The rapid spread of capitalism around the world since the end of the Cold War in 1990 has been accompanied by a sharp increase in global business competition. Many computer technology advances have occurred in the same period, focusing on organizational computer networking. In such an environment, most organizations increasingly face two external pressures to produce and deliver higher quality and cheaper products and services than their competitors, and to adapt to the fast pace of technological change.
As organizational development approaches including total quality management (TQM) and business process reengineering (BPR) have taught us, one of the best ways to surmount these challenges is to make incremental or radical improvements in organizational processes, or sequences of interrelated activities. Although many organizations are presently structured around computer networks in a decentralized and geographically dispersed fashion, most research describing successful TQM and BPR projects makes little reference to the possibility of conducting process improvement in a distributed and computer-mediated fashion.
Current BPR research proposes computer technology as an enabler of radical process improvement, by showing how computer systems can be used to implement radically redesigned processes. Seldom discussed is how computer systems can actually support those who redesign processes. Despite this research oversight, two trends point to the importance of initiating computer-mediated process improvements: the accelerated diffusion of computer networks, which is dramatically changing the way teams communicate in organizations; and the increasing globalization of specialized knowledge.
Since the mid-1980s, computer networks have increasingly provided communication links within organizations through LANs and, more recently, intranets and between organizations through WANs and, more recently, the Internet and interorganizational computer systems. This worldwide diffusion of computer networks has facilitated a rapid increase in knowledge specializations and globalization. Today, one finds professionals with similar degrees of expertise in specific fields, such as expert systems design, or mutual fund management, living in Australia, Japan, the U.S., New Zealand, or Paraguay. The diffusion of computer networks and the globalization of specialized knowledge have allowed many organizations to become independent of geographical constraints. Many such organizations have restructured around communication networks, moving toward organizational paradigms emphasizing flexibility, knowledge accumulation and deployment, and distributed teamwork. One such paradigm is that of the virtual organization [3, 10].
Virtual organization has become one of the hottest management topics of the 1990s, particularly given the possibilities afforded by local- and wide-area computer networks. Advantages of moving from physically aggregated organizational units toward electronically linked ones have been highly publicized. Facilities rental and maintenance costs can be considerably reduced or eliminated. Employees can work from home or from distributed offices close to home. Company offices can spread over large geographical areas, allowing clients throughout the world to purchase goods, services, and information without traveling long distances. All this, and operations data can still be monitored in a centralized manner.
Most organizations using computers in a non-standalone way can be said to have at least some degree of virtuality. Naturally, the more distributed computer-mediated communicating occurring, both within the organization and with the outside world, the higher the degree of virtuality. Will a high degree of virtuality help or hinder organizations? Can processes be improved quickly enough to cope with competition and the drive for change facing virtual organizations today? Based on an analysis of 38 process improvement (PI) groups, this article addresses these questions.
Process Improvement Groups
Process-based improvement is the keystone of approaches such as TQM and BPR, which assume that process-related change, whether incremental or radical, can lead to increases in organizational efficiency and effectiveness. An analysis of these approaches strongly suggests the PI group is their primary change instrument.
Whether aimed at incremental or radical change, PI groups tend to be small, with 320 members; have generally well-defined group stages; and have defined roles played by specific group members. Two examples of PI groups illustrating these common characteristics are quality circles, widely used in TQM in Japan, [6, 12], and BPR groups .
Quality circles were originally developed in Japan in the early 1960s, as small groups of workers from the same organizational areas used statistical control methods to improve the quality of manufactured products. These early groups inspired the development of similar group-based organizational change approaches that underpinned the TQM movement in the 1980s .
Unlike quality circles, BPR groups usually involve line staff and managers from different areas of the organization, and use nonstatistical process modeling tools. Moreover, BPR groups are usually temporary, and may be dissolved as one or more processes are redesigned, while quality circles typically meet once a week over several months or years. Finally, while BPR groups search for radical process improvement, the degree of improvement sought by quality circles is usually incremental .
PI efforts are typically comprised of two stages: conceptual, and practical. In the conceptual stage, a process is conceptually redesigned. This stage consists of three main steps: Identification, Analysis, and Redesign. In Identification, a process is selected for redesign; in Analysis, the selected process is modeled and analyzed; and in Redesign, process changes are proposed.
The practical stage consists of two main steps: Implementation, and Routinization. In Implementation, redesign proposals generated in the conceptual stage are put into practice; and in Routinization, the process changes implemented in the previous step are consolidated as routine operations of the organization and undergo incremental optimizations. The practical stage is typically much longer than the conceptual stage, particularly when processes are radically redesigned (see  for an interesting illustration of this point).
Figure 1 illustrates the two main stages and related steps of a generic PI effort; the vertical dimension of the rectangles provides a general idea of the relative duration of the steps (but is necessarily out-of-scale since the practical stage is much longer than the conceptual stage).
This article focuses on 38 process improvement groups conducted in three organizations between 1993 and 1996. The groups ranged in length from less than a week to several months long. Twenty-six were conducted in a Brazil-based advertising firm with 60 employees and $3.5 million in annual revenues; six in a semiautonomous branch of the New Zealand government with 2,500 employees and $105 million in annual revenues; and six in a tertiary education establishment in New Zealand with 1,300 employees and $84 million in annual revenues. While some of the groups interacted face-to-face, most interacted predominantly through an email conferencing system. Several groups radically redesigned organizational processes; others generated modest and incremental process improvements.
I was particularly concerned with the conceptual stage of PI efforts in this study. Therefore, I collected research data about the formation of PI groups, the group identification of processes to be redesigned, the analysis of those processes, and the generation of redesign proposals by the groups. A comparatively small amount of research data was obtained regarding the implementation and routinization of process redesign proposals.
The majority of the 38 groups communicated predominantly through a simple asynchronous computer conferencing system I developed using XPost, and later Novell Groupwise. This system allowed PI group members to post electronic messages to a mailbox that automatically replicated the messages to the other PI group members. The system also allowed for one-to-one message exchanges between staff within and outside the group. A variety of computer files, including graphs and spreadsheets, could be attached to electronic messages. The conferencing system also allowed public access to PI group discussion archives.
Along with the conferencing system, groups were given a printed guide containing suggestions on conducting PI groups . The decision of how extensively to rely on the system and the guide was left to each group. I participated in all the PI groups, facilitating the use of the system. This allowed me to observe and interview members, and to collect discussion transcripts. I refrained as much as possible from influencing the content, volume of postings, and stage duration of the PI group discussions.
Most PI group members had participated in previous PI groups where meetings were exclusively face-to-face and involved members from one department only. This allowed me to analyze their comparisons between current and previous experiences.
PI Group Results
I found four technology effects of the PI groups to be particularly relevant for virtual organizations:
- A decentralization of incremental improvement initiatives;
- An increase in the possible number of PI groups per unit of time; and
- A reduction in group cost;
- A neutral effect on process redesign quality.
Decentralization of improvement. When I analyzed the more incremental PI groups (which formed the majority of the computer-supported PI groups) from a group leader seniority perspective, I noted an interesting pattern. The asynchronous and distributed nature of PI groups resulted in a considerable reduction in the demand for group leadership seniority, and in general, a decentralization of PI initiatives. In the Brazilian organization, for example, face-to-face PI groups had previously been led by the highest status members of the group. After the conferencing system was made available, however, some PI groups were led by junior staff that included the CEO and one of the company directors as ordinary members.
Similarly, in the New Zealand organizations, several lower-status staff members led their groups through successful analyses and incremental redesigns of processes, which resulted in improvements in process productivity and outcome quality. These group leaders agreed unanimously that distributed computer mediation made it much easier for them to lead their groups.
With the decrease in the demand for leadership seniority enabled by computer support, PI groups could be formed quickly to eliminate process imperfections known to those involved in process execution, but often unrecognized by managers. It was interesting to note the surprise of some managers when serious interdepartmental process-related issues were brought up, discussed, and pragmatically solved in computer-mediated PI groups, without need for managers to coach group members.
The CEO and several senior managers of the organization in Brazil reported great satisfaction with the outcomes of computer-mediated PI groups. Several managers who browsed through records of computer-supported PI group discussions said they were impressed by worker awareness of company problems and their willingness to find solutions. Most managers had thought this type of awareness was nonexistent, since few employees previously spoke out in face-to-face situations.
The computer-supported PI groups also highlighted the previous dearth of communication between managers and subordinates, which often caused simple process-related problems to go unnoticed until they became more severe. PI groups were seen as helping to ease tensions between managers and line staff tensions perceived by both managers and line staff as detrimental to cooperation in inter-level PI efforts.
The three PI groups that were strikingly radical in terms of process redesigns implemented were least likely to embrace decentralization of improvement initiatives. These groups incidentally met two of the criteria proposed by Davenport and Stoddard to unequivocally characterize BPR: An orientation to broad, cross-functional processes; and the need for radical change in process performance. Members of these three groups, led by their company CEO, unanimously refused to conduct any part of their group discussions through the computer conferencing system. They perceived a higher risk associated with contributing electronic postings in comparison with making oral contributions. As a PI group member remarked:
- The problem with [an electronic message] is that it is very easy to think of it as a form of conversation, and yet it is also a written record, and people can easily write something down in [an electronic message] that they later regret. In a verbal situation, an oral situation, people tend to be a bit more accepting of people saying something inappropriate, and if they say something inappropriate often the cues of everyone else will protect them.
Increase in PI groups. In the Brazilian organization, PI groups were forced to meet exclusively face-to-face for about three months while the LAN and computer conferencing system were installed. In the quarter immediately after the conferencing system became available to the PI groups, the number of completed PI groups nearly doubled compared to the previous quarter, from eight to 15 groups.
This increase may be attributed to a novelty effect in the Brazilian organization, but the same cannot be said of the two New Zealand organizations. Here, novelty effect was minimal, since the conferencing system was seen as an extension of email, a tool most group members used daily. In fact, nearly all respondents interviewed referred to the conferencing system as "the email system."
In the New Zealand government branch, approximately 78% of those interviewed, thought computer-supported PI groups had been completed in less time than face-to-face PI groups. The frequency distribution of responses illustrated in Figure 2 suggests a statistically strong trend (p < 0.001) toward the perception that computer support leads to a reduction in group duration which clearly contributes to an increase in the number of possible PI groups per unit of time.
The New Zealand government respondents attributed the reduction in PI group duration mainly to the reduction in group set-up time, or the time needed to bring group members physically together, enabled by the computer support, particularly for groups with members from geographically disparate sites.
In the tertiary education institution in New Zealand, an even more interesting effect was observed regarding the possible number of PI groups per unit of time. This institution had originally established "official PI days" during which staff and faculty were expected to engage in PI group discussions full-time. Groups analyzed current processes and organizational policies, proposing changes to improve the organization. Although these PI group efforts were typically seen as very successful, their frequency was low usually only twice a year. As soon as the conferencing system became available to staff and faculty, five groups were conducted over less than a quarter (an effect similar to that observed in the Brazilian organization). Several PI group members pointed out the availability of the conferencing system made it more efficient for them to set up and conduct PI discussions with minimum disruption for PI group members, and this had been the main reason why they had participated in the computer-supported PI groups.
Group cost reduction. PI group members in the Brazilian organization perceived a reduction in the organizational cost of groups as a benefit of computer support. They noted their easier access to previous group archives, and to generic process-related information, thanks to the conferencing system. Easy access to this information enabled new groups to replicate certain solutions, allowing them to spend less time in process analysis and redesign. While this access to previous solutions may have decreased creativity somewhat, it was beneficial from a cost-savings perspective. In addition, access to process-related information such as customer satisfaction, throughput, and lead-time, helped members identify PI opportunities, and consequently, reduced costs.
PI group members at the tertiary education institution in New Zealand consistently saw group cost as considerably reduced by use of the conferencing system. Their explanation was more direct than the Brazilian organization's, and was related to group communication costs. Group members noted the average time spent in face-to-face PI group discussions was far higher than that spent in computer-supported PI groups. Group leaders, however, did not report this time reduction.
In the New Zealand government branch, structured interviews indicated that approximately 78% of respondents thought computer-supported PI groups cost far less than face-to-face PI groups, as illustrated in Figure 3. The distribution of response frequencies suggests a statistically strong trend (p < 0.001) toward the perception that computer support leads to a reduction in the cost of PI groups.
Group members estimated slightly over 20 hours were needed for exclusively face-to-face meetings, while 1.5 hours was the average time spent by ordinary group members in the computer-supported PI group discussions a reduction of approximately 93%. The average time for group leaders and the facilitator was reduced from 20 to approximately seven hours, according to direct time measurements initially taken by the researcher and then matched against figures provided by the group leaders a time savings of 65%.
This time savings can also be said to have resulted in cost savings. If we conservatively assume a group member costs the organization $15 per hour, and the average number PI group members is nine (based on the PI groups facilitated in this organization), then the cost savings from computer conferencing support would be $2,415. Note that the cost of the facilitator and possible travel expense savings are not included.
Process redesign quality remains stable. Although several group members from the Brazilian organization observed the conferencing system had a strong impact on PI group efficiency, no significant trend was observed regarding variations in PI group outcome quality. The general satisfaction with the overall results achieved by the PI groups, is illustrated by this remark by a senior manager: "...we have never been through such a successful motivational endeavor since the firm was founded..."
I believe a significantly negative impact on process redesign quality would have been noticeable in in-depth interviews. Since this was not the case, I conclude the overall technology effect on process redesign quality at the Brazilian organization was either positive or neutral.
A slight trend toward a perceived increase in process redesign quality was noted in both New Zealand organizations, and more strongly in the New Zealand tertiary education institution. (See Figure 4 for the frequency distribution of answers regarding process redesign quality in both organizations.) But the chi-square analysis does not show a statistically significant trend toward a perceived increase in process redesign quality in either organization. However, the Figure 4 data suggests a strong trend toward the perception of a non-negative effect on process redesign quality. Seventy-two percent of the respondents in the New Zealand tertiary education institution, and 66% of those in the New Zealand government branch perceived either an increase or a null effect in process redesign quality.
Most respondents who perceived an increase in quality attributed it to an improvement in the quality of individual contributions from PI group members interacting through the computer conferencing system as one group member remarked:
- You think more when you're writing something, so you produce a better quality contribution. Take for example what [member's name removed] wrote, she wrote a lot and it seemed that she thought a lot about it before she emailed it to the group. She wasn't just babbling off the top of her head, she tended to think out what she was writing. I know I did it a lot, especially my first message. I really thought a lot to put it together.
Those who perceived a decrease in process redesign quality tended to think use of the conferencing system increased the ambiguity of PI group discussions, particularly during analysis (see Figure 1), when the group analyzed the target process(es) for redesign. During the analysis step, group members must build a shared understanding of the process being analyzed to effectively contribute process change suggestions. As most groups had a multidepartmental composition, variations in technical language abilities among members sometimes caused communication problems. Some members thought the asynchronous nature of the conferencing system, which prevents immediate feedback and the use of nonverbal cues such as gestures, made it more difficult to remove obstacles to shared understanding.
However, the percentage of respondents who thought redesign quality had suffered through use of group technology was comparatively low, suggesting the higher ambiguity caused by the electronic medium may have been offset by the higher quality of individual contributions, leading to a neutral overall effect on process redesign quality.
Implications for Virtual Organizations
These findings suggest organizations with a high degree of virtuality can benefit from computer-supported group PI efforts. This is particularly true for groups searching for incremental improvements in organizational processes as opposed to radical (or BPR-like) improvements. BPR groups are more likely require face-to-face meetings for most of their life duration, as higher perceived risks are associated with conducting these discussions through permanent-record types of electronic media.
Distributed and multidepartmental PI groups searching for incremental quality and productivity improvements can be successfully conducted with computer support. They can be conducted in less time and at a lower cost than equivalent face-to-face groups. They can also occur more often, due to their lower reliance on management leadership, than similar face-to-face groups. And all this can be achieved without any significantly negative impact on the quality of the process improvements generated.
This is good news for organizations with a high degree of virtuality, as well as for those moving toward virtuality. Such organizations should consider implementing computer-supported PI projects targeted at achieving incremental improvements in process quality and productivity. Although BPR-like projects may not benefit as much from distributed computer support, these groups can still benefit from a combination of face-to-face and computer-supported interaction (the key challenge here would be to find the right mix). Moreover, the economic revolution that has taken place in Japan between the end of World War II and the early 1980s suggests a large number of incremental improvement initiatives can lead to quantum-leap competitive advantage over the long run.
5. Gallivan, M.J., Hofman, J.D., and Orlikowski, W.J. Implementing radical change: Gradual versus rapid pace. In Proceedings of the 15th International Conference on Information Systems (Vancouver, Canada, Dec. 1417). ACM Press, New York, 1994, 325340.
9. Kock, N. and McQueen, R. Asynchronous groupware support effects on process improvement groups: An action research study. In Proceedings of the 17th International Conference on Information Systems (Cleveland, OH, Dec. 1618). ACM Press, New York, 1996, 339355.
11. Pickering, J.M. and King, J.L. Hardwiring weak ties: Interorganizational computer-mediated communication, occupation communities, and organizational change. Organization Science 6, 4 (1995), 479486.
©2000 ACM 0002-0782/00/1100 $5.00
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2000 ACM, Inc.