 |
Knowledge and Superstition
Jack Paully is the highly regarded Managing Director of Kokia, a Norwegian multi-national that produces hardware and software for telecommunications.
In Paully's elegant New York office, the following graph is on proud display. It shows the weekly sales trends for Kokia US.
*
Paully, who used to be head of Human Resources, is known for being fair but very demanding.
"Hello, Jack, how are things?" The unmistakable accent of Mark Carter, on a business trip to New York, made Paully look up from the report he was writing.
"Hey, what are you doing here?" Paully
and Carter had been students together when Paully spent a year at the London
School of Economics. They hadn't seen each other for a few years.
"I thought I'd drop in and give you a
surprise; my flight doesn't leave for a few hours. Got time for a
coffee?"
Carter, naturally curious, noticed the graph and asked, "Why do you have that graph up there?"
Antonia, Paully's secreatary, suppressed a smile; nobody had ever dared to ask.
"This graph tells me how things are going." Paully replied, somewhat condescendingly.
"And how are they going?" asked Carter, intrigued.
"Well, some weeks are better than others."
Paully was convinced of the importance of this graph. It was updated weekly. But it gave him no possibility to analyze the data and interpret them beyond a hunch.
The scope of data gathering is to increase the understanding of the phenomenon being examined; in order to have data that can be interpreted, it has to be presented in a way that allows us to
- read the data
- forecast trends
- make good decisions
Interpreting data is a process.
More generally:
A process is the transformation of an input (material resources, people, methods) into an output (result of the transformation).
"Mmmhh," was all the comment Carter had for Paully's statement that, some days, sales were better than others.
"What are you pulling that face for?" Paully seemed surprised by his friend's enigmatic expression.
"Jack, tell me something, what criteria do you use to make your decisions?"
"Aaachoo!" Paully blew his nose.
"Why do you ask, Mark? You've always got your head stuck in a book. Anyway, for your information, we have very detailed reports that tell me every month how the situation's developing. A company's not like university, you know. We can't afford to let things get out of hand."
Mark sighed. It was 4 in the afternoon and his flight left at 21.30. He decided to go for it.
"I'd like to see what those reports look like."
"Antonia," said Paully, lifting up the phone, "bring me the reports of the last few months, would you? I can see you've got time to kill."
"There you go." Paully pulled out the July
report and handed it to Carter.
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
*
"OK, let's have a look," Carter reached for his glasses. "So this is how you keep things under control, is it? Tell me, what have you decided to do this month?"
Paully couldn't help looking at him with an air of superiority.
"Well, first of all I look at the data
that show me excessive variation compared to the previous month…ok, look
here, for instance…the In-Process inventory in Department 17 shows an increase
of 42% compared to the plan; something must have gone wrong. We've
never recorded such a high value; it's 12 percent higher than last July.
Somebody in Department 17 will have to
explain what's going on; I think I know, but I want a plausible explanation
from the boys in 17. You see, people have to take responsibility
for things; that's the only way they can mature and be really involved
in their work.
Why are you pulling that face again?"
"Nothing, I was just listening. Jack, have you got the reports from the last few months?"
"Are you kidding? Of course I do. Otherwise how would I know the variation's not what it should be?"
"Do you mind if I take a look?"
Paully shrugged, "What's all this, Mark? You want to be a manager? You don't know how much I envy you. I mean your life's so peaceful, you teach your courses, you got your students, you go round the world to conferences. What do you care about the real world! Believe me, your much better off where you are."
Mark smiled.
"Here's the last 31 months. Satisfied?"
|
|
| Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec | |
| Year 1 | 19 | 27 | 20 | 16 | 18 | 25 | 22 | 24 | 17 | 25 | 15 | 17 |
| Year 2 | 20 | 22 | 19 | 16
|
22 | 19 | 25 | 22 | 18 | 20 | 16 | 17 |
| Year 3 | 20 | 15 | 27 | 25 | 17 | 19 | 28 |
Carter gave a quick look at the numbers. "Listen, Jack, would you be willing to hear me out on this?"
Paully knew that expression. Carter had been a friend for years. He was undoubtedly one of the people he had the most respect for. He'd just never been able to figure him out fully. But he was sure he was worth listening to.
"OK, I'm really curious to know what you're driving at."
"Right. First of all, let's try and organize these numbers in a different way. Let's put them on a graph, the way you do with sales, and draw a line in the middle at the height of the average value."
Carter pulled out a ruler and a calculator from his bag. A few minutes later he pushed the following graph over to Paully's side of the desk.
*
"Why did you draw a line at the height of the average value?" Paully asked.
"To see if the points plotted follow a particular trajectory. Just looking at it, I'd say it's all pretty regular.
Paully snorted. "The highest value is 28! That's exceptionally high, and there must be a reason."
"Well, if we want to see whether 28 is "exceptional" then we have to measure the variation from month to month."
"Eh?" exclaimed Paully.
"Yes, let's try and see if the difference between one month and another over the last 31 months shows a particular trajectory and…"
"Listen, Mark, if you're late you're going to miss your flight…"
"You said you were going to listen."
"Yes, sorry. It's just that…"
"It's easy, just one moment…there you are.
The procedure is simple, look at this table.
|
|
| Jan
|
Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec | |
| Year 1 | 19 | 27 | 20 | 16 | 18 | 25 | 22 | 24 | 17 | 25 | 15 | 17 |
| Differences | 8 | 7 | 4 | 2 | 7 | 3 | 2 | 7 | 8 | 10 | 2 |
The difference between January and February
in the first year is 8 (27-19), between March and February it's 7 (20-27
you only consider the differences in absolute numbers, not -7 but 7) and
so on. This way we get this series of values
|
|
| Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec | |
| Year 1 | 8 | 7 | 4 | 2 | 7 | 3 | 2 | 7 | 8 | 10 | 2 | |
| Year 2 | 3 | 2 | 3 | 3 | 6 | 3 | 6 | 3 | 4 | 2 | 4 | 1 |
| Year 3 | 3 | 5 | 12 | 2 | 8 | 2 | 9 |
If we put them on a graph, again with a line in the middle that represents the average value, it looks like this!
*
Paully seemed surprised. Even this graph
- Carter called it a "Moving Range" graph - didn't show any particular
trend.
When the graph of Individual Values and
the graph of Moving Ranges were put near each other, it looked like this:
*
*
"But this doesn't tell us a lot," Carter went on. "What we want to know is whether that 42% represents something odd, a special event, or is simply the basic "noise" of the process."
"I don't get it," Paully stammered.
"It's simple. Each of these numbers represents a magnitude that is the result of a measurement. What determines the result is a process behind it. In other words, if we want to know whether the result is odd or not, we have to find a way of interpreting the process that generated it. To put it more poetically, we have to listen to the voice of the process."
Paully looked interested. "Go on."
"To do this," Carter continued, "there's
a simple and very rigorous method. First you identify the "natural" limits
of oscillation in the process, starting from the results this process has
generated."
No sooner said than done. Carter pressed
a few keys on his calculator, and drew a couple of lines on the graphs
- one for Individual Values, and one for the Moving Ranges. The two graphs
looked like this:
*
"Everything's normal," Mark announced. "Your 42% is an absolutely normal value for this process. What I mean is that the amount of In-Process inventory is physiological. It's the system you work with that produces that result.
"Normal my a---!" Jack exclaimed. "I want to know why we have all these jumps."
"The "jumps", or as we should say, variation, are the result of the fact that, even if done in the same way by the same person, no two things will ever be identical. This means that the process that generates them will always be affected by variation. ALWAYS. The only reasonable thing we can ask about a process is if the variation is intrinsic to the process itself, or if it is caused by some special event. If the process oscillates within its natural limits, then we say it's in statistical control. If it goes outside the limits, we say it's out of control."
"Let me check if I've got it." Paully loosened his tie and undid his top shirt button. "You're telling me that the process that determines the amount of In-Process inventory in Department 17 can oscillate between 0.85 and 3.16 tons. That means next month could show a value anywhere between 0.85 and 3.16?"
"Exactly. You see, it's not difficult."
"But this is horrendous. This variation is excessive. We have to get it down. Tomorrow I'm going to call a meeting with the people from Department 17."
"They can't do anything about it."
"What do you mean?"
"This variation is determined upstream, by the general planning, production and distribution system. In other words, the variation is provoked by the company system as a whole."
"I'm the Managing Director, are you saying it's my fault?"
Carter stood up and grabbed Paully's hand to shake it. "My dear Jack, you should see the Millenium Dome they're building in London it's extraordinary."
"Don't change the subject. And I want to know how you calculated those limits."
"Antonia, it's been a real pleasure," Carter shook Paully's secretary's hand with a quick bow. "Bye, Jack. If I don't go now I'm going to miss that plane. Keep in touch."
"But what about those limits…"
A few seconds later the door opened again.
It was Antonia. "Wow, that is one charming man…"
*The tables and charts
in this article come from Don Wheeler's "Understanding Variation", SPC
Press, 1993
copyright 1999 mst-usa LLC