The Statistical Process Control (SPC) chart provides your organization with clear and concise answers to some of the most sought-after questions about performance and improvement for your operations. Analyzing an SPC chart not only tells you if your organization is meeting its goals and targets, but it also helps identify if the metric is improving, getting worse, or fluctuating in a normalized cadence. By applying three simple rules when analyzing SPC charts, you can distinguish between when to step back and analyze and react to a change.
The SPC chart gives you a clear indication when a metric has changed within your operations for better or worse. SPC charts will not tell you why things have changed but will show you the underlying signals needed to hone your investigation. These signals allow you to focus your effort, saving you and your team countless hours overtime. Learning when and when not to react will enable you to build a systematic process improvement culture.
In this article, we'll learn how to interpret a predictable SPC chart and how to identify and react to signals. This skill is fundamental for project leaders that will be using WorkClout for continuous improvement in their processes. As a project leader, you'll apply this skill to help make better decisions that will ultimately save you time by only reacting when necessary to changing metrics to improve more effectively.
Let's first start by observing the example SPC chart below. The SPC chart displays a production line's product output over the past 30 days.
WorkClout can generate these SPC charts for you containing a few helpful horizontal lines used as a guide for interpretation that you can use for any of your WorkClout Projects.
The Natural Process Limit lines represent a boundary to indicate where the predictable metric will generally fluctuate over time unless there is a change in the system that causes a shift.
One important thing to keep in mind for SPC Charts is that The Natural Process Limit lines are calculated automatically based on the existing data set present on the chart. These lines should not be confused with the target goals, which could be set custom by Project Owners for each Project.
SPC charts help identify whether or not your metric is predictable over time. In a predictable system, you can expect future performance to fluctuate within the two the Natural Process Limits lines. The metric will generally spend an equal amount of time on both sides of the average line in a predictable system.
Looking back at the example graph above, we can determine that this is a predictable system. You can deduce that the output of product over time will fluctuate between 2,400 and 4,400 on any given day in the near future.
So what does this mean? Applying your new learnings, you know not to automatically react when there is a change in the product output above or below your target goal so long as it is within the boundaries of the Natural Process Limit lines. For example, a single day where your product output is 3,500 should not elicit any drastic actions.
Understanding that your daily output is predictable helps you understand what next steps to take for improvement. You probably won't find any ideas for improvement by asking for explanations for the daily output changes within the Natural Process Limit lines. The predictable system here shows nothing but routine variations, otherwise known as noise, that does not require immediate action.
There are three scientific rules that help identify "signals" on an SPC chart. WorkClout helps identify these signals by automatically highlighting them for you on the SPC chart. These signals are also known as "exceptional variations" - data point(s) that are likely deviations from a predictable system that can potentially lead to a shift.
Signals identified by WorkClout help you distinguish from the noise on SPC charts so that you can save time and react to only the critical data points that can lead to improvement.
Rule 1 - Any data point that is outside of the Natural Process Limit lines.
Rule 2 - Eight consecutive data points are on the same side of the average line.
Rule 3 - Three or four consecutive data points that are closer to the same limit than the average line.
A data point outside of the limit (Rule 1) is a significant deviation outside of the predictable system; this means that there is almost certainly a root cause for the change that can be studied. This change does not necessarily mean that the deviation will go away or if it will be sustained over time.
A moderate but sustained effect, eight consecutive data points under or over the average line (Rule 2), means that there is most likely a change that has caused the product output to be above or below average consistently. This observation can be seen as an early indication that the predictable system will trend upward or downward in the near future.
A weak signal but sustained effect, 3-4 consecutive data points closer to one of the limits than the average line (Rule 3), means that there has been a recent change that has caused product output to be above or below average.
An alteration of the product output chart below shows all three of the rules - the three signals that can help you identify a predictable system change. Detecting a signal means that there may be a root-cause to be investigated before the signal's occurrence. One question to ask yourself when a signal is identified on WorkClout is, "What happened right before this signal occurred in my operations?".
If you find that the signal is moving the metric in a positive trend, you need to understand its cause and make sure it's not temporary. Finding the cause can also lead to a permanent increase in efficiency if you find that it should be applied as a new or updated SOP.
For example, perhaps after thorough root-cause investigation, you find that an employee has found a new and more efficient way of reducing the amount of time it takes for changeovers. You can now capture this new process and standardize it across the organization to improve overall product output.
Alternatively, if you find that the signal is moving the metric in a negative trend, you need to eliminate the cause and restore performance to the previously predicted levels using both corrective and preventive actions.
For example, you may find that new employees have not been consistent with how they perform the changeover leading to additional downtime after a root-cause investigation. Knowing this, you can now effectively address the issue and prevent the system from continuing its negative trend.
An alteration of the product output chart below shows the dates you might want to investigate and question any operations changes that might have occurred.
As a leader, it's essential to know when and why your KPIs are changing over time. Understanding what drives the increases and decreases in your key metrics gives you the power to focus energy on what matters most. Also, knowing how to detect signals derived from an SPC chart keeps from overreacting to the noise in everyday fluctuations in performance.
.png)
.png)
-min.png)
