upper natural process limit and a lower natural process limit. These limits are set at three standard deviations above and below the center line. The second (or lower) graph shown on the PBC is the moving range. This graph also has an average and an upper limit on it. This graph shows the value of the change from point to point on the upper chart.
While there are exceptions, as a general rule, any data points between the upper and lower limits demonstrate how your process is behaving. Generally speaking, there is no reason to even ask why results between the limits are where they are on the chart. Any variation inside these limits is the routine noise of the process. In other words, it is what it is. This “is” what your process does naturally. There is no reason to “go find out what is wrong.” Additionally, these charts are classified as in-control or out-of-control. While in-control processes are what they are, if a PBC is out-of-control, work needs to be done to stabilize the process and bring it into control so that you can learn from how the process behaves when it is stable.
But, even for controlled PBCs, unless you change the overall process, the results are what you are going to get. Which brings us back to the topic. If you don’t know what your process is doing, and how it is acting, how do you know what to do to improve it? In the lean world of kaizen there is a very tight set of rules you operate by to make improvements. The actions are based upon your VSM set and resulting action plan.
However, very difficult variation and defect problems are not regularly corrected even with lean methodology. You may address some of the problem, but when it gets really tough with variation or defects with no apparent cause, you must find another way. PBCs are the starting point for the “other way.” Before you decide what tools to use and what to do, how about understanding what is happening first?