FACTORS AFFECTING THE SUCCESS AND FAILURE OF AN INFORMATION SYSTEM FOR PRODUCT OUALITY*

Size: px
Start display at page:

Download "FACTORS AFFECTING THE SUCCESS AND FAILURE OF AN INFORMATION SYSTEM FOR PRODUCT OUALITY*"

Transcription

1 INTERFACES Vol. 8, No. 2 February 1978 FACTORS AFFECTING THE SUCCESS AND FAILURE OF AN INFORMATION SYSTEM FOR PRODUCT OUALITY* Daniel Robey Florida International University Tamlami Trail Miami, Florida and Richard L. Zeller General Electric Company ABSTRACT. This paper analyzes the adoption and successful use of an information system by one department, and the rejection and failure of the same system in a similar department of the same company. Since the technical features of the system and the work performed are identical in both departments, the reasons for success and failure are traceable to organizational and human differences. Indeed, the technical features of the system studied are entirely irrelevant to the adoption and rejection of the system. In many studies of implementation, the MIS parameters are not well controlled or measured, and many types of organizations are studied together, in our study these sources of experimental error are controlled, thereby focusing attention on the individual, organizational, and process differences between the adopting and rejecting departments. Introduction In recent years the problems of implementing management information systems and OR techniques have received greater attention from both systems designers and behavioral scientists. The books by Lucas [1] and Schultz and Slevin [4] show the current awareness of behavioral and organizational factors in systems success. As Mason and Mitroff [2] observe, an information system is composed not only of hardware and software, but also a person, his or her psychological characteristics, the nature of the decision problem, the organizational context, data, and the mode of data presentation. Many efforts to implement MIS make rather simplifying assumptions about these factors, and as a consequence implementation may fail for behavioral reasons having little to do with a system's technical features. The purpose of this paper is to illustrate the relevance of some nontechnical variables in system implementation. Description of the system The research was conducted at two neighboring plants ofa division ofa very large U. S. corporation. Tlie division manufactures a wide variety of highly technical equipment for the health care industry, and the demands for quality are enormous. Consequently the Product Quality Section has constantly sought techniques to improve product quality and reduce quality costs such as inspection and warranty re- Portions of this paper were presented at the 8th Annual Conference of the American Institute for Decision Sciences, San Francisco, November Copyright 1978, The Institute of Management Sciences 70

2 pairs. With this goal in mind the development of the Quality Information System (QIS) was initiated in late Similar models were in use in other divisions ofthe corporation, but the model concept was new to the division in question. The primary objectives of the QIS were stated as follows: 1. reduce the cost and time required to gather quality data; 2. provide detailed data and history on problem parts, areas, and defects for corrective action purposes, thereby reducing scrap and rework costs; 3. provide accurate and timely data which would improve management visibility into quality performance; 4. provide more efficient use of inspection resources to concentrate on problem areas. The QIS used inputs from various inspection operations to compile data into six reports on a daily basis, with summary reports being generated each week, month, and year. The six reports were as follows: 1. Operator Defect Distribution, 2. Product Effectiveness, 3. Quality Report, 4. Top Five Defects by Part Number, 5. Top Five Defects by Operator Number, 6. Top Five Defects by Type. These various reports were used to determine department quality levels, identify problem areas and assist the allocation of inspection manpower to facilitate corrective action. The QIS model was introduced into the two separate quality departments of the neighboring plants in April of The organization charts ofthese departments are shown in Figure 1. After approximately six months the model was rejected and discontinued by one department and accepted and used extensively in the other department. We therefore sought to determine the factors associated with the success and failure of the system in the two locations. Research methods To learn about these factors we conducted interviews with personnel in each department. Interviews were supplemented by an attitude questionnaire developed by Schultz and Slevin [3]. The questionnaire identifies seven areas that system users are concerned with. Responses show the user's perception of how favorable the system is for each area of concern. Questionnaire data can thus provide evidence of specific user problems. For the QIS, the seven areas of concern measured by the questionnaire were: I. Effect of QIS on user's job performance and perfonnance visibility. II. Effect of QIS on interpersonal relations. III. Effect of QIS on organizational changes. IV. Effect of QIS on goals. V. Support for QIS from top management and other departments. VI. Relationship between QIS users and QIS developers. VII. Urgency and importance of QIS in the organization. INTERFACES February

3 ; ; 5 ff > [ m 1 I 9 m > m -* Ih 1II * tit * INTERFACES February 1978

4 The questionnaire was administered in October to four users in the adopting group and to seven users in the rejecting group. Since these attitudes were assessed after the respective departments had chosen to adopt and reject QIS, one cannot logically identify attitudes as the cause of success or failure. However, it was possible to show which attitudes were the important ones. The purpose of the interviews was to gather general information about the implementation process in each department. In addition, the interviews and company documents were used to identify some differences in structure between the departments. Attitudes of users Daij from the questionnaires were analyzed with the Mann-Whitney U test. Basically, this nonparametric statistical test compares the two groups to determine if one group ranks higlier than the other on each attitude. A significant difference here means that the adoptors bad more favorable attitudes than the rejecting group. The results of this analysis (Table 1) show that the department which adopted the Quality Information System was more favorable on attitudes I and VII: the effect ofthe QIS on individual performance and performance visibility, and the perceived urgency and importance of QIS, respectively. No other attitudes were significantly different between the groups. TABLE 1. (Differences in attitudes between the adopting group and the rejecting group) Area of Concern Level of Statistical Significance I. Performance.055 II. Interpersonal Relations n.s.* III. Organizational Changes n.8, IV. Goals 113. V. Support from Others n.s. VI. User-Developer Relationships n.s. VII. Urgency and Importance.036 No significant difference between groups, according to Mann-Whitney U Test results. Conctusion: The differences on Attitudes I and VII indicate that the adopting group's attitudes were more favorable toward QIS than were the rejecting group's. This analysis provides some understanding of group reactions to the implementation of the QIS. While we cannot say that attitudes were the cause of behavior in this case, certain attitudes, and not others, are related to behavior Departments differed on the utility of QIS in aiding personal performance and on the importance and urgency of QIS implementation. Other attitudes having to do with interpersonal relations, organizational changes, goals, outside support, and user-developer relationships were not discriminating factors between the groups. It would appear from these INTERFACES February

5 results that a primary user concern is the role of MIS in improving user performance, and that many of the other hypothesized concerns are less important and perhaps not critical to eventual implementation success. Organizational differences A review of the interview results and the organization charts ofthe departments (Figure 1) demonstrates the impact that selected structural variables had on implementation. As Figure i shows, the department which rejected QIS is larger and consists of a greater number of occupational specialities than are present in the adopting department. Other research on innovation shows that this complexity can bring a variety of bases of information and knowledge to bear that can increase the probability that innovative changes will be proposed. However, at the implementation stage., high complexity also makes it more difficult for the organization to handle conflict and implement smoothly [5]. It is understandable that successful implementation was achieved only in the less complex department. Differences in formality of structure were also found between departments. The group which rejected QIS had recently been reorganized into two groups-engineering and Inspection (See Figure 1). Related to this change, very few formal rules had yet been estabished to govern the activities of members. At the implementation stage of innovation, low formalization is likely to lead to role conflict and role ambiguity [5]. Since implementation requires a clear execution of responsibilities, low formalization could have been a factor contributing to the failure of QIS in this department. The interviews also showed that the adopting group had more centralized authority than the rejecting group. Decentralization appears to be more appropriate in gathering and processing information during the initial stages of change. During implementation, however, the more centralized authority structure in the adopting group was probably a strong factor in this group's positive decision. Subordinate users were well aware of management's position on QIS, and this probably led to adoption. In the other group, strong support of a centralized figure was absent, and the QIS failed. The implementation process The interviews with members of both departments revealed some differences as wel! as similarities in the process of implementation. One interesting result was that no formal implementation program was used in either department. The developers of the QIS neither introduced it in a planned manner nor followed the system through the implementation process. A manual on QIS had been "dropped off on people's desks, and users were asked to give their reactions to the system. "Hie initial reaction of people in both departments was one of uncertainty. At this point some differences between the two departments emerge clearly from the interviews. In the department which eventually accepted and used the system extensively, the Manager of Quality Control took the initiative in explaining QIS to the other prospective users and strongly supported its use. This person had been exposed previously to the system in another application. In contrast, considerable conflict existed in the group which eventually rejected QIS. This conflict was associated with a general mistrust of the accuracy of QIS outputs and the cost ofthe system 74 INTERFACES February 1978

6 in comparison with its value. No one was influential enough to convince the others that the QIS should he adopted by the department. This comparison of the two processes of implementation reveals the importance of a strong supporter of MIS within the user group. It also reinforces our discussion about the role of organizational structure in successful implementation. The larger size and greater complexity of the rejecting department made it more difficult to resolve conflicts, particularly since the QIS had no formal introduction and no strong supporter. In the smaller and less complex department, strong management support emerged, and implementation met no serious barriers at any stage in the process. Conclusions Our case indicates the importance of several factors in MIS implementation. 1. At the individual level, certain attitudes were found to be more important than others. 2. The organizational factors of complexity, formality, and centralization also affect implementation. 3. Lack of involvement by system developers is not sufficient to ensure failure if the vital function of explaining the system to ultimate users is assumed by some other knowledgeable person in the group. 4. Strong management support is instrumental to system adoption. Although the research reported in this paper is of Umited scope, it emphasizes the need for a rather broad analysis of the factors affecting implementation of MIS. As a research strategy, the comparative case study reported here strikes a balance between large sample survey techniques and one-shot case studies. It neither possesses the severe limitations of these forms of research nor makes full use of their advantages. The essential advantage of our research design is its ability to compare attitudes, organization structure and implementation process while holding the MIS features, production technology and corporate policy constant. By controlling the effects of these other factors, the effects of the behavioral and organizational factors can be more readily identified. REFERENCES 111 Lucas, Henry C. Why Information Systems Fail, New York, Columbia University Press. 197S. 121 Mason, Richard O. and Mitroff, Ian I., "A Program for Research on Management Information Systems," tilanagement Science, Vol. i9. No. S (January. 1973), pp I 3] Schultz, Randall L. and Slevin. Dennis P., "Implementation and Organizational Validity: An Empirical Investigation," in Schultz and Sievin, eds., Implementing Operations Research/ Management Science, New York. American Elsevier, i975. [4] and, eds., Implementing Operations ResearchlManagement Science, New York. American Elsevier, (5 Zaltman, Gerald, Duncan. Robert, and Hoibek, lonny. Innovations and Organizations. New York, John Wiley, i973. INTERFACES February 1978

7