# Control Chart

**URL:** https://kaizumi.com/dictionary/control-chart

**Description:** A control chart is a statistical tool that monitors process variation over time, distinguishing between common cause and special cause variation.

**Category:** six-sigma

**Tags:** metrics, quality, visual-management

## Definition

A control chart is a time-series graph used to monitor process performance and detect changes. It plots data points over time against statistically calculated control limits (typically ±3 standard deviations from the mean). When all points fall within the limits in a random pattern, the process is "in control"—variation is due to normal, inherent causes. Points outside the limits or non-random patterns signal "special causes" requiring investigation. Control charts are fundamental to Statistical Process Control (SPC) and the Control phase of DMAIC.

## Examples

### Manufacturing

> A machining process monitors shaft diameter hourly. The control chart shows UCL of 10.06mm and LCL of 9.94mm. When two consecutive points exceed 10.05mm, the team investigates and finds a worn cutting tool—catching the issue before producing out-of-spec parts.

### Healthcare

> An ER tracks daily patient falls on a control chart. For months, data varied randomly around 0.8 falls/day within control limits. When a new flooring was installed, falls dropped to 0.3/day—a positive special cause. The improvement was verified as statistically significant, not random.

### Administrative

> A call center monitors average handle time weekly. The control chart shows stable performance at 4.2 minutes. When a new product launch causes a spike to 5.8 minutes (above UCL), management knows to investigate rather than treat it as normal variation.

## Key Points

- Control limits are calculated from process data, not from specifications or targets
- Points outside limits or patterns (trends, runs, cycles) indicate special causes
- Different chart types exist for different data: X-bar/R for continuous, p-chart for proportions, c-chart for counts
- Process must be "in control" before capability can be assessed

## Common Misconceptions

**Control limits are the same as specification limits.** Control limits show what the process IS doing; specification limits show what it SHOULD do. A process can be in control (stable) but still produce defects if control limits exceed specification limits.

**Any point outside limits means the process is broken.** Approximately 0.3% of points will fall outside 3-sigma limits by chance alone. Rules for detecting special causes consider patterns, not just individual points.

## Related Tool

Where a control chart tracks variation over time, a Yamazumi chart reveals how work-element times stack against takt time across operators — build one with the [operator balance chart builder](/tools/yamazumi).

Source: https://kaizumi.com/dictionary/control-chart
Licence: free to quote and cite with attribution to Kaizumi.