Management teams expend tremendous amounts of time and effort determining what should be measured and how. Some of these metrics are so important to the organization they are deemed Key Performance Indicators (KPIs). Once agreed upon, resources are then marshaled to monitor and act upon movement (or lack of movement) in these KPIs.
Unfortunately, many teams then proceed to make a simple mistake and lose a significant amount of the value created in the KPI effort. The outcome of this error is wasted resources on failed projects and the business impact of not improving the KPIs.
The mistake is, while establishing the metric’s baseline, they don’t take into account the variation inherent within the process being measured. There are two primary methods of making this error: ignoring and hiding.
Standard Practice #1 – Ignoring Variation
Assume the chart below graphically displays the results of a process over time (in this case higher is better). Also assume the management team did not have any information prior to the point they selected. We can now discuss the potential impact if the organization establishes their baseline using only the time period A, B or C.
Point A – Point A seems to show good results and is actually in the top three. If our management team chooses this data point as their baseline, they will quickly become concerned when the results begin trending down. One typical reaction is to quickly assemble an improvement team.
However, if the improvement team is looking for a single root cause (or a “vital few”) for the poor performance, they will likely be unsuccessful. Eventually the effort will lose steam and fizzle out with performance remaining below the level management, or the market, demands.
Point B – This point looks very low. If the baseline performance is deemed too low and an improvement effort is chartered, it will quickly and easily show positive results. However the results will likely plateau around 20. But if the goal were higher than 20, the team may have a difficult time; especially if their search is focused on finding “the” issue. Again, the initiative will likely fizzle out with no real benefit from the time, effort and expense.
In the above scenarios management is reacting to Common Cause Variation (CCV) as if it were Special Cause Variation (SCV). CCV is simply the variation caused by the interaction of numerous relationships and is to be expected within a certain range of results. SCV is when something “special” happens within the operations and it leads directly to results outside of the typically expected range.
This is important to understand because when a process is experiencing CCV, improvement efforts will most likely need to continuous improvement flavor. Instead of single “big win,” progress will most likely be from many smaller improvements and involve a longer time commitment. But if management expects dramatic improvement immediately, they will quickly grow frustrated with the improvement team’s results.
The cause of management’s inappropriate reaction is that the measurement system isn’t helping them (i) understand what type of variation they are experiencing, (ii) determine if action is warranted and, if so, (iii) identify the type of action to take.
Point C – If the organization selected point C as their baseline, then any improvement effort will appear to show positive impact because the process is experiencing SCV (the results are outside of what we would typically expect). As soon as the special cause is no longer present, performance will improve even if nothing is changed.
Standard Practice #2 –Hiding Variation
Some organizations deliberately hide the variation within their processes. They use averages, rolling averages, deleting outliers, etc. Such practices are specifically intended to conceal variation and deny valuable information to management.
Rather than ignoring or hiding the variation, a better option is to try to understand it. A control chart can aid this understanding. It is easy to construct and simple to use. Below is an example of one type of control chart using the previous data.
One of the many benefits of a control chart is it visually displays whether or not a process is experiencing CCV or SCV.
Using this approach, our management team would have known points A and B were simply CCV (they do not violate established rules) and there was very little use in starting an improvement effort that would look for the “big win.”
Likewise, management would have known point C was SCV and rightfully expected an improvement effort to identify the “special cause” and make recommendations to keep it from happening again.
The important point is the management team is reporting results in a way that facilitates understanding. With this information, they are able to more effectively manage the expectations of improvement efforts both in terms of time and impact. This will result in fewer resources wasted on unsuccessful projects.
Content contributed by hiSoft Technology International Limited, a consulting services firm. For more information, contact Darrell Letourneau, Senior Managing Consultant in the U.S. Consulting Division, at email@example.com or 704-944-3155 or visit www.hisoft.com.