Free interactive learning tool

The Process Behaviour Chart

Every month your organisation reads a number next to last month's number and writes a story about the difference. Two numbers are never identical, so the story gets written every single month — whether or not anything happened. Here is the one chart that tells you which months deserved a story.

Guess first — you are in the meeting, this is the pack

First-response time on new claims just went from 6.3 to 8.0 hours. That is a 27% jump, the worst month in over a year. The room wants to know what went wrong. Before you scroll: is this a real change, or is it nothing?

Mean first response, hours — as the monthly pack shows it
MonthHoursvs. prior
The chart — same numbers, with the limits drawn
Show the numbers as a table
The plotted series
Period Value Moving range Verdict
What the chart says
Draw the limits
Right now this is a run chart — a line with no reference points. Every wiggle looks like it might mean something, which is exactly the problem.

The arithmetic

Points
–
Centre line (mean)
–
Average moving range
–
Upper limit
–
Lower limit
–

limits = mean ± 2.66 × mR̄
mR̄ = mean of |each point − the one before|

Where this bites — the same habit, wearing different clothes
The monthly pack

A story for every month

Any report that shows this month beside last month manufactures a narrative on a fixed schedule, because two numbers never match. The commentary column is the problem — it has to be filled in whether or not anything happened.

Operations

Tampering

Adjusting a stable process in response to noise makes it worse, not neutral — you add your corrections on top of the variation that was already there. The shower with the slow water heater: turn, wait, scald, overcorrect, freeze.

People

Ranking the ordinary

League tables of teams or regions on a stable process rank them on noise. Someone is bottom every quarter and someone is top, reliably, with nothing to explain — and the performance conversation that follows is about a coin landing.

Improvement

Claiming a win

Intervene after an unusually bad month and the next month is usually better — regression to the mean, not your initiative. Without limits you cannot tell which changes worked, so the ones that didn't get rolled out everywhere.

A point inside the limits is the process being itself. Reacting to it adds variation rather than removing it — and it costs you the meeting, the review, and often someone's week. A point outside is worth your whole attention, because something genuinely changed. Learn to draw this by hand from any twelve numbers, and ask the question that ends most monthly commentary: "is that outside the limits?"