What is it?
A Control Chart shows how a measure behaves over time and adds statistically calculated limits. It helps distinguish the routine variation built into the process from unusual signals that deserve investigation.
What's it really for?
It is really a decision aid for knowing when to act and when to leave the process alone. That matters because reacting to ordinary variation usually adds more variation, while ignoring an unusual signal can allow a real change to pass unnoticed.
You know when…
Yesterday's result is worse, so somebody asks for an explanation. Today's is better, so the explanation is quietly forgotten. The team keeps adjusting the work in response to individual points and performance becomes more erratic. A Control Chart gives the movement some boundaries.
Have a go
- Choose a measure and chart type suited to the data. ☐ Plot observations in time order. ☐ Calculate the centre line and control limits from the process data. ☐ Apply the same data definition throughout. ☐ Investigate signals; improve the system when variation remains routine.
Why it works
Without limits, every rise and fall can look important. The chart establishes the range a stable process would normally produce, making sustained shifts, points beyond the limits and other unlikely patterns visible.
Look for…
Points beyond a control limit, sustained runs on one side of the centre line, steady trends and other patterns unlikely to occur by chance. Also look for a stable process that is consistently failing the customer's requirement.
Watch out…
Control limits are not targets or specification limits. Recalculating them whenever a point looks awkward defeats their purpose. A stable process can still be incapable, and an unstable process should not be described by a reassuring average.
Worth noticing…
The chart changes the question from ‘Why was Tuesday bad?’ to ‘What sort of variation are we looking at?’ That is usually a better place to begin.
Try next…
Use a Run Chart when a simpler time-series view is enough. Use Process Capability when a stable process must be compared with requirements. Use 5 Whys or Fishbone Diagram to investigate a genuine special cause.