Six Sigma for the Apparel Industry, Richard ATWELL


Six Sigma for the Apparel Industry
Richard ATWELL

 

The trade agreements are passing, the buzzwords are flying, and every manufacturer is looking for the next certification to use as a badge of honor. And so enters “Six Sigma” to the rescue. The Superhero is really the mild mannered Clark Kent in Guatemalan manufactured tights and cape. The “supercharge” comes from computer software that allows the average apparel engineer or quality manager to perform analyses that previously were done only by college educated statisticians. Take away the computer and there is a solid basis of statistical methods that have been in the “bag of tools” for decades.


The prevailing motto for Six Sigma is “Perfection is Possible”. The subtitle would be “Even when things go wrong, the product is still good”. There are two major qualifiers to the concept of perfection. The major concept is that perfection specifically means that the product meets the customer's accurately determined specification. Therefore perfect production of a size “large” shirt, which has a specification of 42 inches +/- ¾ inch, does not mean that all shirts produced measure 42” but rather that all shirts measure between 41 ¼ inches and 42 ¾ inches. The second major concept of Six Sigma is that the variation of the output is very slight. Thus, when size “large” shirts are produced the measurements of multiple shirts will be so closely gathered around 42 inches that when something goes wrong, such as bad cutting or a gauge being misadjusted, the end result is that shirts still measure within the tolerance.


The rest of the Six Sigma process is understanding the statistics behind the charts and controls, learning the methods to determine what is working properly, and how to identify and fix problems that do occur. The concept as a statistical term has been frequently used since the 1930's and gained popularity in Japan and the USA beginning in the 1950's with Crosby, Deming, and Juran. However, Motorola is credited, by most, for bringing the term to the general public in the 1980's. Oh yes, the statistical term. Six Sigma manufacturing says that the variation in the product is so slight that 99.73 % of the output falls within the tolerances. Actually, the quality is so good, and multiple measurements so closely centered around the specified number, that even when the measurement is off by three times the standard deviation, in either direction, it is still within tolerance. Thus the measurements fall within Six Sigma or six standard deviations of the mean and the quality yield is 99.73%.


Because many companies are seeking badges or medals to show to prospective clients, frequently the managers are disturbed that Six Sigma certification is for individual people rather than companies. The green belt and black belt designations are awards for proficiency in using the tools of the program. The Six Sigma methodology is a structured program for improving quality through a process with the acronym of DMAIC. Define, Measure, Analyze, Improve, and Control. Portions of each of these steps can be performed using manual techniques that are found in any statistics textbook. The computer programs included in many of the certification classes provide much more powerful tools and the capability of calculating in minutes what might take days to calculate manually.


Convincing apparel executives to accept the process, other than as a selling tool, has been a difficult task. About the only agreement reached is that every garment can NOT be manufactured to the perfect specification. However, the typical manufacturing plant is producing apparel at about a 3 sigma level with 2.5 to 4 % defects. That is 4 defects per 100 not the 3.4 defects per 1,000,000 produced by a Six Sigma manufacturer. The gap is wide enough that significant improvement can be made in any such plant. Two financial facts are important to note. Historical studies have shown overall savings in the $10,000 to $20,000 range for an improvement of just one Sigma. Apparel managers are generally astounded by the “True Cost of Quality” in their manufacturing facility. Typically, these costs are hidden in overhead but include inspection and marking, sorting, transport, reinspection, supervisor time, downstream operator repair, cleaning, and irregulars. Therefore, any improvement in quality has a triple effect of reducing indirect labor, lowering total fabric cost, and improving customer satisfaction. Direct Labor Productivity will normally show a 10 to 15% improvement due to less repair time, correction of imbalances, and higher throughput.


Within the DMAIC process, the management team must Determine the actual quality requirement through discussions with the end customer. Such determination must eliminate the “Give them an inch- they will take a mile” attitude. Record the needed result, train both operators and auditors, and then hold the appropriate people accountable.


The measurement and analytical processes use tools and charts that are already available in many plants. These include some capabilities within basic spreadsheet computer programs such as Check Sheets, Histograms, Pareto Charts, and Scatter Diagrams. Flowcharts and Fishbone diagrams are easily drawn by hand.


The actual analysis and improvements come through human interpretation of the information and a logical thought process. There are two areas within this section that can be greatly assisted by the newer computer programs available. The first item is the determination of whether changes observed are statistically relevant, thus avoiding unnecessary adjustments to a existing process. The second area, with less use in the apparel industry, is the Design of Experiments capability to see what combination of changes will give the best end result. Experienced and motivated people given the time for thought and experimentation are the critical resource for analysis and improvement.


Controlling the change is often a matter of “What gets Measured and Posted is What Improves”. While computer programs can calculate the control limits required to know when a process is out of control, the limits have less value in an environment utilizing human operators. Manual calculations and charting will allow quick recognition of changes in the average apparel plant.


Should a company have people certified as Green Belts or Black Belts? The answer is “YES, IF”. If the company needs the computer program, needs a show of management support for Quality Improvement, or lacks personnel with Engineering or Quality Management training.


However, any company can reap significant financial benefits from a focused quality program utilizing the statistical tools that have been available for years. The requirement is to provide a motivated team of people with a structured process, an agreed upon set of standards, an expectation of continuous improvement, and then hold them accountable.