Read an Excerpt
Six Sigma DeMYSTiFieD
By Paul Keller The McGraw-Hill Companies, Inc.
Copyright © 2011The McGraw-Hill Companies, Inc.
All rights reserved.
ISBN: 978-0-07-176082-9
Excerpt
CHAPTER 1
Deployment Strategy
What Is Six Sigma?
Sigma (σ) is the Greek letter used by statisticians to denote the standard deviation for a set of data. The standard deviation provides an estimate of the variation in a set of measured data. A stated sigma level, such as Six Sigma, is used to describe how well the process variation meets the customer's requirements.
Figure 1.1 illustrates the Six Sigma level of performance for a stable process. The process data are represented by the bell-shaped distribution shown. Using the calculated value of the standard deviation (σ), the distance from the process centerline to any value can be expressed in sigma units. For example, consider the teller station at a bank whose average customer wait time (or time in queue) is 7.5 minutes with a standard deviation of the wait time calculated as 1 minute. Six standard deviations, or 6?, from the average is 1.5 minutes (in the negative direction) and 13.5 minutes (in the positive direction).
Separately, through the use of customer surveys, focus groups, or simple feedback, customer requirements may have been established for the process. In this case, the process is likely to have only an upper specification limit defined by the customers; there is no minimum limit to desirable wait times.
If this upper specification coincides exactly with the plus 6σ level (i.e., 13.5 minutes), then the process is at the Six Sigma level of performance. The implication is that the customer wait time will exceed the customer requirements only a very small percentage of the time. Similarly, if the maximum allowable customer wait time is 10 minutes, then the process would be operating at only a 2.5σ level of performance, indicating an increased risk of customers exceeding this maximum wait time.
Although the normal distribution tables discussed later in this text indicate that the probability of exceeding 6 standard deviations (i.e., z = 6) is two times in a billion opportunities, the accepted error rate for Six Sigma processes is 3.4 defects per million opportunities (DPMO). Why the difference? When Motorola was developing the quality system that would become Six Sigma, an engineer named Bill Smith, considered the father of Six Sigma, noticed that external failure rates were not well predicted by internal estimates. Instead, external defect rates seemed to be consistently higher than expected. Smith reasoned that a long-term shift of 1.5σ in the process mean would explain the difference. In this way, Motorola defined the Six Sigma process as one that will achieve a long-term error rate of 3.4 DPMO, which equates to 4.5 standard deviations from the average. While this may seem arbitrary, it has become the industry standard for both product and service industries.
These concepts have been applied successfully across a broad range of processes, organizations, and business sectors with low and high volumes and millions or billions in revenue and even in nonprofit organizations. Any process can experience an error, or defect, from a customer's point of view. The error may be related to the quality, timeliness, or cost of the product or service. Once defined, the Six Sigma techniques can be applied to methodically reduce the error rate to improve customer satisfaction.
Using the curve shown in Figure 1.2 (Keller, 2001), any known process error rate can be converted directly to a σ level. Most companies, including those with typical total quality management (TQM)-type programs, operate in the 3σ to 4σ range based on their published defect rates. In Figure 1.2, airline baggage handling, order processing, tech center wait time, and flight on-time performance fall in the general area from 3σ to 4σ
Notice that the y axis, representing DPMO, is logarithmically scaled. As sigma level is increased, the defects per million opportunities decreases exponentially. For example, in moving from 3σ to 4σ, the DPMO drops from 67,000 to 6,500 and then to just over 200 at 5σ.
Moving from left to right along the curve in Figure 1.2, the quality levels improve. Companies operating at between 2σ and 3σ levels cannot be profitable for very long, so, not surprisingly, only monopolies, government agencies, or others with captive customers can afford to operate at these levels.
It's clear that significant improvement in customer satisfaction is realized in moving from 3σ to 4σ. Moving beyond 4σ or 5σ involves squeezing out every last drop of potential improvement. Six Sigma is truly a significant achievement, requiring what Joseph Juran termed breakthrough thinking (Juran and Gryna, 1988).
There is some criticism of the DPMO focus, specifically with the definition of an opportunity. In counting opportunities for error in a deposit transaction at a bank, how many opportunities are there for error? Is each contact with a customer a single opportunity for error? Or should all the possible opportunities for error be counted, such as the recording of an incorrect deposit sum, providing the wrong change to the customer, depositing to the wrong account, and so on? This is an important distinction because increasing the number of potential opportunities in the denominator of the DPMO calculation decreases the resulting DPMO, increasing the sigma level.
Obviously, an artificially inflated sigma level does not lead to higher levels of customer satisfaction or profitability. Unfortunately, there will always be some who try to "game" the system in this manner, which detracts from the Six Sigma programs that estimate customer satisfaction levels honestly.
Since DPMO calculations can be misleading, many successful Six Sigma programs shun the focus on DPMO. In these programs, progress is measured in other terms, including profitability, customer satisfaction, and employee retention. Characteristics of appropriate metrics are discussed in more detail later in this section.
The financial contributions made by Six Sigma processes are perhaps the most interesting to focus on. The cost of quality can be measured for any organization using established criteria and categories of cost. In Figure 1.3, the y axis represents the cost of quality as a percentage of sales. For a 2σ organization, roughly 50 percent of sales is spent on non-value-added activities. It's easy to see now why for-profit organizations can't exist at the 2σ level.
At 3σ to 4σ, where most organizations operate, an organization spends about 15 to 25 percent of its sales on "quality-related" activities. If this sounds high, consider all the non-value-added costs associated with poor quality: quality assurance departments, customer complaint departments, returns, and warranty repairs. These associated activities and costs are sometimes referred to as the "hidden factory," illustrating the resource drain they place on the organization.
For many organizations, quality costs are hidden costs. Unless specific quality cost identification efforts have been undertaken, few accounting systems include provision for identifying quality costs. Because of this, unmeasured quality costs tend to increase. Poor quality affects companies in two ways: higher cost and lower customer satisfaction. The lower satisfaction creates price pressure and lost sales, which results in lower revenues. The combination of higher cost and lower revenues eventually brings on a crisis that may threaten the very existence of the company. Rigorous cost of quality measurement is one technique for preventing such a crisis from occurring.
It's not uncommon for detailed quality audits to re
(Continues...)
Excerpted from Six Sigma DeMYSTiFieD by Paul Keller. Copyright © 2011 by The McGraw-Hill Companies, Inc.. Excerpted by permission of The McGraw-Hill Companies, Inc..
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.
