Summary
This resource explains how businesses should use Statistical Process Control (SPC), which provides a way to understand the variations found in every process, including business processes that involve Key Performance Indicators (KPIs).
Using an actual business case scenario, the article shows how businesses can discern whether the numbers 1) hint at a stable or unstable process, 2) indicate variation that is inherent to the process or outside the process, 3) call for a process change in order to get better results, or 4) require businesses to identify the actual cause and take corrective action.
To illustrate, the article looks at a typical cumulative year to date indicator, which is compared to a target and the traffic light associated with it. Two hypothetical situations are used to explore what the cumulative traffic light position would report and what an SPC chart based on individual months might show.
Through hypothetical situations, the article points out that the key difference between traffic light assessment and SPC is attention versus action. Traffic light assessment can give an indication of where attention might be placed. SPC can not only suggest where attention might be placed but also play a key role in determining what action might be appropriate.
Exploring Statistical Process Control
SPC is all about understanding the different kinds of variation you find in every process:
| Controlled Variation – Common Cause |
• Stable, consistent pattern
• Constant causes / “chance” |
• process is stable
• variation is inherent to process
• therefore, process must be changed in order to get better results |
| Uncontrolled Variation – Special Cause |
• Pattern changes over time
• Special cause variation / “assignable” |
• process is unstable
• variation is outside the process itself
• cause should be identified and “treated” |
The cumulative year to date indicator, which is compared to a target and a traffic light associated with it, is one that is often used in reports. Shown below are two different hypothetical examples exploring what this cumulative traffic light position would report and what a spc chart based on the individual months might show. Currently we report using just the traffic light version.
The indicator shown is % of Elective Admissions where the procedure was cancelled – a key NHS measure.
The SPC chart plots the actual % on a monthly basis.
The traffic light indicator below shows a cumulative position and is based on less than 1.39% is amber and above is red which are the targets we use.
-
The traffic light deteriorates to red flagging up that our cumulative performance has worsened to above 1.39%
-
Our response on the basis of this should be to investigate the exceptionally bad months. We should not see this red light on its own as a need to redesign the whole process as what we’ve seen is special cause variation. Something happened in those two months which we need to understand to ensure it doesn’t happen again
-
We may well decide that the underlying performance isn’t adequate but then we need to pursue a common cause strategy. (And a question might also be why did we flag something up as amber if we weren’t satisfied.)
An alternative scenario is shown below (again using hypothetical figures):
-
Here our current way of traffic lighting would show deterioration to red in March
-
However what we would have missed is the trend from July to December (signalled by the C – 6 points in an upwards or downwards trend)
-
Unless we’re looking elsewhere at this indicator over time, our current reports would not pick this up. And although we do show graphs on an exception basis – we define an exception in general as being a red light
Comment on how useful traffic lights and spc charts are in the examples above.
Traffic light assessment can give an indication of where ATTENTION might be placed based on some target. However even here there are some caveats:
-
Early months can create a distorted picture of a cumulative performance because we always start reporting from the beginning of each year. A bad month early in the year could give you a red until you’ve gone far enough into the year for the cumulative to smooth this out.
-
We might miss important signals (e.g. where a trend has happened but it hasn’t yet caused a traffic light change.)
SPC can play a key part in determining what ACTION might be appropriate e.g. whether we need a special cause or common cause strategy. It can also play a part in suggesting where ATTENTION might be placed eg if we spot a trend which hasn’t yet had an impact on a traffic light assessment but could do.
|
