What Is a Run Chart? Definition and Uses

A run chart is the simplest useful picture of a process: measurements plotted in time order, usually with a center line drawn at the median. No control limits, no constants, no assumptions about distributions — just the data, in the order it happened. It is the first chart most improvement teams should draw and, because it demands nothing but a pencil, the one with no excuse for skipping.

How it’s used

The run chart’s power comes from a basic fact: if a process is not changing, points should fall on either side of the median at random, like coin flips. Patterns too orderly for coin flips are evidence of change. The commonly used conventions: a shift is six or more consecutive points on the same side of the median; a trend is five or more consecutively rising or falling points; and too few or too many crossings of the median also betray non-random behavior. Each rule is simply a pattern unlikely to occur by chance in a stable process.

Because it needs little data and no software, the run chart shines early in a project — in the Measure phase, when the team has a few weeks of history and needs to see the process before modeling it. It also communicates: anyone can read one without training.

A worked example

A café tracks the daily average time from online order to handoff, plotting each day on a whiteboard chart with the median drawn in. Six weeks in, the owner redesigns the menu. Over the following two weeks, nine consecutive daily points sit above the median — a run far too long to be chance. The chart cannot say why the new menu slowed service, but it establishes that service genuinely slowed, when the slowdown began, and that the effect persists. The investigation that follows starts from evidence instead of impressions.

• Keep strict time order — sorting or grouping the data destroys everything a run chart can tell you.
• Annotate known events directly on the chart: changes, holidays, new staff. Context turns patterns into explanations.
• Give the chart enough history before reading patterns — around twenty points is a sensible floor.
• Do not paste a specification line on it and call it a control chart; without calculated limits it cannot separate signal from noise point by point.
• Graduate to a control chart when decisions start riding on single points — that is what control limits are for.

Before you compute anything, plot the data in time order. Most process stories are visible to the naked eye.

The run chart is a foundational tool, taught in our Yellow Belt program ($129, about 14 hours) and used constantly at Green Belt, where it grows into full control charting. Our free White Belt course shows where it sits in the improvement toolkit — a reasonable place to start, at no cost.