Download the Article (PDF, 73 KB)
Miklos Biro, Budapest University of Economic Sciences and Public Administration, Richard Messnarz, ISCN, and Alfred G. Davison, QPAC
Adopting the practices of the worlds leading economy is a desirable and sensible strategy, since these practices are based on the largest existing pool of experiences. The question is whether these practices will be equally effective in other areas of the world where different national cultures condition different value systems. Capability Maturity ModelIntegrationSM (CMMI) is a framework from which models can be generated for different organizations. It is open to considerations regarding all kinds of circumstances including differences in cultural value systems that were characterized in G. Hofstedes seminal work. The statistically identified clusters of characteristics of national cultures are: power distance, individualism vs. collectivism, masculinity vs. femininity, uncertainty avoidance, and long-term vs. short-term orientation.
With this article, the authors intend to contribute to the international success of CMMI and of all other process improvement methods by proposing to work toward a third dimension of the CMMI/SPICE architecture: the cultural dimension, in addition to the process and capability dimensions of the existing models.
Key words: capability maturity model, management practice, national culture, software process improvement, tolerance, value system awareness
EDITORS NOTE: Many projects now extend beyond a single national setting, drawing upon team members in varying cultural settings. Despite the influences of globalization, significant differences exist in local outlook and expectation. The following treatment is a preliminary exploration of that situation for software process perspective, as well as a call for additional, more rigorous research.
Supporting Details of the Sociological BackgroundCharacterizations of national cultures are not to be understood as meaning to stereotype any nation or its citizens, but to be a shorthand representation of particular clusters of outlooks and expectations.
The seminal work (Hofstede 1994) identifies the generic factors that characterize value systems in different national cultures, including those of software and systems developers, applying statistical cluster analysis. The analysis was based on questionnaires from more than 50 countries. Each country could be given an index score for each of the following factors or dimensions of national cultures:
All four factors are a continuum between two extremes, and no national culture is at one or the other extreme. Furthermore, the index scores are statistical results, which means there are always individuals who do not conform to the general model. By consequent, the model is not meant to and should not be used to stereotype people from various cultures.
There is in fact a fifth factor identified only later due to the natural western cultural bias of the experts themselves compiling the questionnaires used for the study. This is:
THE THIRD DIMENSION
The following discussion is based on the Capability Maturity Model Integration (CMMI)SM framework. It can, however, be applied to any method designed to improve the processes of an organization. CMMI is a framework from which models can be generated for different organizations. Although process areas depict behavior that should be exhibited in any organization, practices must be interpreted using an in-depth knowledge of the CMMI model, the organization, the business environment, and the specific circumstances involved (CMMI 2000).
The previous statement bears a striking resemblance to Mary Parker Folletts way of thinking (Metcalf and Urwick 1940), which is not surprising because CMMI was developed in the U. S. cultural environment, and the essential characteristics of national culture are enduring despite all industrial and societal changes. Mary Parker Follett (1868-1933) was a U. S. pioneer of organization theory. It is fortunate that, at the same time, this statement clearly demonstrates the intention to be open to considerations regarding all kinds of circumstances, including differences in cultural value systems.
CMMI continuous representation is often illustrated as having a two-dimensional structure with the process dimension on one side and the capability dimension on the other. This is a sound approach, since each (process area, capability level) pair has a well-defined meaning and implies suitable improvement actions.
With this article, the authors propose to work toward a third dimension (see Figure 1) in the CMMI architecture: the cultural dimension. This model extension is valid, since the national cultural position of the company may determine a different meaning and suitable improvement actions for every (process area, capability level) pair, or even every (process area, specific or generic practice) pair using a finer granularity of CMMI where process capability levels are achieved by performing the specific or generic practices on the process areas. In short, the effectiveness of various practices depends on the national culture where they are performed. In fact, this model extension can be fully accommodated by even the current version of the CMMI, since all practices are only expected and not required model components (CMMI 2000).
The cultural dimension itself is multidimensionally determined by Hofstede, but there are other models of national cultures that are also valid and that the authors intend to leave the new model open to.
To remain as consistent as possible with the structure of the CMMI continuous model, the new cultural dimension has model components similar to those of the capability dimension of CMMI. Nevertheless, the authors use the old maturity level terminology instead of capability level for the cultural dimension, since it is a better description in this case. The model has also similarities with the ancestor of CMM: Crosbys Maturity Grid (Paulk 1995; Messnarz and Tully 1999; Crosby 1979).
Cultural maturity level 0 (closed)
The cultural maturity level of a process area at a given capability level is 0 if the specific and/or generic practices, leading to the achievement of the specific and/or generic goals of the process area at the given capability level, are prescriptive to the extent where no differences in cultural value systems are allowed.
Cultural maturity level 1 (open)
The cultural maturity level of a process area at a given capability level is 1 if the specific and/or generic practices, leading to the achievement of the specific and/or generic goals of the process area at the given capability level, are open enough to allow for differences in cultural value systems.
Cultural maturity level 2 (model based)
The cultural maturity level of a process area at a given capability level is 2 if the consideration of cultural differences is based on a scientifically established model. Hofstedes multidimensional model is an example, but other models are also acceptable in case they are useful in distinguishing the differences in cultural value systems that have an impact on the performance of the specific and/or generic practices.
Cultural maturity level 3 (comprehensive)
The cultural maturity level of a process area at a given capability level is 3 if the scientifically established cultural model is comprehensively applied to all specific and generic practices leading to the achievement of the specific and/or generic goals of the process area at the given capability level.
Cultural maturity level 4 (tailored)
The cultural maturity level of a process area at a given capability level is 4 if the depth and complexity of the application of the cultural model to the specific and generic practices is based on quantitatively managed experience and business needs.
Cultural maturity level 5 (competency driven)
The cultural maturity level of a process area at a given capability level is 5 if the cultural model applied to the specific and generic practices is refined, extended, or fully changed on the basis of competency acquired through quantitatively managed long-term model experience and business needs.
Level 1 Generic Cultural Goal
At level 1 the generic cultural goal is to create an organizational culture where specific and generic practices allow for differences in cultural value systems when the need is identified. A level 1 Generic Cultural Practice (GCP) is to scan specific and generic practices for cultural restrictions and relieve the identified restrictions.
An excellent example of a practice in CMMI where cultural restrictions are already partially relieved is Generic Practice (GP) 2.4: Assign Responsibility (CMMI 2000). The description of the practice contains the following: Responsibility can be assigned using detailed job descriptions . Strongly uncertainty avoiding German engineers will love this approach as opposed to weakly uncertainty avoiding Swedish engineers. GP 2.4 states, however, that Dynamic assignment of responsibility is another legitimate way to perform this practice . One can claim by consequent that GP 2.4 can actually be adapted to both weak and strong uncertainty avoiding national cultures.
Level 2 Generic Cultural Goal
Specific and generic practices take cultural differences into consideration on the basis of a scientifically established model. A level 2 GCP is the application of an element of the cultural model to selected specific and generic practices.
In the case of the Hofstede model, GCP could be the following:
In CMMI GP 2.4 of the previous example, the consideration of the individualism vs. collectivism factor of the Hofstede model (GCP 2.2) leads to another cultural restriction of this practice that should be relieved.
The assignment of responsibility is strongly related to the individualism vs. collectivism dimension. Should the responsibility be assigned to people or teams? The wording of GP 2.4 reflects cultural conditioning: Confirm that the people assigned to the responsibilities and authorities understand and accept them (CMMI 2000).
To illustrate the importance of this issue, the authors quote (Hofstede 1994) referring to a management researcher from the United States, Christopher Earley, who performed an enlightening laboratory experiment on a group of 48 management trainees from southern China and 48 matched management trainees from the United States. Half of the participants in either country were given group tasks, the other half individual tasks. Also, half of the participants in either country, both from the group task and from the individual task subsets, were asked to mark each completed item with their names. The other half turned them in anonymously. The Chinese collectivist participants performed best when operating with a group goal and anonymously. They performed worst when operating individually and with their name marked on the items produced. The American individualist participants performed best when operating individually and with their name marked, and abysmally low when operating as a group and anonymously.
In addition to GP 2.4 and many other parts of CMMI, the previously mentioned experiment has a profound impact on such Integrated Product and Process Development (IPPD) such as Organizational Environment for Integration and Integrated Teaming.
Level 3 Generic Cultural Goal
The scientifically established cultural model is comprehensively applied to all specific and generic practices. A level 3 GCP is the systematic application of an element of the cultural model to all specific and generic practices. In the case of the Hofstede model, GCP could be the following:
CMMI GP 2.8: Monitor and Control the Process and GP 2.10: Review Status with Higher-Level Management are both affected by GCP 3.1 and GCP 3.4. CMMI GP 2.7: Identify and Involve Relevant Stakeholders is, on the other hand, affected by GCP 3.3 and GP 3.2: Collect Improvement Information by GCP 3.5.
GCP 3.1: Power Distance considered in GP 2.8: Monitor
and Control the Process and GP 2.10: Review Status with Higher-Level
Both of these generic practices require a review involving communication with either the immediate level of management or higher-level management. Power distance has a determining impact on the communication considered appropriate in a given culture.
Referring again to (Hofstede 1994) here is a quote of a senior Indian executive from the United States. (Negandhi and Prasad 1971): What is most important for me and my department is not what I do or achieve for the company, but whether the Masters favor is bestowed on me.... This I have achieved by saying yes to everything the Master says or does.... To contradict him is to look for another job.... I left my freedom of thought in Boston.
It is obvious now that the way of performing generic practices 2.8 and 2.10 must take into account the power distance index in the national culture where the organization is located.
GCP 3.4: Uncertainty Avoidance considered in GP 2.8: Monitor
and Control the Process and GP 2.10: Review Status with Higher-Level
Track corrective action to closure is an important subpractice of GP 2.8. Precision and punctuality required by this subpractice is a natural characteristic of strongly uncertainty avoiding cultures, which will by consequent be better in this respect.
GCP 3.3: Masculinity vs. Femininity considered in GP 2.7:
Identify and Involve Relevant Stakeholders
The article (Atwong and Lange 1996) gives account of a virtual classroom experiment with students of the California State University-Fullerton and Lappeenranta University of Technology, Finland. The subject of the experiment was a marketing research project, which is irrelevant in this context. The importance is that the project combined the American and Finnish students into one virtual classroom with cross-national teams. Students used the Internet extensively for data collection and conducted Internet chats with foreign team members when necessary. The message of the story can be summarized with the opinion of a Finnish student: It was interesting to see the effect of cultural differences, even in a relatively simple project like this. When we first established contact with our American teammates, they wanted first to introduce themselves and chat about their interests and hobbies, which we thought was strange. Later we realized that this was their way to establish rapport with small talk. The Finns are used to getting immediately down to business. In the oral presentations, the American students seemed to emphasize presentation technologies more than us. However, in my opinion the quality of the work was roughly equal.
The previous observation is due to the difference between the United States and Finland on the masculinity vs. femininity scale, which is the only dimension where the United States and Finland are significantly different. Assertiveness is a characteristic that correlates more with masculine characteristics, while modesty correlates more with feminine characteristics. The involvement and the resolution of the conflicts of stakeholders from even these two otherwise close cultures requires a careful handling of the cultural differences without which both the assertive U. S. students and the modest Finnish students may find each other ridiculous, strange, shocking, or even hateful.
This issue, of course, must be taken into account while stakeholder involvement is planned in the project planning process area.
GCP 3.5: Long-Term vs. Short-Term Orientation considered
in GP 3.2: Collect Improvement Information
Process improvement as a whole, and especially this GP 3.2, clearly requires long-term orientation, that is, persistence, establishment, and observation of priorities, and thrift. Short-term orientation, which attributes high value to the protection of face and the respect for traditions, acts against process improvement.
Level 4 Generic Cultural Goal
Experiences with the consideration of cultural differences are quantitatively managed, and the depth and complexity of the application of the cultural model is based on the quantitatively managed experiences and business needs.
Experience may indicate the need for a deeper application of the Hofstede model. In fact, Hofstede himself examined pairs of cultural factors in addition to the single ones. GCP 4.1 considers such a pair. The number of pairs of the five cultural dimensions is 5C2 = 10. One must be careful, however, with increasing the number of considered combinations, since the number of cases may increase exponentially, which is not useful for ones purposes.
CMMI GP 2.7: Identify and Involve Relevant Stakeholders will be affected by GCP 4.1.
GCP 4.1: Power Distance and Uncertainty Avoidance considered
in GP 2.7: Identify and Involve Relevant Stakeholders
The preferred way of resolving conflicts between stakeholders can be predicted from the position of a culture in the two-dimensional space of power distance and uncertainty avoidance.
(Hofstede 1994) describes the results of an organizational behavior course examination reported by Owen James Stevens, an American professor at INSEAD business school in Fontainebleau, France. A mixture of French, German, and British students received a case study where they had to resolve a conflict between two department heads within a company. A sales and a manufacturing manager, for example, usually have conflicts, since sales tries to satisfy changing customer demands, while manufacturing is more efficient if batches are larger and changes are less frequent.
The results were striking.... The solution preferred by the French was for the opponents to take the conflict to their common boss, who would issue orders for settling such dilemmas in the future.... The solution preferred by the Germans was the establishment of procedures. The British solution was the registration of both department heads to a management course to develop their negotiation skills.
In summary, the French with large power distance and strong uncertainty avoidance prefer to concentrate the authority and structure the activities, the Germans with strong uncertainty avoidance but smaller power distance want to structure the activities without concentrating the authority, while the British with small power distance and weak uncertainty avoidance believe in resolving conflicts ad hoc.
Level 5 Generic Cultural Goal
The cultural model is refined, extended, or fully changed on the basis of competency acquired through quantitatively managed long-term model experience and business needs. For example, there may be a need for the consideration of cultural factors which are impractical to evoke from the existing cultural model. Following are the discussions of GCPs for two such factors.
In (OSuilleabhain et al. 2000), a practical example of the Johari window, is described showing how discrepancy in perceived understanding is measured in the OSIRIS European project.
The Johari window is a tool originally designed for conceptually distinguishing between four different possible states with regard to knowledge of oneself; these states are shown in Figure 2.
The same principle can be applied to measure four different possible states with regard to knowledge of ones culture, and the understanding of other cultures in the project team.
Respondents to the OSIRIS survey were asked to rate, on a scale from 1 to 5, the discrepancy between the understanding cultures in the ISIS project have of themselves, and the understanding others have of them. The questions were based on the Johari Window.
Figure 3 shows the perceived discrepancies between each cultures understanding of itself and the understanding other cultures have of it. Note that the ratings for each culture are by the other two cultures not by their own culture.
The bigger the discrepancy, the smaller the incidence of perceived shared understanding and, implicitly then, the fewer issues open for discussion. It is interesting that Germany was the only culture with an above-average mean score, which may imply that other cultures more often than not find their own understanding of German culture to be at odds with that of the Germans own and that blind spots, which can hinder effective interaction, may be greater than the Germans themselves believe.
A typical sign when this Johari problem happens is that during a certain project some results have to be reagreed a number of times, although the Germanic group feels that clear definitions have been agreed upon already.
This problem can be solved by moderating workshops in a way that allows people to explain the different viewpoints and identify synergies, finally resulting in common solutions.
GCP 5.1. German
Do the managers understand that the once agreed-upon business model might change, driven by stakeholder inputs and new demands, and that cultural difference requires understanding different viewpoints on the same.
As outlined in (Davison 2001) the movement of a large multinational organization toward higher capabilty is a long-lasting agreement process where step by step, one must win the confidence of the managers and staff. Many tactics are needed to achieve this. For example:
GCP 5.2. Large
Are the stakeholder agreements achieved step by step so that all involved parties are convinced and the joint mission is clear?
CMMI intends to be effective in all national cultures. By consequence, it should consider the way its practices can most effectively be performed in different national cultural environments.
This article draws attention to the deep relevance of this issue to the international success of CMMI and other process improvement methods. The eye-opening ideas are validated in a general sense and are meant to justify undertaking a more extensive study, which would require considerable resources and a worldwide collaboration.
The authors of this article encourage the establishment of a forum of people and organizations interested in sharing their ideas and experiences on process improvement in different cultural environments for improving effectiveness and suppressing intolerance and conflicts due to unawareness of differences in value systems.
Atwong, C. T., and I. L. Lange. 1996. How collaborative learning spans the globe. Marketing News 30, no. 17: 16-17.
Biró, M. 2000. Cultural environment protection in the information society. In Project Control: The Human Factor, Proceedings of the combined 11th European Software Control and Metrics Conference and the 3rd SCOPE Conference on Software Product Quality, eds. K. D. Maxwell, R. J. Kusters, E. P. W. M. van Veenendaal, and A. J. C. Cowderoy. Maastricht, The Netherlands, Shaker Publishing B.V.: 415-421.
CMMISMSE/SW/IPPD, V1.02 Continuous Representation. 2000. Capability Maturity ModelIntegrated for Systems Engineering/Software Engineering/Integrated Product and Process Development, version 1.02 Continuous Representation (CMU/SEI-2000-TR-031). Pittsburgh: Software Engineering Institute, Carnegie Mellon University.
Crosby, P. B. 1979. Quality is freeThe art of making quality certain. New York: McGraw-Hill.
Davison, A. G. 2001. The process psyopsMoving a large company towards SPI. In Proceedings of the EuroSPI 2001 Conference, eds. B. Hindel, C. Jorgensen, J. Elliot, M. Christiansen, R. Messnarz, R. Nevalainen, T. Stalhane, and Y. Wang. Limerick, Ireland: Limerick Institute of Technology: 10-12.
Hofstede, G. 1994. Cultures and organizations, software of the mind: Intercultural cooperation and its importance for survival. London: McGraw-Hill.
Metcalf, H. C., and L. Urwick. 1940. Dynamic administrations: The collected papers of Mary Parker Follett. New York: Harper & Row.
Messnarz, R., and C. Tully. 1999. Better software practice for business benefitPrinciples and experience. Washington, D. C.: IEEE Computer Society Pres.
Negandhi, A. R., and S. B. Prasad. 1971. Comparative management. New York: Appleton-Century-Crofts.
OSuilleabhain, G., R. Messnarz, M. Biró, and K. Street. 2000. The perception of quality based on different cultural factors and value systems. In Proceedings of the EuroSPI 2000 Conference, eds. B. Hindel, C. Jorgensen, J. Elliot, M. Christiansen, R. Messnarz, R. Nevalainen, T. Stalhane, and Y. Wang. Copenhagen: Copenhagen Business School: 2-32 2-45.
Paulk, M. C. 1995. The evolution of the SEIs capability maturity model for software. Software Process Improvement and Practice 1 (Spring): 3-15.
Siakas, K. V., and B. Balstrup. 2000. A field-study of cultural influences on software process improvement in a global organization. In Proceedings of the EuroSPI 2000 Conference, eds. B. Hindel, C. Jorgensen, J. Elliot, M. Christiansen, R. Messnarz, R. Nevalainen, T. Stalhane, and Y. Wang. Copenhagen: Copenhagen Business School: 2-20 2-31.
Dr. Miklós Biro is an associate professor at the Department of Information Systems and the Technology Transfer Center of the Budapest University of Economic Sciences and Public Administration. In addition to a doctorate in mathematics (operations research) from the Loránd Eötvös University in Budapest, he has an executive MBA degree from ESC Rouen, France, and a master of science in management degree from Purdue University.
Biro is a Bootstrap and SPICE assessor. He initiated and managed the Hungarian participation in numerous European multinational projects and organizations committed to software process improvement. He was an expert consultant for many Hungarian and international organizations (European Commission, Hungarian Committee for Technological Development, Investment and Trade Development Agency of Hungary, Hungarian Airlines, UNIDO, International Software Consulting Network). He has numerous publications in international scientific and professional journals and conference proceedings. He is the author of Hungarian and English language book chapters on operations research models, software engineering, software process improvement, and business motivations.
Biro is a member of the editorial board of the journal Software Process Improvement and Practice, and founding president of the professional division for Software Quality Management of the John von Neumann Computer Society. He is the Hungarian member of Technical Committe 2 (TC-2) Software: Theory and Practice of the International Federation for Information Processing (IFIP). Biro can be reached at email@example.com .
Dr. Richard Messnarz is executive director of ISCN. He has a masters degree from the University of Technology in Graz, Austria. He also received a doctorate from the University of Technology for his work on QUESA Quantitative Quality Evaluation System.
Messnarz was the technical coordinator in process improvement initiatives such as PICO and Victory and the coordinator of a software engineering group within the pilot project for building a prototype of a virtual university. He is the editor of the book Better Software Practice for Business Benefit: Principles and Experience. Messnarz is a chairman and the main organizer of the EuroSPI conference series. He can be reached at firstname.lastname@example.org .
Dr. Alfred G. Davison is a senior engineer/manager with a background in aeronautical, marine, telecommunications, and process control industries. He has broad-ranging experience, IT knowledge, and technical expertise. He has documented leadership, motivational, planning, strategic, tactical, and operational skills. He has fulfilled numerous roles, including CEO, director, R&D manager, and QA/QC chief. His experience also includes project management, product management, team building, business analyst, and design engineer. He is currently engaged in research on the organization of virtual software companies across geographical and cultural boundaries together with a number of universities across the world. Davison can be reached at email@example.com .
Capability Maturity Model and CMM are registered trademarks in the U. S. Patent and Trademark Office. CMM Integration and CMMI are service marks of Carnegie Mellon University.
If you liked this article, subscribe now.