NEWS AT SEI
This article was originally published in News at SEI on: April 1, 2008
“Successful improvement in a multimodel environment is essential to building the resilient organization. And resilient organizations have the agility to achieve and maintain competitive advantage,” Mike Phillips of the Software Engineering Institute (SEI)1 told participants at a recent workshop sponsored by Lockheed Martin Corporation and the SEI.
Organizational growth and evolution test that agility and underscore the need for an integrated approach to using the multiple improvement processes, according to Ray Johnson, chief technology officer of Lockheed Martin Corporation. In his keynote address at the May 8, 2008, Hard Questions for Process Improvement in Multimodel Environments workshop, Johnson remarked that his company plans to bring in as many new employees over the next 10 years as it has today. Some new hires will represent company growth; others, replacements for those leaving the company. All will need to understand how process improvement is central to the company’s success.
More than 80 managers, process improvement professionals, technical professionals, and researchers from 40 organizations participated in the workshop. The participants represented U. S. Department of Defense and civilian government agencies, consulting firms, commercial businesses, universities, and research centers. Based on the experience of those participants, three out of every five larger organizations are already facing the challenges of using multiple technologies to meet customer satisfaction, business profitability, market share, product and service quality, cost reduction, and other objectives. Overall, 85% of participants reported that their organizations use one or more of the Capability Maturity Model Integration (CMMI) constellations (CMMI-DEV, CMMI-ACQ, and CMMI-SVC). About half of the organizations represented use an ISO standard, Six Sigma, Lean, or ITIL. In all, the participants brought experience with 30 process improvement standards, frameworks, technologies, models, and practices.
Emphasizing the overarching importance of taking a strategic approach to multimodel process improvement, the workshop featured two rounds of process strategy sessions. In the first round, all of the participants worked in concurrent process strategy sessions discussing scope, solutions, needs, and priorities. The results were then funneled into the second round where two groups continued discussing process strategy and three groups separately discussed process architecture, technology relationships, and implementation and deployment.
From these sessions, participants recommended that organizations in multimodel environments form a strategy that measures process improvement against business objectives and create an architecture that reveals where additional technologies are needed. They called for process model developers and the process improvement community to develop guidance for combining popular technologies and reduce duplication of effort by defining the differences between single-model and multimodel environments. Participants strongly urged developers in particular to work together toward building interoperability into standards, models, and technologies to make their transition into multimodel environments easier.
“We believe our discussions here can help to catalyze conversation between process improvement model developers and their customers about designing interoperability into the models,” Jeannine Siviy of the SEI said.
Align Performance and Business Goals (Sessions on Process Strategy)
The process strategy groups identified the need to align performance strategy with business objectives as the primary focus. They suggested the need for:
- a cookbook for what is needed to address multiple models
- a common terminology
- an open architecture (Process improvement is not about modeling, but how models are implemented.)
- a management paradigm for process improvement comparable to that of the rest of the organization
The group also noted that some solutions for the challenges of developing a strategy for process improvement in a multimodel environment might come from other domains such as service-oriented architecture (SOA).
Combine Technology Elements (Session on Technology Relationships)
A way to combine technologies that delivers the best of each as demanded by the organization is important in a multimodel environment. However, the technology relationships group reported there is a paucity of solutions for combining improvement technologies.
One proposed approach is to use element classification, suggested Patrick Kirwan of the SEI. Improvement technologies can be classified according to their primary element type—good practice, improvement method, or institutionalization. According to the white paper Improvement Technology Classification and Composition in Multimodel Environments that was provided to workshop participants:
- Good practice elements define what or how an organization needs to improve.
- Improvement method elements drive the change and facilitate the technology transition processes in the organization.
- Institutionalization elements help an organization in sustaining achieved improvements.
The technology relationship session called for research to provide:
- practical guidance for combining the top 10 most popular improvement technologies
- an open forum for community discussion
- ways that appraisals and audits can enable convergence of technologies
- a taxonomy for common aspects among models and increased understanding of the differentiating elements between them
Map the Delta (Session on Process Architecture)
The value of process architecture, according to the group, is that it allows an organization to “map the delta”—that is, to determine how to add needed technologies at a higher level where there is greater flexibility in making tradeoff decisions.
At the beginning of the workshop, M. Lynn Penn, director of quality systems and process management at Lockheed Martin, presented a case study attesting to the value of creating a process architecture. According to Penn, Lockheed Martin’s Information Systems and Global Services (IS&GS) organization of 10 companies and 52,000 employees uses process architecture to “instantiate compliance to industry standards via a single organizational standard process.”
Among the benefits that this approach has delivered for IS&GS and Lockheed Martin are a “30% cycle-time reduction, a 20% lowering of software costs” and a robustness that has made it “easy to build in new models and practices,” according to Penn.
The process-architecture group suggested that more examples like Penn’s experience are needed. Such case studies would help to draw the distinction between architecture and a mere collection of processes. The group also identified a need for languages, methods, and tools to define an architecture and subsequently define the processes that are ready for users.
Reduce and Minimize Duplicate Efforts (Session on Implementation Issues)
A multimodel environment challenges the process-improvement engineers and developers charged with implementation because of more complicated sharing of audit and appraisal processes and data, greater coordination in training, and potentially overlapping roles and responsibilities, among other aspects. Those kinds of challenges can duplicate effort and foster waste.
To gain understanding about how to reduce duplication and waste, the implementation working group recommended that research be conducted in several areas, including the following:
- cost-effective and transparent means to get integrated feedback
- case studies that document the cost comparison between multimodel and single-model environments
- an integration framework that shows Venn diagrams of various models and standards in public domain
Multimodel Environment Issues are Pervasive
Workshop participants emphasized that issues in process strategy, technology relationships, process architecture, and implementation and deployment arising in multimodel environments cut across organizational levels, roles, responsibilities, and practices. The Process Improvement in Multimodel Environments (PrIME) project that the SEI proposes to lead will span the range of those topics and others in establishing an approach to harmonizing process-improvement technologies in multimodel environments. “Multimodel improvement is our reality,” Siviy said. “Our processes should be engineered with the same attention we bring to the engineering of our products.”
For Further Reading
Visit http://www.sei.cmu.edu/prime/ for the workshop presentations—including the proposed SEI approach to model harmonization—and the white paper series that examines problems organizations encounter when operating in multimodel environments and the current process improvement approaches such organizations must consider.
PrIME Project Sponsorship
The SEI is looking for organizations now to fund the PrIME efforts for the full three years of the project. Project sponsors enjoy benefits such as influence on the order in which models and research themes are addressed, royalty-free license for the duration of the project, one-on-one implementation meetings with the SEI, early access to research outputs, and participation in selected workshops. For more information on sponsorship opportunities, contact us using the link in the For More Information box at the bottom of this page.
1 Along with Jeannine Siviy and Patrick Kirwan, Mike Phillips is the co-technical lead for the Process Improvement in Multimodel Environments (PrIME) project. The SEI is proposing PrIME as a three-year effort to establish an approach to harmonizing process improvement technologies in multimodel environments.