Department of Software Engineering

1

Six Sigma

Nicholas Blank

Department of Software Engineering

University of Wisconsin – Platteville

Platteville, WI 53818

blankn@uwplatt.edu

Abstract

Six Sigma is a business management strategy aimed to improve a process. Originally developed for assembly lines, Six Sigma has been adapted to improve any type of process. This process was created in the 1980s by Motorola to address and improvethe quality of their products. A process that is meeting Six Sigma standards will produce no more than 3.4 defects for every million opportunities. People who follow Six Sigma have two processes to analyze their process. There is amethodology for improving an existing process and anotherfor beginning a process. Six Sigma uses a certification system that utilizes the belt color in martial arts. This paper will be a look into how Six Sigma works, benefits, criticisms, and how it relates to software development.

Introduction

Six Sigma

Six Sigma is a business management strategy originally developed by Motorola. Through minimizing variability in manufacturing and business processes and removing causes of defects, Six Sigma seeks to improve the quality produced by a process. By using statistical methods, Six Sigma uses a set of quality-management methods and designates an infrastructure of people who understand and implement Six Sigma.

Overview of Six Sigma

The term six-sigma is used in statistical process control and is more recently known to be a wide-ranging set of databased process improvement techniques. Six Sigma is named due to its mathematical background and goal of process variation within six standard deviations. A standard deviation is a parameter that characterizes a set of measurements, such as an average could be a characteristic to describe a set of data. Sets of data can be modeled with the normal distribution or bell curve. One standard deviation is equivalent to about two thirds of all values in a normal distribution. This standard deviation is centered on the average of the curve. A larger coefficient in front of the sigma means that more of the distribution is included and the more mature the manufacturing process or fewer defects the process has. Six Sigma is about measuring defects where a process in the six-sigma level will yield results within the tolerance interval of 99.99966%. This equates to 3.4 defects per million opportunities[2].

History of Six Sigma

Six Sigma started as a set of practices designed to improve manufacturing processes and eliminate defects. Application of Six Sigma has been extended to other types of business processes as well. Six Sigma was started development at Motorola in the 1970s, evolving from senior executive Art Sundry’s criticism of the company’s low quality. Due to this criticism, the relationship between increases in quality and decreases in cost were discovered. During this time, it was always believed that a high quality product wasmore expensive to produce. Most cost savings were seen with the cost of repair or control shrinking. Bill Smith formulated the particulars of the methodology at Motorola in 1986[1].

Roles of Six Sigma

Six Sigma utilizes a martial arts convention for naming its professional certified roles. Leaders and Champions usually receive high-level training on the technical aspects of Six Sigma and specific training on how to lead an initiative. At each belt level, a candidate is assigned an initial training project that he or she will work on during the formal training period.

Yellow Belts

Yellow Belts are given as a title to anyone involved in a Six Sigma project. A Yellow Belt is only required to be introduced to the main methodologies of Six Sigma. The knowledge required for this level of Six Sigma is approximately equivalent to reading this paper [1].

Green Belts

After personnel have gone through some formal training in Six Sigma, they become a Green Belt. This part-time professional participates on a team with a Black Belt team lead. This is the first level of certification and consists of a two-week course in the Six Sigma methods and basic statistical tools [1].

Black Belts

Black Belts are full-time professionals who act as team leaders in a Six Sigma project. If the project is small enough, Black Belts will be the project lead. Their training is four to five weeks long in methodologies, statistical tools, and team skills. This is additional training after the professional has achieved Green Belt certification [1].

Master Black Belts

The Master Black Belt is the highest certification that is offered. These professionals are highly experienced and successful in Six Sigma and have managed several projects. Master Black Belts are experts in all areas of Six Sigma methods and tools. Similar to most experts, Master Black Belts are responsible for the training of all other belts and to keep Champions and the project on track [1].

Deployment Champions

A Deployment Champion supports the deployment strategy within a line of business, customer segment, or core process teams. This position is commonly filled by a Master Black Belt, but Six Sigma does not require any certification to fulfill this role. Depending on the size of the company, there may be only one Deployment Champion or a senior deployment Champion with multiple Deployment Champions reporting to that senior person [1].

Project Champions

Project Champions are chosen from managerial leadership and will ensure that organizational systems are in place to support the Six Sigma initiative. Project Champions are strong advocates of the Six Sigma program. Due to their role in management, they reveal the benefits of Six Sigma to the organization [1].

Using Six Sigma

Six Sigma Methodologies

A Six Sigma process that will only have 3.4 defects per million opportunities will have employed one of two methodologies. These two methodologies are DMADV and DMAIC. DMADV is an acronym for define, measure, analyze, design, verify and is used for a process that is beginning with Six Sigma. DMAIC is an acronym for define, measure, analyze, improve, and control and is for a process currently in place wishing to improve by using Six Sigma. Some of the similarities between the two methodologies are that they are both data intensive. Intuition is eliminated and only facts remain. Both are implemented by trained professionals at the Green Belt, Black Belt, and Master Black Belt certification levels.

DMAIC

DMAIC is the method to improve quality of an existing process. This five-step process increases a product or process quality to the level of six sigmas. The DMAIC method should be used over the DMADV method when a product or process is in existence at your organization but is not meeting customer specification or is not performing adequately [2].

Define

The first step in the Six Sigma DMAIC process is to define the project goals and customer deliverables. In this step, the critical-to-customer and critical-to-quality requirements are defined within the scope of the Six Sigma project. Team members in this stage will develop the problem statement, goals, and benefits. Also, in this step, the key organizational support is identified and evaluated. This includes the needed resources being identified. Roles and titles of personnel are identified in the define stage to include the champion, process owner, and team members. By the end of this phase, it is important that the team has accomplished the development of the project plan, milestones, and the high-level process map. A couple of tools that are commonly used in this step are project charter, process flowchart, stake-holder analysis, and voice of the customer [1].

Measure

The second step in the Six Sigma DMAIC process is to measure the current process and determine current performance. This is where the Six Sigma team will start to quantify the problem. In this step, the definitions of what a defect, opportunity, unit, and metrics are identified by the Six Sigma team. Theses definitions are then used to develop the data-collection plan. Throughout the collection of the data, the validation of the measurement system is evaluated. Also, in this step, the initial sigma level of the process is evaluated. Some of the tools that Six Sigma professionals use in this step include reevaluation of the process flowchart, a data collection plan, benchmarking, and process sigma calculation [1].

Analyze

The next step in DMAIC is to analyze and determine the root causes of the defects in the process or product. This is the step where the Six Sigma team will define the performance objectives. Another important task is to identify value or non-value process steps. This is where knowledge of Lean, a production improvement practice that essentially helps simplify a process, is applied. The team will also identify the sources of variation in the product or process. This helps in determining the root causes of defects. In this step, a y = f(x) relationship is constructed. This is an overall idea to identify cause and affect relationships where y is the defect in the process or product and each x represents things that cause or influence the defect. Some of the tools that Six Sigma teams utilize in this step are the histogram, pareto chart, run chart, scatter plat, regression analysis, cause and effect diagram, and hypothesis testing. This step utilizes statistical analysis to analytically find the problems in the process or product [1].

Improve

The fourth step in the Six Sigma DMAIC process is to improve the process by eliminating defects. Part of this step is for the team to develop potential solutions to the causes of the defects in the process or product. One of the areas that products or processes can be improved is tolerances. In this step, the potential system-operating tolerances are defined. Furthermore, the failure modes or shortcomings of potential solutions are evaluated. The potential solutions then need to go through validation through the use of pilot studies. The potential solutions should be corrected and then reevaluated until the potential solution is a benefit to the process. Some tools that Six Sigma professionals are using in this step are brainstorming, mistake proofing, pugh matrix, failure modes and effects analysis, and simulation [1].

Control

The last step in the Six Sigma DMAIC process is to control the future process performance. The Six Sigma team will define and validate the monitoring and control system. The standards and procedures of the improvements are defined in the control step. Also, the team is responsible for the determination and implementation of the statistical process control. Another aspect about a process or product in the control step is for the team to determine the process capability after the changes have been implemented. At the end of this step, the Six Sigma team will verify benefits, cost savings, profit growth, and document the process. Some of the tools that the team will utilize during this step are the process sigma calculation, control charts, cost savings calculations, and the control plan [1].

DMADV

The common five-step process to start a process with Six Sigma in mind is DMADV. This process is used when a product or process is not in existence at an organization and one needs to be developed. This is a very common methodology when professionals are talking about design for Six Sigma (DFSS). DMADV can also be used when the existing product or process exists and has been optimized using DMAIC and still does not meet the level of customer specification or Six Sigma level. This most commonly occurs when a project was planned for DMAIC and incremental process improvement but really should have been a DMADV methodology improvement. The first three steps of DMADV and DMAIC are the same [2].

Design

The fourth step in Six Sigma DMADV is design details, optimize the design, and plan for design verification. Being the first step that differs from DMAIC, at this point, the data collected will assist in the design of the process to meet customer demands. Some tools that are used during this step are specification limits, simulation model, and the measurement and control plan[1].

Verify

The last step in the DMADV methodology is to verify the design, set up pilot runs, and implement the production process. Six Sigma professionals will have to make sure that the process is meeting standards. Also, the training of the personnel that will maintain the process is conducted during this step. As with most successful processes, documenting the process is extremely important. The tools commonly used during this step are pilot runs and documentation [1].

Results of Sigma

Benefits of Six Sigma

The biggest benefit to the organization that implements Six Sigma falls under profits. Black Belts are estimated to save a company approximately $230,000 per project and can complete four to six projects a year[1]. For example, after the first five years of implementation, General Electric has seen over ten billion dollars in benefits. Raytheon implemented Six Sigma in 1999 and has seen 1.8 billion in gross financial benefits, 500 million improvements in operating profit, and generated $865 million in cash flow[1]. The other added benefit is the increase in customer satisfaction.

Criticisms of Six Sigma

As with any system,Six Sigma does have some weaknesses. Some of these problems deal with the demand for aggressive performance tracking and accountability that are required for results. To further the problem of data intensity, employees need to be willing to learn and use statistical tools. After the technical side of the problems with Six Sigma, there are still issues with theoverselling of Six Sigma. There are consulting firms that are claiming to be masters at Six Sigma without any true experience implementing the methodologies. The next criticism is the cost of training. For Master Black Belt certification, Motorola University charges $17,000. This is crippling to a small company to train so many employees to carry out Six Sigma. The last criticism is that thesix sigmas of deviationarearbitrary. Why so many? Why not increase it to eight sigmas? Some industries, such as pacemakers, need a higher standard than Six Sigma [4].

Six Sigma and Software

Differences between software and manufacturing

In software development, the difference is that process variation can never be eliminated or even reduced below a moderate level. Something is wrong with software development when a team is making the same module repeatedly. Another ideal that Six Sigma holds that is not true for software development is that skills and experience vary between developers. Tolerance is not one of the things that there is in software development. Software products do not fail because of loosely tolerant components in the software. The last point where software is different than the processes for which Six Sigma was developed is that there is no priority behind a defect. There is no way to differentiate between a catastrophic failure and a minor defect that does not interfere with the system[3].

How software Six Sigma can work

First, the most common things that are measured in software development are the amount of time required to perform a task, the size of the product, and the number and type of defects. With these metrics, a company can utilize Six Sigma but it still has some other criteria for which the methodologies will work best. Six Sigma works best with incremental and evolutionary product development. This is because there are more code chunks to analyze and will fulfill the data demands of Six Sigma. Another important aspect is that the software development team needs to be ready to change their process. It may seem obvious but developers can be unwilling to change their ways[3].

Concerns implementing Six Sigma

There are plenty of concerns when implementing Six Sigma into software development. A couple of reported problems really have to relate to the methodologies not being directly related to software development. It’s a gap that many developers cannot bridge by themselves. Some of this resistance is in the developer’s reluctance to believe that the one size fits all Six Sigma approach will work, especially with the differences in manufacturing and software development. Also, an alarming aspect of Six Sigma enthusiast is the lack of understanding of the CMMI and ISO 9000 standards. Too many developers end up viewing Six Sigma as a competing initiative, and current methodologies do not address this concern [3].

Software Six Sigma

The software development companies that areclaimingsuccess with Six Sigma are using it to improve the processes within the standards of CMMI and ISO 9000. Consequently, for an individual process, CMMI would identify what activities are expected in the process and Six Sigma would identify how the activities might be improved. Looking at this from the organizational infrastructure, CMMI would identify how activities might be implemented and Six Sigma would utilize one of the five-step methodologies. Some of the benefits that development teams are reporting are defects are found earlier in the software development life cycle andthus become more predictable to the extent that they can predict, with relative accuracy, the number of defects. Also, scheduling tasks and project tracking has seen significant improvement[4].