Journal of Global Information Management, Vol. 8, No. 1.
Cultural Problems in Applying SSM for IS Development
Moores, T. T. * and Gregory, F.H. **
* Division of Accounting and Computer Information Systems, Emporia State University, Emporia, KS 66801, USA. E-mail: moorestr@emporia.edu.
** Independent Consultant, Room 1002, 103 Condominium Project 4, No. 35, Soi Wat Pratanporn, Chiang Mai 50200, THAILAND. E-mail: isfrank@loxinfo.co.th.
ABSTRACT
Soft Systems Methodology (SSM) is a general problem structuring approach that seeks to incorporate multiple stakeholder views in the analysis of a given problem. When applied to IS development, the method requires negotiation and debate between the stakeholders when exploring the feasibility of developing an information system. The applicability of this approach depends, however, on the willingness of participants to enter into such an open discussion. This paper reports on a case study that highlights the problems of applying such a confrontational method in an Eastern culture, such as Hong Kong. Three main problems were identified: 1. Group discussions were avoided; 2. Interviews were conducted in multiple languages; and, 3. High staff turnover made it difficult to develop and maintain mature stakeholder views.
INTRODUCTION
Soft Systems Methodology (SSM) is a general problem structuring method devised by Checkland (1981) that requires the building of rich pictures, root definitions and conceptual models of the human activity under study. The root definitions and conceptual models are built by the stakeholders in the problem domain in an iterative debate organized by a facilitator. This process allows the features of the problem area to be explored. In short, SSM takes a system’s approach to explore problem situations where there is no fixed outcome. If a solution has already been determined, there is no need to apply the SSM method.
With the importance placed on identifying stakeholder viewpoints, SSM would seem to be capable of recognizing the cultural values that may be central to an organization. These cultural values would include the beliefs, assumptions, and values shared by members of the organization. The need for debate, however, and the expected conflict between different stakeholder views, relies on the fact that the participants are willing to enter into such an open discussion, putting forward their personal views of whether the system is needed, and if so, what the system should do. But this, in itself, rests on the Western cultural assumption that debate is constructive and the most effective means of eliciting the information required.
Following Hofstede (1984), it is generally accepted that Western cultures are more individualistic, tolerant of different behavior and opinions and have a lower regard for status differences than Eastern cultures. Eastern cultures, on the other hand, tend to be more collectivist, have higher uncertainty avoidance and have an acute regard for status levels. The issue to be addressed here is the impact this cultural difference has on the early stages of an information system (IS) development project. Hong Kong is perhaps a unique place to investigate this problem, since, until recently, it has been Western governed for more than 150 years, but has remained an essentially Chinese culture.
The cultural problems inherent in IS development will be demonstrated by presenting the results of a case study based on an M.A. project (Lau, 1996) supervised by one of the authors. The project attempted to apply SSM in order to assess the feasibility of developing an Executive Information System (EIS) for a marketing unit in Hongkong Telecom. Three main problems were observed. Firstly, no group debate took place between the stakeholders. The student, being Chinese, avoided the confrontation inherent in this approach and conducted interviews on a one-to-one basis. Secondly, these interviews were typically conducted in Cantonese, but written in English. A feature of Hong Kong is that three languages are in common use: Cantonese, Putongua and English. As such, interviews could be conducted in different cultural modes depending on the language used. Finally, another feature of Hong Kong is the relatively high rate of staff movement among executives and professionals. Before the study had been completed all of the original stakeholders had been transferred. This clearly has a drastic effect on the ability of mature, experienced stakeholder views to be developed or captured.
The next section will outline the use of SSM in the development of information systems, and the Logico-Linguistic Modeling (LLM) enhancement used to determine the information sources required. Section 3 will then describe the case study, while Section 4 will identify the key issues that prevented a "normal" implementation of the debate and exploration process expected.
SSM AND ITS LLM EXTENSION
Soft Systems Methodology (SSM) is often used as an information system design method. Developing the SSM models requires an analyst to determine the principal components of the activity or system under study. These principal components are guided by CATWOE analysis. CATWOE involves the identification of:
? the Customer, who is the victim or beneficiary of the system;
? the Actors who carry out the transformation;
? a Transformation that underpins the system;
? the World view (Weltanschauung) that makes the root definition meaningful;
? the Owner, who can stop the system; and,
? the Environmental constraints that subsume the activity.
For instance, analyzing a University may identify the transformation as the process of transforming ungraded students to graded students. In which case, the customers might be the future employers, and the actors might be the lecturers. The world view may be that grades are necessary to identify the best employable people, while the owner might be the fee-paying students or the funding Government, and the environment might be the job market.
Wilson (1990) extended the method by showing how it could be used to support the detailed restructuring of an organization (Organizational Mapping), and the use of Information Categories and the Maltese Cross to support the conversion of a SSM model into a detailed information system design. Information categories show the required information inputs and outputs from the activities in an expanded conceptual model, while the Maltese Cross is a matrix of these categories that shows where the new information processing procedures are required. The systems devised are meant to be culturally feasible and systematically feasible.
Shortcomings of the method include the fact that Wilson’s method is restricted to the design of transaction processing systems and provides no facility for building other types of information system, such as decision support or knowledge based system. The method also assumes that the people in the organization know what information is needed to support their activities. This is particularly striking since the method often calls for new activities to be included in the model. These shortcomings are also found in other methods linking SSM with information systems design.
Logico-Linguistic Modeling (LLM)(Gregory, 1993, 1995) attempts to add a strong logical foundation to the SSM method by making a distinction between information that is empirically true and information that is true by definition. When this is done, the logically enhanced LLM-SSM models can be developed into a knowledge-based system that has learning capability.
THE CASE STUDY
The study described here centers on the application of LLM-SSM to derive the requirements for an Executive Information System (EIS). The analyst involved in the project, Lau Siu Pong, was undertaking a part-time Masters degree in the Information Systems Department at the City University of Hong Kong. He was a market analyst within Hongkong Telecom, which had been the sole provider of telecommunications services in Hong Kong, until the franchise was opened up with the granting of two other licenses in July, 1995.
The market for communication technology and services is particularly lucrative in Hong Kong, with mobile phones and pagers being carried by most sectors of the community, from University students to sales and company executives. In the face of increased competition, Lau was interested in developing an EIS to support the activities of the Business Market Business Unit (BMBU). The BMBU is charged with planning the marketing of new business products. It was not envisaged that the project would immediately result in a fully computerized EIS. Rather, it was intended to undertake an analysis to determine if an EIS would be suitable, and if so, to broadly define its configuration.
The project proceeded in a manner very close to the conventional seven-stage model of SSM (Checkland & Scholes, 1990). Stages 1 and 2 (situation entry and problem expression), normally requires an outside consultant to act as an analyst/facilitator. This is meant to bring an independent, and possibly fresh, perspective to the problem situation. In this case, however, Lau acted as both the analyst and as one of the stakeholders of the system. While this ensured access to the people concerned, Lau’s own views about what should be done had the potential to cloud his judgement about the nature of the system required. For this reason, the role of facilitator was taken by one of the authors, Frank Gregory. The debate between the analyst (Lau) and the facilitator (Gregory) often required the analyst’s assumptions and his evaluation of the system to be constantly challenged and criticized.
Stage 3 involves the formulation of root definitions of relevant systems of purposeful activity. Stage 4 involves building conceptual models of the systems named in the root definitions. The situation was slightly unusual in that the analyst was himself one of the "Actors" defined in the CATWOE analysis. The root definitions and conceptual models produced by the analyst were discussed with Gregory, colleagues and managers within Hongkong Telecom.
Stage 5 involves comparing the model to the real world in order to determine the extent to which actions suggested by the model actually occur. Based on differences between the model and the real world, Stages 6 and 7 involve identifying changes that are needed that are also acceptable to the stakeholders. In this case, a number of actions were proposed, including setting up a news clippings database, establishing feedback channels from sales staff, and increasing the flow of marketing information between departments.
It was during Stages 3 and 4, where discussion and debate was required between stakeholders, that the cultural features inherent in applying the method in Hong Kong became apparent.
THE CULTURAL ISSUES
Culture is often defined as a pattern of basic beliefs, assumptions and values held by the people concerned (Deal & Kennedy, 1982). In an organizational context, culture is also taken to include the control and exchange mechanisms inherent in the organization (Jones, 1983; Wilkins & Ouchi, 1983), that is, the "way things are done around here" mentality. Although Chinese management has a more group-orientation (Lockett, 1988), where collective ideals are emphasized, Hong Kong comes relatively low in terms of uncertainty avoidance in Hofstede’s model (Hofstede, 1984; Leung & Bond, 1989), suggesting they are willing to take risks. This suggests that Hong Kong, while having a basically Chinese group-oriented culture, has also adopted an entrepreneurial spirit, where business is as dynamic and high risk as in any other Western country.
The question, therefore, is whether the culture defined for Hong Kong by other studies also includes the willingness to carry out an open debate, as required by SSM to derive the rich pictures and root definitions of the problem area being investigated. Furthermore, would there be any other features of Hong Kong culture that might create problems when attempting to develop an information system? While IS projects often encounter problems, there were three issues that were clearly rooted in the way things are done in Hong Kong.
Avoidance of Group Debates
In a context that involves Western and Chinese elements, a "debate" can involve a head-on clash of values. For a Westerner, a debate can be seen as constructive and beneficial. When people have adversarial positions, a debate may help the truth to emerge and so be considered as beneficial to all parties, even for those that "lost" the debate. For Chinese, however, the display of adversarial positions depends very much on the social standing of the two adversaries. An argument between peers is acceptable, while a junior arguing with a senior is seen as unacceptable.
In SSM, the building of root definitions and conceptual model is meant to proceed by means of an iterative debate involving the stakeholders and a facilitator. It is assumed that much of this work will take place in group meetings in which the stakeholders will express their individual viewpoints. In these circumstances, it can be expected that the viewpoints of two or more stakeholders will be antagonistic. The role of the facilitator, therefore, is to mediate and produce a model that accommodates as many viewpoints as possible.
In this case, however, it was clear to both analyst and facilitator that the confrontational debate approach needed to be avoided if colleagues and managers were to be persuaded to participate in the project. The analyst insisted that such group debates would not be acceptable to the stakeholders that had been identified. The result was that the analyst avoided the group debate approach and conducted one-on-one interviews in order to build the SSM models. While this diverged from the principle of stakeholder debate required by SSM, the one-on-one interviews was seen as a more culturally-sensitive means of eliciting opposing views. The ‘debate’, in other words, was carried out on paper where status levels could be more easily respected.
The implication here for other IS development methods is that any technique used to elicit user requirements that involves debate between non-peers would be doomed to failure in an eastern culture. This would suggest that techniques such as Joint Application Design (JAD) and the Nominal Group Technique (NGT) could not be successfully implemented in Hong Kong, since both approaches require debate.
The JAD approach is specifically designed to use group meetings between users, stakeholders, and IS professionals, in order to debate the form and function of a proposed information system. It is exactly this phase that was avoided during the SSM project. The NGT approach requires participants in a group meeting to write down ideas and suggestions anonymously, before a facilitator summarizes and presents the ideas for debate. The purpose is to prevent the meeting from being hijacked by one or more dominant individuals, and to allow all views and opinions to be heard. However, it is entirely possible that sub-ordinates would resist criticizing an idea in the subsequent debate just in case the idea originated from one of their superiors.
In short, while it is not possible to say whether the paper-based debate strategy used was entirely successful, it is clear that the power distance dimensions of Chinese culture make formal debates an unacceptable approach to eliciting system requirements.
The Use of Multiple Languages
The analyst reported that the SSM models, written in English, would be discussed with his colleagues in Cantonese, although the discussion itself would be written down in English. It is not uncommon for two Cantonese speakers to intersperse English words into their conversation if the speaker is familiar with the fact that the English word requires less effort than the Cantonese alternative. The ‘effort’ here is the number of syllables required to pronounce the word or phrase.
This sprinkling of English is only made, however, when the speaker believes the listener is sufficiently familiar with English words that the basic meaning will still be communicated. Because of the recent history of Hong Kong, such linguistic acrobatics are quite common. In particular, while the local Chinese dialect is Cantonese, University teaching is conducted in English, although the recent transition in Sovereignty has meant Putongua (i.e., Mandarin) is now being taught in Cantonese secondary schools.
If philosophers of language and thought are correct, however, the language one uses will determine the breadth and depth of ideas that can be communicated. For instance, Whorf (1956) suggests that what one can think of is constrained by the language that one has to express these thoughts. The often quoted myth attributed to Whorf is the suggestion that North American Eskimos have a richer picture of frozen or freezing water because they have over 70 words for concepts such as snow, ice, slush, etc. The intimate connection between language and thought is also central to the philosophy of Wittgenstein (1953).
On this basis, Cantonese has a richer linguistic set than Putongua. Cantonese has nine tonal levels, while Putongua has four. A simple mistake in tone can change the Cantonese phrase "Cigarette lighter" to the more sinister "Kill the waiter!" In common with Putongua-speakers in Taiwan, Cantonese uses the older Chinese script, while the Mainland uses a more modern, simplified character set. In terms of tonal levels and complexity of script, therefore, Cantonese is perhaps unrivalled in its Whorfian ability to express thoughts, with subtle differences in tone or pen stroke communicating radically different ideas.
Because of the use of multiple languages in Hong Kong, however, it is not even clear which language is being used to express the beliefs, assumptions and values of the stakeholders. Since the analyst and stakeholders share a common language, the view of the stakeholder is likely to have been communicated effectively.
Given that something is always lost in translation, however, developing the models in English suggests the stakeholders’ views must then be transformed into an English concept before it could be represented in the SSM diagrams. Even if the analyst understood the viewpoint of the stakeholder, it is unlikely there would always be an English word to match the idea expressed in Cantonese. To compound this problem, when the analyst discussed the same diagrams with the English facilitator, one would have to assume the shared culture is Western, since the facilitator (Gregory) does not speak Cantonese.
One of the likely consequences of this shift is that the facilitator would bring to bear a more precise understanding of the English words used. In this case, there is the possibility that some of the ideas inadequately expressed in English by the analyst were misinterpreted by the facilitator in subsequent discussions. In reviewing the models, some of the words may have been changed to express an idea closer to the English description given by the analyst, although the subsequent rewording may have shifted the SSM model further from the original viewpoint. In turn, when discussing the revised models with the stakeholders, the analyst would then be faced with translating the new words into Cantonese. Given this continual shift from Cantonese to English (and back again), it is unlikely that the models developed could ever fully represent the (Cantonese) views of the stakeholders.
While most IS projects struggle to derive models that faithfully reflect the views of those concerned, it would seem to be an extra twist of the knife to require a further transformation from one language to another. Formal methods that use mathematical notation rather than free text, such as Z, would not be a solution here, since there is still a transformation from free text to the Z notation.
Furthermore, it would not be acceptable to suggest that an information system should only be developed in one linguistic context, since the development of the World Wide Web and electronic commerce has opened up the boundaries of business. As information systems stretch across the globe, a better solution might be to include aspects of Hofstede’s work in the curriculum of IS management courses. Information system projects of the next millenium would seem to demand more training in issues of culture shock, ethnocentrism, stereotyping, and the differences in language and humor. This suggests that IS development projects that tap more than culture should include a cultural feasibility study, to determine the extent to which any cross-cultural issues could jeopardize the project.
Rapid Turnover of Stakeholders
While the problems outlined above can be overcome to some extent, there was one problem that neither the analyst nor the facilitator could do much about. By the time the feasibility study had been completed, Hongkong Telecom had transferred all of the stakeholders involved in the model building project. Indeed, Hong Kong’s high growth and relatively low unemployment has helped to produce an environment in which staff movement between companies is also relatively high. Recent Government figures indicate that staff movements among executives and professionals is nearly 5% per year (CSD, 1997), and it is generally accepted that junior staff will change jobs every 18-24 months.
In this case, the Head of the BMBU was transferred a few weeks before the project was completed, and within three months the analyst and all other members of the unit had been transferred.
While the project was successful in determining the scope of the EIS required, it has to be admitted that the SSM project itself was a failure. The key issue here is that SSM is intended to work because the stakeholders produce and "own" the solution. With their viewpoints inherent in the models, the stakeholders are meant to recognize the solution as being of benefit to them, and thus, to support the implementation of the solution. If the original stakeholders have moved on, then the new staff will not be the owners of the solution, and thus, nobody can be expected to implement the solution. The SSM project would have to be started again.
The time scales involved in staff movements suggests that, perhaps, no SSM project will be successful in Hong Kong. Although there are no published figures on the duration of SSM projects, six months elapsed time is an appropriate benchmark. SSM evolved through a learning cycle using Masters students at the University of Lancaster. Most of the Masters students completed an action learning project lasting approximately six months. Much of the development of SSM is based on what was learned during these projects. However, of the six SSM projects conducted by the facilitator in this case study (all of six months duration), four were badly affected by the movement of major stakeholders within the organizations concerned.
It would appear, therefore, that for SSM to be applied in Hong Kong, the project time would need to be significantly reduced to cope with staff movements, or significantly increased, to include a second iteration that includes the new stakeholders. These approaches may not solve the problem, however. Shortening the project time impacts on the ability of the project to gain all the necessary information, resulting in an incomplete analysis. Even if the analysis has been completed satisfactorily, the rapid transfer of staff suggests there would be no stakeholders left to recognize their part in the development of the system and encourage its use. The alternative solution of adding a second iteration to include the new stakeholders runs the risk of spiraling the project out of control, since, presumably, if the project is again badly affected by staff turnover, another iteration would be required in order to include the newer stakeholders.
The need for a stable development team would seem to be a basic requirement in the development of any information system. It should be pointed out that the individuals in this case were transferred rather than left the company, but the effects are the same from an SSM standpoint. Other development method may be less affected by this problem, since development teams would typically have representatives of stakeholders, users and developers.
In this case, it is possible to have turnover in these members as long as the same viewpoints are being expressed. However, SSM is specifically aimed at clarifying the nature of a problem from a number of stakeholder viewpoints, and a particular viewpoint is clearly tied to a particular person and their experiences and perceptions of the problem. In short, for any organization with a high staff turnover, it must be accepted that most SSM projects will not be completed satisfactorily.
CONCLUSIONS
This paper has presented the results of a case study that applied LLM-SSM to analyze the feasibility of developing an Executive Information System (EIS), for a business unit within the newly de-monopolized Hongkong Telecom. The EIS was intended to support the Business Market Business Unit, a unit that plans the marketing of new products. The LLM-SSM approach was adopted at the outset because it was felt the method incorporated a more culturally sensitive approach to investigating the feasibility of a project. That is to say, the method requires that multiple stakeholder views are an essential element in constructing models of the problem situation, and that any solution must fit in with the way things are done in that organization.
As such, it was felt that LLM-SSM had a better chance of coping with any cultural problems that might be encountered during the project. While it was not known at the outset exactly what these problems would be, an English facilitator leading a Cantonese analyst in the application of a Western-derived method for a project in Hong Kong were sufficient ingredients to suggest that there would be some.
The main problem encountered was the need to involve stakeholders in a debate over the nature of the problem, the causes, and potential solutions. Such debates are common in IS development methods in order to elicit a full set of requirements. In this case, the debate centered on the root definitions and conceptual models developed in Stages 3 and 4. It was at this point that the analyst suggested that such direct confrontational debates could not be conducted within his unit. It was not that the analyst was incapable of conducting such discussions, rather, the potential for disagreements between non-peers suggested the technique was culturally inappropriate.
The solution was to effectively conduct a ‘virtual’ debate, where ideas and suggestions were recorded on paper and discussed on a one-to-one basis. If this approach had not been taken, it is likely that the project would have failed, since the stakeholders would have refused to participate in the debate.
The use of multiple languages raises the issue of how well the English-labeled models can faithfully reflect the viewpoints of the stakeholders. Given that the stakeholders communicated their ideas in Cantonese, should the models also have been developed in Cantonese? If the models had been developed in Cantonese, however, this would have ended the project, since the analyst (Lau) needed the help of the facilitator (Gregory) in order to develop the models, but the facilitator does not speak Cantonese. The result was that the ideas expressed in Cantonese, recorded and debated between the analyst and facilitator in English, revised, and then re-translated into Cantonese in order for the analyst to discuss the revised models with the stakeholders. Thankfully, the analyst was sufficiently fleet of foot to perform this linguistic dance, but the semantic precision of the models is open to question.
The departure of most of the stakeholders from the unit is one issue that could not be overcome. If the people involved in discussing the problem move on, then the SSM requirement that the stakeholders own the solution and are responsible for its implementation is lost. Four of the six SSM projects carried out by the facilitator in this case study were badly affected by staff movements. While it would seem appropriate to either shorten the project time or add additional iterations to take account of the new stakeholders, neither approach is likely to be successful. It must be accepted, therefore, that SSM is a risky approach where there are significant levels of staff movement. Such turnover is a feature of Hong Kong, but this has critical implications for the likely success of any project more than a few months in duration.
Even given the problems encountered, it should be noted that the general seven-stage SSM approach was applied successfully to the extent that the problem situation was analyzed by incorporating multiple stakeholder views. The logico-linguistic models produced were also successful to the extent they allowed problem areas and potential solutions to be identified. This suggests that the problem does not lie with the ability of people in Hong Kong to carry out a SSM project, but in the techniques applied to achieve the goal of analyzing the problem situation.
This case study suggests that no analysis technique should be used that requires direct debate between non-peers. In terms of expressing their views, the language used by those involved will be of crucial importance in determining what is said and how well it can be communicated. How well ideas can be communicated when multiple languages are being used is an open question. Finally, for any organization where there is a high turnover of staff, it is difficult to imagine any IS project being completed satisfactorily.
In conclusion, as the global markets expand and information systems are created that connect the world, the possibility of IS projects failing through culture shock increases. The ‘shock’ here involves the realization that people around the world think and behave differently. When it becomes apparent that some people do not and will not behave as expected, the problem of what to do next becomes of crucial importance. The recommendation being made here is that the method used to develop a global system must be examined to determine its propensity for culture shock. As this case study has shown, the IS project can be conducted successfully, as long as the means to achieve those ends are also culturally appropriate.
REFEFENCES
Checkland, P.B. (1981). Systems Thinking, Systems Practice. Chichester: John Wiley.
Checkland, P.B. & Scholes, J. (1990). Soft Systems Methodology in Action. Chichester: John Wiley.
Deal, T. & Kennedy, A. (1982). Corporate Cultures. Reading: Addison-Wesley.
Gregory, F.H. (1993). Cause, effect, efficiency and soft system models. Journal of the Operational Research Society, 44, 333-344.
Gregory, F.H. (1995). Soft system models for knowledge elicitation and representation. Journal of the Operational Research Society, 46, 562-578.
Hofstede, G. (1984). Culture’s Consequences: International Differences in Work-Related Values. Beverly Hills: Sage.
Jones, G.R. (1983). Transaction costs, property rights and organizational culture: An exchange perspective. Administrative Science Quarterly, 28, 454-467.
Lau, S.P. (1996). Logical soft system models for information requirement analysis/elicitation. MA Thesis, Department of Information Systems: City University of Hong Kong.
Leung, K. & Bond, M.H. (1989). On the empirical identification of dimensions for cross-cultural comparisons, Journal of Cross-Cultural Psychology, 20(2), 133-151.
Lockett, M. (1988). Culture and the problems of Chinese management, Organization Studies, 9(4), 475-496.
Whorf, B.L. (1956). Language, Thought and Reality. London & New York.
Wilkins, A.L. and Ouchi, W.G. (1983). Efficiency cultures: Exploring the relationship between culture and organizational performance, Administrative Science Quarterly, 28, 468-481.
Wilson, B. (1990). Systems: Concepts, Methodologies and Applications. Chichester: John Wiley.
Wittgenstein, L. (1953). Philosophical Investigations. Oxford: Blackwell.