Articles by CBAssociates

Published, Newspapers and Technology, July 2001

DO THEY TALK TO YOU OR ARE THEY JUST PRETTY PICTURES?

"What is the purpose of this chart?" I asked a production manager, referring to a graph similar to Chart 1. "We keep track of our production," he said.

The chart, a basic Excel graph that can be automatically created, kept a faithful record of the inserter output for each day. "I see that it records the data, but how do you use it?" I asked.

As he proceeded to explain what each data point meant as if I did not understand basic charting, it became clear that we were not communicating.

"I understand how the chart was created and can read it, but what do you do with it?" I again asked. He again told me, in a manner suggesting that I do, in fact, need the hearing aid my family has suggested, "that it tracks our production so we know what we did."

After several more exchanges, he finally said, "We just put it on the bulletin board every day. We don't do anything else with it."

Whenever I see a simple graph that looks like this one, I'm fairly certain that it is not used on a daily basis and is just a pretty picture. These are called trend charts and are useful for tracking stock, but their value is limited in manufacturing operations.

It was not the production manager's fault that he did not use the information; it simply was not a useful operational tool. And there is no sense wasting very limited staff time making pretty pictures.

To make a chart useful, it needs additional information so it becomes more like a control chart. A simple method some people use is to draw a line across the graph representing the expected output. For example, the goal can be the average of the operation over time, as Chart 2 shows. By graphically indicating the goal, a person can very quickly determine if it has or has not been met.

In an inserting operation, however, this is only useful when the operation being tracked is consistent, for example the same size jacket and the same number and type of inserts. Because the number of inserts is not consistent throughout the week, Chart 2 reflects a seven-day up and down cycle.

Chart 2, as used for inserting, is misleading. It would drive the inserting manager crazy if the boss complained every time the goal designated on the chart was missed. In reality, production was not bad; the target was just wrong. If the inserting manager had an enlightened boss, he would receive an equal number of pats on the back as kicks in the bottom. This example illustrates how seemingly good tools can actually provide bad information.

Had the person who set up Chart 2 asked, "What is the purpose of the chart?" then made sure that it met those goals, he would have constructed it differently. Every chart needs a purpose.

If a chart's purpose is to show how output compares to expectations for the shift, then Chart 3 might be the result. The light line indicates the goal. The heavy line indicates actual output. At a glance, this graph shows the manager if the goal was met, exceeded, or missed, and by how much.

Charts used in manufacturing must include expectations. In most cases, there should be a different expectation for each different event. Startup waste for a 24-page and 64-page paper should be different. As a percentage of newsprint, they might be the same if the run lengths are the same.

Chart 3 became even more useful by changing the scale. The previous charts set the axis in increments of 1,000 copies, which makes it a bit hard to figure out the value of the data points. By changing the increments to 500 copies, it is easier to determine the number of copies.

What is the purpose of the chart? How will I use it? These two questions are the starting point for designing a truly useful tool.

A variety of charts can be used, such as rolling charts, control charts, charts with warning tracks, and charts using bars. They all have their place in the manager's efforts to cost and improve efficiency, but that is for another day.

Great questions elicit great answers, and the converse is true. The value of the answer is in direct correlation to the quality of the question. This is as true in life as it is in charting.

[HOME]