BACK TO BASICS
That’s So Random—or Is It?
by Cecelia McCain
Assignable cause, random variation, process stability, process capability—these terms can easily confuse those new to process management. Does random variation in a process automatically require process improvement? Can a process be stable and capable with an assignable cause present? Not knowing the answers to questions like these can cause process owners to overengineer and possibly overcorrect their processes.
A simple example of random variation is the process of driving to work. Let’s say your typical drive time is 20 minutes, but it’s not exactly 20 minutes every day. There are random variables or causes that impact your drive time, including the speed at which you drive, traffic patterns and the time you leave the house. So the random process variation may extend from 18 to 25 minutes.
One day, however, your drive time increases to
55 minutes (see Figure 1). You investigate the cause and
determine there was an accident, which caused the process to
spike out of control. The accident is the assignable cause of the
out-of-control situation. In other words, you know what happened
and can assign a cause to the out-of-control situation in the
process. Even though you know what caused the increase in drive
time, you wouldn’t revise the process because of this one
The same theory applies to business process management. You need to establish a random process variance for each of the processes for which you’re responsible, such as the drive time variance of 18 to 25 minutes. You need to know what you can expect and plan for every day. Then, when random, assignable cause incidents that negatively impact the process occur, you need to investigate the cause, assign a cause, document the cause and continue to monitor process performance.
An assignable cause log (see Figure 2) can
effectively document assignable cause incidents, allowing you to
capture the information for use in process performance reviews or
If an assignable cause incident becomes chronic or recurring, you will need to apply a process improvement to bring the process back into control or to the previous acceptable level of process variation. To illustrate this point, let’s say your drive time increased to 55 minutes per day for five days in a row, causing you to be late for work all week, which is unacceptable. Now you have a chronic, recurring, out-of-control situation, so you need to investigate the cause and implement a process improvement.
After investigating the cause, you determine the train schedule changed, which negatively impacted the flow of traffic. Therefore, you need to adjust or improve the process to remove the now chronic assignable cause. Simple improvements include changing your driving route or work schedule or leaving home earlier or later to mitigate the negative impact of the changed train schedule on your driving time. More complex and drastic improvements include selling your home or changing jobs.
Random cause variation is inherent in every process, from simple order taking to complex automobile assembly. Look at the processes in your own organization and identify the extent to which random variation exists by plotting the data over time. Use simple control charts and establish upper and lower control limits to maintain a stable process.
Assignable cause will occur in every process sooner or later. The key to effective process management and stability is to investigate and identify all assignable cause incidents, while remembering you don’t need to make process improvements after every occurrence.
CECELIA MCCAIN is the director of quality program management at Compuware Corp. in Detroit. She is a member of ASQ and a certified Six Sigma Black Belt.