Plot your data in time order, draw the median line, and automatically test for non-random patterns — trends, shifts, clusters — that signal a special cause in your process.
Paste your data on the left and press Plot & Analyse to draw your run chart.
Watch: Run Charts: Spot Real Process Changes Fast
Is your process drifting, cycling, or just bouncing around? Enter the lab to plot data over time and spot the patterns.
Paste your measurements in time order (one per line or comma-separated). You can add an optional target line to see how the process compares to a goal.
The calculator plots your points, connects them, and draws the middle line (median). Points on the same side form a 'run'.
Four pattern tests check for trends (6+ in a row), shifts (8+ on one side), clusters, and unusual points — flagging anything that doesn't look random.
Use the calculator above to plot your process data over time, find the median, and automatically test for non-random patterns. A run chart is the simplest entry point to statistical process control — it tells you whether your process is behaving randomly or whether something has changed.
A run chart is a time-ordered line graph with a median drawn through it. It is one of the first tools to reach for when you want to understand process behaviour over time. Unlike a control chart it needs no distribution assumptions and can be built from as few as 10 data points. The four run tests tell you whether any patterns are statistically unlikely to have occurred by chance.
A run chart plots your data in time order with a median line through the middle. Each point lands above, below, or on that line, and the tool looks for patterns too orderly to be random chance. It checks four things: too few or too many switches across the median (clustering or oscillation), a steady trend of six or more points climbing or falling, a shift of eight or more points stuck on one side, and any obvious freak outlier. Any of these flags a special cause worth investigating; otherwise the process is just showing normal variation.
A team plots 20 consecutive daily measurements. The average looks stable — no alarm points. But a run chart analysis flags that 9 consecutive points are all above the median, far more than you'd expect by chance.
This is called a 'run' — and 9 in a row suggests the process has shifted to a new level rather than randomly fluctuating.
What to do with this: Look at what changed around measurement 10 or 11 — a new material batch, an operator change, a machine tweak. Run charts are great for catching slow trends that traditional control charts miss.
A run chart turns a column of numbers into a story. You can see at a glance whether a process is getting worse, improving, or simply bouncing around its average. That distinction drives very different management decisions — and stops people reacting to noise as if it were signal.
Use a run chart early in a DMAIC project to establish the baseline and track improvement. It is ideal whenever you collect data at regular intervals: daily defect counts, weekly cycle times, monthly complaint rates. If a run test fires, investigate the special cause before calculating capability or setting targets.
Run charts sit in the Measure phase of DMAIC as a simple baseline tool, and in the Control phase to confirm that improvements have held. A process with no run-test signals is said to be in a state of statistical control — the starting point for any meaningful capability analysis.
When a run test fires, mark the approximate time the pattern started and look for anything that changed — materials, people, machines, methods, or the environment. If it is a beneficial shift (improvement), lock in the change. If it is a deterioration, remove the cause and confirm the fix with a follow-up run chart. Once the process is stable, move to a control chart for ongoing monitoring.
Combine the run chart with a control limits calculator, Pareto analysis, and DMAIC structure for a complete data-driven improvement workflow.
A run chart is a line graph of data plotted in time order, with a median line drawn across it. It is one of the simplest tools in quality improvement because it shows whether a process is stable or whether something has changed over time. Run tests detect non-random patterns — trends, shifts, and clusters — that suggest a special cause is at work.
A run chart uses only the median to test for non-random patterns, whereas a control chart also calculates upper and lower control limits (UCL/LCL) based on the process variation. Run charts are simpler and require no assumptions about distribution; control charts are more sensitive to individual out-of-control points and are preferred for ongoing monitoring once a process baseline is established.
You need at least 10 data points to get meaningful run tests, and 20 or more to detect subtle patterns reliably. With fewer than 10 points, the number-of-runs test lacks statistical power. For ongoing process monitoring, collect data at regular intervals and add points as they come in rather than waiting to have a large dataset before plotting.
A run is a sequence of consecutive points on the same side of the median line. When a point lands exactly on the median it is excluded from the count. The number of runs expected by chance depends on the total number of plotted points; too few runs suggest a shift or cluster, too many suggest oscillation, and six or more consecutive points trending in one direction signals a trend.
A trend is six or more consecutive points all going in the same direction — all increasing or all decreasing. In a stable (random) process this is extremely unlikely to happen by chance, so a trend is strong evidence that something is systematically changing the process, such as tool wear, a gradual process drift, or an environmental shift.
A shift is eight or more consecutive points all on the same side of the median. Again, by chance this is very unlikely in a random process, so a shift suggests the process level has genuinely moved up or down. Common causes include a material change, a different operator, a machine adjustment, or an environmental change that was not noticed at the time.
An astronomical point is a data value that is so far from the others that any experienced observer would flag it immediately — it stands out visually. Unlike control chart signals, which use statistical limits, astronomical points are identified by visual judgement. They are usually caused by a measurement error, a data entry mistake, or a genuinely unusual event that needs investigating separately.
Using this calculator is the first step. The real value comes from applying improvement tools, structured training, and practical templates to fix the root cause and sustain the gain.
Explore more practical calculators designed to support improvement, quality, statistics, and operational decision-making.
Want to know how to use run charts to spot trends and shifts in your process? The Yellow Belt covers run charts, process behaviour, and the core improvement toolkit.
View Black Belt →