NEWS AT SEI
This article was originally published in News at SEI on: April 1, 2006
Organizations throughout the world are turning to an ever-increasing set of international standards and models, such as Capability Maturity Model Integration (CMMI), Six Sigma, IDEAL, and others, in their effort to achieve competence in the processes used to manage their businesses, increase customer satisfaction, and achieve and maintain competitive advantage. The success of the Capability Maturity Model for Software (SW-CMM) and CMMI models, among others, has contributed to this trend.
The SEI’s work with industry organizations reveals, however, that many organizations encounter difficulties in implementing process improvement successfully in multi-model improvement environments. These interactions have led the SEI’s research team to develop an approach for successfully implementing process improvement in this challenging environment.
As organizations in multi-model improvement environments attempt to implement the standards and models they have chosen to meet business objectives and those that are mandated by regulators, they typically encounter a common set of barriers. Such barriers are frequently found in the automotive, financial, telecommunications, and service sectors. These barriers include the following:
- the proliferation of models that the organization must consider
- the differences in structure and terminology across the standards and models
- difficulty in recognizing similarities among standards and models
- conflict between different improvement programs within the organization, as each attempts to champion and foster improvement based on its standard or model
- a lack of sustainability and institutionalization of achievements made by individual improvement programs
- a proliferation of the number and types of audits, assessments, and benchmarking that the organization must undergo while performing the functions needed to manage its business
- the perception of high risk associated with process improvement, especially in a multi-model environment
Because organizations are typically structured to implement individual model-based improvement initiatives independent of one another, the risk that a new improvement initiative will fail is high. In addition, improvement programs compete for the resources available in the organization. Because improvement programs and the improvements resulting from such programs are not coordinated, those charged with the day-to-day execution, maintenance, and management of the organization’s processes begin to experience fatigue.
The challenge is to succeed in implementing process improvement despite the additional difficulties found in a multi-model improvement environment.
The Unified Process Improvement Approach
To succeed, organizations must change the way they approach process improvement. Improvement initiatives in multi-model environments must be integrated across relevant standards and models, and this integration must continue. A first step of integration is to recognize that despite the different structures and terminologies and despite different levels of abstractions, the standards and models used in the organization share common element types. The challenge is to examine the models to identify the common elements and then to organize implementation of these standards and models in the organization’s processes in relation to these common elements.
The SEI examined a series of standards and models and identified three common element types:
- good-practice elements
- improvement methods
- institutionalization elements
Every standard and model relevant to an organization will contain some or all of these element types. Once an organization starts to view the standards and models using element types, it has made a significant step toward a more effective approach to process improvement in a multi-model environment.
Good-practice elements define what an organization must improve in any particular area. The specific practices of CMMI are an example of good-practice elements. Other models with such elements include ISO 9001, COBIT, and ITIL.
Having identified good-practice elements across the relevant standards and models, the organization must view these elements as requirements on the organization’s process. This set of requirements then becomes the focus of change instead of the many individual models. Thus, the many good-practice elements across the relevant models become the relevant requirements to be satisfied by the organization’s processes. Therefore, these requirements can be approached as such when planning improvement cycles and structuring the organization to support process improvement.
Improvement methods are the model elements that help an organization effect its needed changes and facilitate its technology-transition process. These methods help the organization master the change-management process, communicate the vision of the future organization, plan and execute the resulting improvement initiatives, and manage them over time. The improvement methods also provide guidance for forming organizational structures, roles, processes, and methods needed to effect change. Examples of improvement methods include Total Quality Management, IDEAL, Six Sigma, assessments methodologies, and audits.
Organizations should select the components of any of the improvement methods available that are useful to them or familiar to them and that are best suited to their needs and cultures. The most important change is that organizations apply a single, uniform approach across all improvement initiatives. This does not mean using, for example, only Six Sigma; rather, it means that the organization applies its own uniform improvement methodology across all improvement initiatives in the organization no matter what standard or model or combination it implements. In particular, the organization should avoid using different transition methods for different models. Using one transition method increases the organization’s ability to communicate across models and individual improvement initiatives.
While helping industry partners implement CMMI-based improvement in multi-model improvement environments, the SEI saw that the strong emphasis in CMMI on institutionalizing the improvements was missing or only weakly represented in many other models. To sustain and institutionalize improvement, it is critical to apply a unified, common set of goals and practices related to institutionalization across the good-practice elements and improvement methods of all the standards and models used by the organization. The Generic Goals and Generic Practices of CMMI offer an excellent basis for institutionalization and may easily be extended by describing elaborations for the good-practice elements and improvement methods of other standards and models. This unified approach to institutionalization also contributes to a common understanding of how to sustain hard-won improvements across all standards and models, helping management, process users, and process professionals within organizations communicate with each other more effectively.
Managing Process Relationships in a Multi-Model Improvement Environment
If an organization revises its process-improvement efforts using this approach, it will establish a sound basis for implementing improvement in multi-model improvement environments. A major obstacle to success, however, remains: the complexity of the relationship among standards, models, and the organization’s own processes. For this reason, the unified process improvement approach stresses the importance of placing the organization’s own processes at the center of the improvement effort—because the standards and models can ultimately be implemented only in the organization’s process.
The standards and models must be mapped to the subprocesses, procedures, and instructions of the organization’s process. This mapping of individual standards and models to the organization’s own process must take place at the level of abstraction embodied in the standards and models. This work should be supported by the use of metadata, an annotation mechanism, to reference individual parts of an organization’s own process model to the corresponding parts of each relevant standard or reference model. This metadata-based mapping allows individuals in different roles in an organization to generate context-specific views of the organization’s process in relation to the standard or model currently in focus. It also helps organizations prepare more effectively for the many audits, assessments, and benchmarking activities to which they are committed because they can generate the relevant view easily.
Building the Experience Base
Organizations working in multi-model improvement environments may benefit from examining relevant standards and models in terms of good-practice elements, improvement methods, and institutionalization elements. By implementing process improvement in terms of these elements and linking the relevant standards and models to their own processes using metadata to allow the generation of context-specific views, organizations will have an effective approach to driving improvement in multi-model improvement environments.
The SEI continues to develop and refine the unified process improvement approach and to expand the experience base with its partners in industry.
About the Authors
Patrick Kirwan is as a senior member of the technical staff at the Software Engineering Institute in Europe. He is currently working on the challenges associated with the effective transition of new technology to industry in multi-model improvement environments. His research interests also include the challenges associated with building large interoperable systems of systems as part of the Integration of Software-Intensive Systems (ISIS) initiative.
Urs Andelfinger is a visiting scientist at the Software Engineering Institute in Europe and a professor of software and business engineering at Darmstadt University of Applied Sciences, Germany. He is currently working on effective process improvement adoption approaches with a focus on the financial industry. He has 12 years’ industry and consulting experience in the banking and automotive domains.
Hans Sassenburg is a part of the SEI’s Software Engineering Process Management group and works as a visiting scientist for the Software Engineering Institute in Europe. In 2002 he started doctoral-level research studying the area of software-release decisions. This work was finished in January 2006.
André Heijstek has almost 20 years’ experience with process improvement all over Europe. He led the corporate CMM program at Ericsson and consulted in finance, automotive, utilities, telecommunications, IT, and defense companies. Heijstek is an SEI authorized lead appraiser and trainer for CMMI. His current research interest is process adoption for sustained business benefit.