A run chart is a simple graph that plots data points over time, with a center line running through the middle, to reveal whether a process is improving, getting worse, or staying the same. It’s one of the most widely used tools in quality improvement because it turns raw numbers into a visual story about performance. If you’re tracking something like patient wait times, infection rates, or manufacturing defects week after week, a run chart helps you see patterns that spreadsheets alone can’t show.
How a Run Chart Is Built
A run chart has four basic components. The horizontal axis (x-axis) represents time, broken into whatever units make sense for your process: days, weeks, months, or individual measurements. The vertical axis (y-axis) represents the thing you’re measuring, such as minutes, percentages, or counts. Each data point is plotted where its time and value intersect, and those points are connected with a line so you can follow the trajectory.
The fourth component is the center line. This is typically the median of your baseline data, meaning the middle value when all your data points are arranged from lowest to highest. The median is preferred over the average because it isn’t pulled off-center by extreme outliers. If one week had a wildly unusual result, the average would shift, but the median stays stable. Half your data points should fall above it and half below, which is exactly what makes the interpretation rules work.
What a “Run” Actually Means
The term “run” refers to one or more consecutive data points that fall on the same side of the center line. If five data points in a row land above the median, that’s a single run of five. When a data point crosses from one side of the median to the other, a new run begins. The number, length, and pattern of these runs are what tell you whether your process is behaving randomly or something meaningful has changed.
Four Rules for Spotting Real Change
Random variation exists in every process. The challenge is distinguishing random noise from a genuine signal. Run charts use four probability-based rules to make that call.
- Shift: A run of consecutive points all on the same side of the median that exceeds a prediction limit (commonly six or more points in a row for typical sample sizes). This suggests the process center has moved.
- Trend: A sustained sequence of points continuously going up or continuously going down. Five or more consecutive increases or decreases is a common threshold. This signals a directional drift in performance.
- Too many or too few runs: If the data crosses the median far fewer times than expected, it suggests clustering, meaning the process may be shifting between two different levels of performance. Too many crossings can indicate other types of non-random behavior.
- Astronomical data point: A single value so far from the rest that it clearly doesn’t belong to the same process. This is the most subjective rule, but experienced teams can usually spot these immediately.
These rules were originally derived from probability theory developed by statisticians Swed and Eisenhart. The logic is straightforward: if a coin is fair, getting heads ten times in a row is technically possible but extremely unlikely. The same principle applies to data points falling on one side of the median. When the pattern is too unlikely to be random, something real is happening in your process.
Where Run Charts Are Used
Run charts are especially common in healthcare quality improvement. Hospitals use them to monitor infection rates on specific wards month over month, track patient length of stay in days, measure emergency department wait times, and follow discharge volumes. The Institute for Healthcare Improvement considers run charts one of the most important tools for assessing whether a change actually led to improvement.
Beyond healthcare, run charts appear in manufacturing, education, government services, and any field where teams need to know if a process is getting better after they’ve made a change. Their simplicity is a major advantage. You don’t need statistical software or advanced training to build one. A spreadsheet or even graph paper will do.
How Run Charts Differ From Control Charts
Run charts and control charts look similar at first glance, but they serve different purposes. A control chart adds two additional horizontal lines: an upper control limit and a lower control limit, calculated from the data’s statistical variation. These limits allow you to quantify how much variation is normal for your process and flag points that fall outside expected boundaries.
A run chart lacks these control limits entirely. It relies on the median and the four probability rules described above rather than on calculated boundaries. This makes run charts easier to create and interpret, which is why they’re recommended in the early stages of improvement projects when teams are still gathering baseline data and testing changes. Control charts become more useful later, when you need to monitor a stabilized process with greater statistical precision.
One important distinction: the four run chart rules (shift, trend, too many or too few runs, astronomical point) should only be applied to run charts, not control charts. Control charts have their own separate set of detection rules based on their control limits.
How to Create One
Start by defining what you’re measuring and what time intervals you’ll use. Collect your baseline data before any changes are made. Calculate the median of that baseline data and draw it as the center line. Then plot each new data point as it comes in, connecting them chronologically. Annotate the chart with notes about when you made specific changes to your process. These annotations are what transform a run chart from a passive graph into an active decision-making tool, because they let you see whether your changes coincided with shifts in the data.
The more data points you have, the more reliable your interpretation becomes. With too few points, the probability rules lose their power because short sequences can easily appear random. Most quality improvement guides recommend collecting enough baseline data to establish a stable center line before attempting to interpret patterns.
Why Run Charts Work for Improvement Teams
Run charts serve three practical functions. First, they help teams set goals by showing exactly how a process is currently performing. If your baseline chart shows average wait times of 45 minutes with wide swings, that’s a concrete starting point for an improvement aim. Second, they provide near-real-time feedback as you test changes. You don’t have to wait for a quarterly report to know if something worked. Third, they show whether improvements are lasting. A single good week might be random variation, but a sustained shift below the median after a process change is strong evidence that the improvement is real and holding.
The visual nature of run charts also makes them effective communication tools. Showing a team or leadership group a chart with a clear downward shift in infection rates after a hand-hygiene intervention is far more compelling than presenting a table of numbers. The pattern speaks for itself.