Patterns

A pattern is essentially a recurring “thing”. It could be behaviors, defects, markings, traffic, or anything else that can be observed or monitored.

The relevance to Lean is that the pattern is caused by something. Pure randomness is actually surprisingly uncommon in nature, and even less common in the workplace. Nearly everything has a cause if you look hard enough.

And in the case of a pattern, when the cause repeats over and over, it presents an outstanding opportunity to make an improvement. Because of their repetitive nature, they tend to be items that show up far to the left of Pareto Charts. That means that finding patterns of problems can give you the best bang for the improvement buck.

Some patterns are very easy to spot. If every fifth item has a defect, and there are five machines making the product, it doesn’t take rocket science to check if one of the machines is the problem.

Other situations, however, are much more difficult to uncover. In a work environment, there are often many things happening at the same time, so attention is divided. It can be extremely hard to correlate things together. For example, let’s say that Tuesday morning attendance is very poor in the fall in facilities on the west coast. It might be hard for some people to recognize that this is related to Monday night football.

For this reason, tracking data and using a variety of problem solving tools can help identify patterns far more readily than the naked eye can.

Pivot tables can help uncover linkages, but the right data must be plugged in, and the investigator still has to come up with a hypothesis about what is going on. Run charts are great at showing time-based patterns. Pareto charts show common problems that may be associated with a pattern of errors.

Broken patterns are also good indicators of problems. If a demand chart shows a repetitive pattern based on the day of the week, and suddenly Wednesday’s orders are way down, it bears investigating.

In this series of letters, which one would you take a closer look at:

ABABABABXABABAB

You would obviously look at the one that stands out. In a Lean environment, you also look for breaks in patterns that could indicate a change has occurred.

Words of Warning

• Investigate whenever a pattern appears or disappears. It means something has changed.
• Use data crunching tools to uncover the hidden patterns but remember that there is a cost to data collection. Forming educated guesses about what might affect a process can help reduce the data that you need to collect.