Six Sigma
Last Word


Six Sigma and Beyond
Thomas Pyzdek

Real-Life Six Sigma

These synopses of actual Six Sigma projects provide useful examples.

For many people, it's easier to learn by reviewing examples than by reading generalized descriptions. So in this column, we'll look at several quick overviews of Six Sigma projects to provide you with an idea of what Six Sigma deployment looks like at the project level. Of course, these synopses skip over significant factors, such as the leadership and infrastructure necessary to make the projects succeed. It's not that these factors are unimportant; to the contrary, they're critical. For additional information on the critical success factors, you may wish to review my earlier columns at www.pyzdek.com/pdf.htm and www.qualitydigest.com , or in my book The Six Sigma Handbook (McGraw-Hill, 2001).

  Accounts receivable—The Six Sigma team was tasked with improving the accounts receivable department's collection process. The project sponsor was the CFO, and the top-level dashboard item that generated the project was improved cash flow. The team decided to use the average age of uncollected accounts on the last business day of the month as their metric. Using X-bar charts because the histogram showed a very non-normal pattern, the team determined that the process was in statistical control with a mean of 57 days. They made a flowchart of the as-is AR collection process and used it to guide an observational study.

 The team noted and corrected several discrepancies, and several obviously stupid things were changed. For example, a team member from billing asked why the term "Net 30 Days" was used. An experiment was conducted where the term was changed to "Due on Receipt" for a random sample of invoices. The results showed that the average time to collect for the experimental group was 45 days, vs. 57 for the control group. The difference was highly significant, both financially and statistically.

 Next, the team contacted randomly chosen customers who had paid late and asked why they had been late. Fully 70 percent of the reasons for late payment were factors under the company's control (e.g., invoice errors or the bill being sent to wrong address). The team constructed a Pareto diagram and set about correcting the biggest problem areas. Within six months, the average age of uncollected invoices dropped to 37 days. The resulting savings were substantial.

  Printed wiring board components—The Six Sigma team received its project from the material review board. The MRB identified the project as a significant and chronic contributor to the problem of failures at final product test. The assembly was a complicated piece of hardware and final test failures caused shipping delays, resulting in penalties and loss of customer goodwill. The team's project focused in the PWB assembly area. There were three major subprojects: errors at manual insertion, errors at automated insertion, and errors at semi-automated insertion. A few examples of the issues addressed include kitting errors, the layout of the manual insertion workstation, the positioning of axial lead parts on the automatic insertion machine's parts tape, and the speed at which semi-automated insertion was performed. Problems were prioritized and addressed, leading to dramatic reduction of test failures.

  Cycle time—The team was chartered by a program manager to help the company introduce new programs more quickly. The company would often introduce a new design into manufacturing only to find that it couldn't be produced, which resulted in quality and schedule problems. The team's project—one of several—involved establishing the capability of complex numerically controlled machining equipment. This was important because the company manufactured a tremendous variety of complex parts in very low volume. Standard SPC was difficult because production runs were both short in duration and small in quantity. The Six Sigma team wanted to develop the ability to determine in advance if a particular engineering design could be produced at all and, if so, which CNC machine should produce it. To solve their problem team members designed a special test part that put each CNC machine through a complete series of tasks. The parts were then inspected and the results used to determine machine capability for each type of machine movement (e.g., drilling small holes, milling a surface or machining a groove). This data was used to evaluate proposed engineering designs for manufacturability, for make-by decisions and to select CNC machines to produce specific parts.

  Injection molded parts—The Six Sigma team was chartered to evaluate a problem with field failures of molded plastic parts. Members began by replicating the problem with production parts. The problems were resolved within a few weeks when the team identified a new process as the cause of the core problem. The process mixed two different plastic components at the injection-molding machine, as opposed to the single hopper and pre-mixed material of the previous process. If not properly mixed, the carbon black component would stratify and the product would fracture at the stratification when exposed to low temperatures.

  Wire bond—The Six Sigma team received its project from a senior executive who had received direct communication from an important customer who was upset about wire-bond failures of a particular critical part (a thick-film hybrid microcircuit.) The project had two major areas of focus: the wire bonding process and the testing process. Problems addressed included the metallurgy of the gold wire; the preparation and set-up of the process; the dressing of the tip on the machine tool; pressure, time and other settings; the hook used to pull the wire; the angle of the pull; the rate at which force was applied.

  Purchase order process—A Six Sigma team was chartered to streamline the process of obtaining a purchase order. The process took six weeks, creating delays and customer dissatisfaction. The Six Sigma team created a process map and used historical data to show the time taken by each step. By hand-carrying 10 PO requests through the process, the team was able to determine that, more than 99 percent of the time, a PO request spent in the system was nonvalue-added time—mostly waiting time. The team subdivided the project according to the type of PO being requested. Members were able to eliminate PO requests completely for a common type of PO and dramatically reduce processing time for the others.

  Etched circuit boards—The Six Sigma team was directed by a senior executive to solve the problem of photoresist breakdown. This problem occurred at the very end of a long sequence of process steps, which produce a bare printed wiring board. The problem was sporadic, and when it occurred, it resulted in delays throughout the production process. This wreaked havoc with schedules and resulted in extensive overtime work, shipping delays, penalties and angry customers. Through data mining, the team was able to focus the project on work which took place in the "yellow room," where the photoresist was applied. The project eventually focused on the settings of the lamination and the expose processes. The root causes of the problems were identified, and the problem was completely eliminated.


About the author

 Thomas Pyzdek is a consultant in Six Sigma. He has written more than 50 books, software and training products, including The Six Sigma Handbook (McGraw-Hill). Learn more about Six Sigma at www.pyzdek.com . E-mail Pyzdek at Tom Pyzdek .


Menu Level Above

This Menu LeveL

Menu Level Below

[Contents] [News] [WebLinks] [Columnists]
[Harrington] [Crosby] [Six Sigma] [Godfrey] [Townsend] [Marash] [Last Word]

Copyright 2000 QCI International. All rights reserved.
Quality Digest can be reached by phone at (530) 893-4095. E-mail:
Click Here

Today's Specials