Control charts

A control chart is a graph used to study how a process changes over time. Data are plotted in time order, with a central line (average) and upper and lower control limits (usually ±3 standard deviations) that show the expected variation of a stable process. It is one of the seven basic quality tools, developed by Walter Shewhart.

Parts

Value
                              * (out of control)
UCL ---------------------------------  upper control limit
      *       *
 CL --*---*-*---*---*---*--*---------  centre line (mean)
         *       *
LCL ---------------------------------  lower control limit
      1 2 3 4 5 6 7 8 9 10 ...  sample / time

Purpose

• Determine whether a process is in control (stable, with only random/common causes) or out of control (special causes present).
• Monitor the stability of project processes, such as defects per build, response time or schedule variance.
• Prevent defects by detecting trends early.
• Decide whether corrective action is needed.

Interpretation

• A point outside the control limits means an assignable cause that needs investigation.
• Rule of seven: seven consecutive points on one side of the mean show a non-random pattern (trend) even if they are within limits.
• Trends or cycles indicate a problem.
• Control limits differ from specification limits (the customer's requirement). A process may be in control but still not meet specifications.

Types

Steps to prepare

1. Select the characteristic to measure.
2. Collect samples at regular intervals.
3. Calculate the mean and standard deviation; set UCL = mean + 3σ, LCL = mean − 3σ.
4. Plot and interpret.

Example: A team plots the number of bugs found per release. A point above the UCL after a new library was introduced shows a special cause, which leads to a review of the library.