What is Variation in your process and how to identify variation

This post is about process variation. In the last post, I talked about what a process means. A process will have some suppliers, who provide inputs to the process, and customers, who uses output from it. All the steps or activities that are done on the input to convert or change it into desired output is a process. We took the example of preparing a cup of tea.

Lets continue with the same example and try a simple experiment. Whenever you prepare tea henceforth, for the next 30 iterations, try to time yourself using a timer or a stopwatch. Once you do this, you will have 30 data points. These data points are the times that you took to prepare tea in each of those 30 iterations, represented by the term ‘transaction time’ or ‘cycle time’ for this particular process of preparing tea. Cycle time or transaction time is the time taken to perform all the steps or activities in that process. (I will take more about various TIMES in a process with respect to LSS in a separate post.) Your data set will look somewhat similar to the one shown below ( and believe me, it will, try it… )

What is process variation?

As you can see from the sample data set above, the time that you will take to prepare tea varies from, lets say, a minimum of 8 minutes to a maximum of 15 minutes. It’s the same person, preparing the same tea, using the same inputs, in the same kitchen, following the same process, still there is a lot of difference in time taken to do so. This is nothing but the variation in the process. And it applies to all processes.

Sometimes, the variation is very high. Which means the spread of data, or the difference between the minimum and maximum values in the data set, is huge. Sometimes it might me low. But there is always some difference. Even in the most precise of the processes.

Lets take an example of an Original Equipment Manufacturer (OEM) who manufacture a part for a jet engine, which need to be of 2 centimeter diameter. You will agree that this particular part needs to be of the exact specifications as required to fit in the engine. If it’s too big, more than 2 cm, it wont fit in the engine. If it’s too small, less than 2 cm, it again wont fit in that engine. So its obvious that the manufacturer needs to produce this part with a diameter of exactly 2 cm. However, the question is, does it happen in reality???

Well, the answer is, NO. Even in such a sophisticated industry, with all the automation and cutting edge technology in use, there is still some variation. Although, it might not be huge. It is always within the acceptable limits. The diameter may vary from 1.9990 cm to 2.0010 cm, however, there is variation.And it will look somewhat like this;

That’s the law of nature. Any process, may it be as simple as preparing tea or as complicated and advanced as manufacturing a surgical equipment, there will always be some variation in every process, the extent of which may again differ.

Types of Variation in a process

Variation is of two types, process variation and measurement system variation. Please read my post about measurement system variation to understand what it is and why it’s important to remove it before you analyze any data set. I have also clarified the 2 types within process variation in this post.

There are various ways to measure the variation in your process, depending on the data type, which we will look at in a separate post. But before that, we need to understand why variation is so important for every process, why it should be reduced to as low as possible and how it impacts the customers of your process. That will be the topic of my next post.

Sachin Naik

Passionate about improving processes and systems | Lean Six Sigma practitioner, trainer and coach for 14+ years consulting giant corporations and fortune 500 companies on Operational Excellence | Start-up enthusiast | Change Management and Design Thinking student | Love to ride and drive

Read More about Sachin