Sigma Application to Software Quality

¶ … Sigma Application to Software Quality Management

In an increasingly globalized marketplace characterized by fierce competition, identifying the most effective approach to quality management and software development has assumed a new level of importance for most businesses today. Not surprisingly, in recent years, a number of management techniques have emerged that are designed to help companies improve their processes and operations across the board, including Total Quality Management, Six Sigma and the Capability Maturity Model. While these approaches differ, they all provide busy managers with a framework in which to approach continual improvement by identifying quantifiable areas of interest, soliciting feedback from affected and interested stakeholders, and by attempting to discern trends over time. While these techniques have all proven to be effective in helping a wide range of enterprises improve their processes and software, these all have certain inherent limitations as well. This paper provides a review of the relevant and scholarly literature to determine how Six Sigma and the Capability Maturity Model can operate in tandem through the oversight of a project management office to provide a synergistic effect that avoids many of the constraints of these techniques while drawing on the best of what both have to offer. A summary of the research and key findings are presented in the conclusion.

Application of Six Sigma to Software Quality Management

Introduction

While the application of Six Sigma to manufacturing has been well established, the application of Six Sigma for Software has received much controversy. Some observers maintain that it is inappropriate to implement Six Sigma in software development for a number of important reasons. To this end, this research paper provides an overview of Six Sigma and Quality Management principles, a discussion of the difficulties of implementing Six Sigma in software quality, followed by an examination of the applicability of employing the capability maturity model (CMM) to the Six Sigma to provide an integrated approach to improve software quality. A summary of the research and key findings will be presented in the conclusion.

Difficulties of Applying Six Sigma in Software Development

Background and Overview.

According to Hahn, Hill, Hoerl, and Zinkgraf (1999), "Six Sigma improvements -- a highly disciplined and statistically based approach for removing defects from products, processes, and transactions, involving everybody in the corporation -- has been adopted as a major initiative by some of our leading companies" (p. 208). Likewise, Bartels (2003) reports that, "Six Sigma is a management framework that, in the past 15 years, has evolved from a focus on process improvement using statistical tools to a comprehensive framework for managing a business. The results that world-class companies such as General Electric, Johnson & Johnson, Honeywell, Motorola, and many others have accomplished speak for themselves" (p. 1). The proponents of Six Sigma maintain that the approach is invaluable in its ability to improve quality, reduce costs, improve customer loyalty, and help achieve bottom-line results for a wide range of enterprises (Bartels, 2003).

According to Hahn and his colleagues, "More important than the technical definition is the concept of Six Sigma as a disciplined, quantitative approach for improvement -- based on defined metrics -- in manufacturing, service, or financial processes. This drives the process of selecting projects based on their potential to improve performance metrics, and identifying and training the right people to get the business results" (p. 208). In reality, this is not as complicated as it sounds; in many ways, the Six Sigma approach mimics many aspects of the Total Quality Management (TQM) and Capability Maturity Model approaches in that it continually seeks to identify opportunities for improvement in an organization's operations and processes by developing quantifiable measures that can be tracked over time to discern trends. Furthermore, these types of management techniques have been facilitated by the introduction of computer-assisted applications that provide managers with the graphic representations of this data that help determine where improvements need to be made and who should make them.

In this regard, Thorn (2003) reports that the Capability Maturity Model (CMM) is one of most commonly accepted standards and measures of effective project management. Developed by the Carnegie-Mellon University (CMU) Software Engineering Institute (SEI) with sponsorship by the Department of Defense (2002), the CMM provides timely descriptions of the primary elements of an effective software development and project management process that can be used to provide the basis for continuous process improvement. The Capability Maturity Model has since become the standard that is used to appraise the quality of the development process in a wide range of industries (Meyer, 2002). This author adds that, "The hypothesis underlying CMM is that effective practices must be built on one another in logical progression rather than adopted scattershot. Software developers participate voluntarily in the CMM assessment program in order to gauge their rate of improvement" (Meyer, 2002, p. 149). According to Thorn (2003), the CMM is an iterative function that defines a consistent path for improvement that can be used for organizational development from their origin as a set of ad hoc, immature procedures to a consistent, mature, disciplined process. The CMM also describes five specific levels of progressive process maturity: (a) Initial, (b) Repeatable, - Defined, (d) Managed, and (e) Optimizing; these five areas also have a recommended Process Areas (PA) that is addressed at each level with level 2 (Repeatable) specifically focusing on project management issues with an emphasis on the ability to establish and follow consistent processes (Thorn, 2003), which are discussed further below.

Integrating Six Sigma with the Capability Maturity Model.

One well established approach to improving the integration of the management and development techniques is to implement a project management office to oversee coordination of the software development. According to Thorn (2003), a project management office (PMO) will typically development applicable standards of performance for a given topic of interest called metrics. As noted above, this process is not all that mysterious: "On the surface," the author advises, "this method appears to be straightforward -- identify key parameters of project progress and success and track them. And, in fact, most organizations do just that with regard to the standards of scope, schedule, and cost" (Thorn, 2003, p. 48).

Simply applying a given metric to an area targeted for improvement is frequently insufficient to ensure success, though. Therefore, this approach is frequently augmented by the use of industry-specific "best practices" (loosely defined by Thorn (2003) as being lessons that were learned the hard way by someone else). Closer examination suggests the following issues may also become relevant during this process:

While best practice principles may appear universal, their application tends to be situation-specific. Stated differently, the relevance and accuracy of metrics to a particular situation or requirement is of paramount importance. Measurement for measurement's sake or use of the wrong parameter is of no value at all;

Even with appropriate measures, metrics often reflect symptoms of problems -- not their sources;

When viewed from the perspective of the project office, metrics can often appear to be a matter of governance instead of operational control. This is because project offices frequently do not use metrics (especially those of a summary or composite nature) as much as provide them to higher levels of an organization. As a consequence, the project office (and others contributing to the project effort) may not have complete understanding of the purpose and use of the metrics, causing it to view them as a burden or distraction as opposed to a benefit. Even worse, it may view the metrics as a type of "policing action" by higher levels of the organization, leading to a potentially adversarial relationship;

The resulting wariness of the project office can be offset to a large degree by clearly linking project objectives to organizational goals. Building this basis for understanding is essential for the creation of trust vs. The development of a suspicion of oversight and evaluation from afar (i.e., outside the project office) (Thorn, 2003, p. 48).

As a first step, the combined use of metrics and best practices has a distinct potential for success but an equally distinct opportunity for mistrust and failure (Thorn, 2003).

An extension of the CMM model that provides an improved integrated approach involving affected stakeholders in the development process is appropriately known as the Capability Maturity Model Integrated (CMMI). According to one authority, "CMMI is a global standard being used by large organizations in determining their ability to develop and maintain software; it is also a model for organizational improvement. Adopting the CMMI is important in achieving high maturity in the country's software development sector" (NCI to conduct software process, 2004, p. 7). Not surprisingly, because of its ability to provide substantive improvements, the CMMI approach is also becoming widespread in defense applications today (McCabe, 2004).

According to Thorn (2003), the fundamental advantages of using a project management office to provide a more comprehensive approach becomes clear when the PMO is combined with (at least) Level 2 of the Capability Maturity Model (CMM); this approach provides the ability to consistently replicate proven project management and control practices that lead to shared competencies and organizational success, as measured by the previously discussed project metrics (Thorn, 2003). "More to the point," another authority advises, "the objective of a PMO is to establish the procedures, processes, and standards that lead to a Center of Excellence -- not so much for itself as for the projects and organization it supports" (Pohlman, 2002, cited in Thorn, 2003, p. 48).

In this regard, Thorn points out that there are three basic approaches that can be used to establish a PMO to help facilitate the integration of Six Sigma or other management techniques with the CMM:

Project Repository -- The PMO serves as the primary source of information on project methodology and standards. It assumes, rightly or wrongly, that the organization has established and follows a cohesive set of tools for project design, management, and reporting. Often a first step in establishing the idea of consolidating or sharing management practices, it still falls short, both in direct project oversight and as a Center of Excellence;

Project Coach Model -- The PMO extends its repository role by moving to share project management and control practices across business functions through monitoring, if not controlling, project communications. This model is usually a permanent structure with dedicated staff and some degree of responsibility for direct management and control over project initiation, execution, and implementation;

Enterprise Program Office -- Representing the highest level of development and status, this model concentrates project management within a PMO with direct oversight and responsibility for all projects with regard to scope, schedule, and cost of projects regardless of where they originate in the organization. This model also assumes a fully developed governance process with well-defined metrics, as discussed (Thorn, 2003).

By all accounts, though, the foregoing are first steps and a commitment to follow through on these initiatives is needed from top management and all affected stakeholders. For example, according to Cook and Semouchtchak (2004), "Commitment to standardizing the software development process is required. Only 24% of organizations implementing the Capability Maturity Model have reached level 3, and only about 6% have achieved level 5" (p. 12). Furthermore, Thorn emphasizes that the project management office should not be considered to assume all responsibility for the initiative, but rather serves as a clearinghouse for feedback from stakeholders and a coordinating agency to help determine how best to approach the fix to the problem. Many authorities suggest that this is an essential ingredient in any successful CMM initiative: