Jonathan Grudin

University of California, Irvine / University of Oslo (visiting)
Information and Computer Sciences
Irvine, CA 92697-3425

Phone: (714) 824-8674   in Oslo: +47 22852429
Fax: (714) 824-4056    in Oslo: +4722852401
Email: grudin@ics.uci.edu
URL: http://www.ics.uci.edu/~grudin/


Computer-Supported Meeting Scheduling


Computer support for meeting scheduling is a focus of active research in several areas, including distributed AI and intelligent agents, software engineering, and information systems. It is also of very high applied interest: After a decade of very modest success, commercial scheduling software is taking hold in some organizational settings. In fact, meeting scheduling is most significant for being one of the first desktop group support applications that is widely used in some large organizations.

We are conducting empirical studies of the use of this group support technology in settings where existing technical infrastructure and organizational practices enable all or most employees, numbering hundreds or thousands, to fully participate in the use of shared systems. A deep understanding of successful practice will extend our knowledge of organizational behavior in this new technological context, and can guide the research and application of technologies and approaches developed in the diverse fields mentioned above. We are conducting detailed interviews and broad surveys, and plan longitudinal observational studies, to determine:


Domain. Our particular focus is on the use of on-line calendar applications and their involvement in scheduling meetings and coordinating activity. On-line calendaring has been available for over fifteen years, but it has not been widely used. Even in organizations where some people keep calendars on-line, most people work with paper calendars or no calendar at all. This greatly limits or precludes the use of on-line calendars as a coordination tool. Now, at some firms, calendar application use has reached a "critical mass" that extends beyond developers to include managers, administrators, and in many groups all employees. This level of use creates a range of new coordination activities centered on the technology, with the potential for discernibly changing the culture within these organizations.

Application of results. These studies have immediate theoretical value in addressing unresolved issues surrounding the adoption of group support technologies, which in turn should influence approaches to their design. Our initial studies have been in the high-technology firms, that although special are the first to widely deploy and successfully use these technologies. In the future, this level of ubiquitous access may be present in a range of organizations. Our initial observations reveal patterns of successful use that differ across high-technology firms, so the results may provide a broad understanding of the choices and outcomes that these technologies will offer diverse organizations.

Summary of work in progress. We have collected data, now in various stages of analysis, at Microsoft, Sun, Boeing, and Lawrence Livermore National Laboratories. Each of these organizations deploys a range of group support technologies including calendaring and scheduling that are used by many or even most employees. The first of these studies have provided answers or tentative answers to some important, decade-old questions related to groupware adoption and use.

Single-user applications are designed with a 'discretionary use' model. In contrast, for large systems, upper management support has been considered crucial to adoption. Which applies to groupware? The relatively low cost of groupware reduces high-level visibility, but some have argued that social dynamics will force mandated use if success is to be achieved-the large system approach. Our initial interview studies of on-line meeting scheduling at Microsoft and Sun found successful, near-universal use achieved without managerial mandate. Versatile functionality and ease of use associated with discretionary products appeared to be factors leading to adoption. Individuals began adopting calendar use because of personal information management features. The presence of an organization-wide infrastructure then made group uses possible, and substantial peer pressure developed over time and seemed to have a substantial effect in achieving near-universal use in these groups. Following the interviews we distributed surveys which were returned by over 2000 employees in these organizations. The analysis of these surveys is now being completed and provides support as well as elaboration for the interview results.

This study was undertaken without even a conjecture as to whether a top-down mandate or bottom-up adoption path would be found. The result, if general, is crucial for the design of groupware. If top-down mandate is necessary, the designers might focus on providing maximal benefits for the primary beneficiaries of meeting scheduling in management. Based on this data, the emphasis should instead be on providing benefits for those least profiting from ease of scheduling, individual contributors. From the perspective of a technology adopter, it is still advantageous to have the strong support of management when introducing a technology. Nevertheless, we can surmise that emphasis in documentation and training should be given to the individual productivity features, something that often does not occur when the training is aimed primarily at the administrative assistants or secretaries who will be the heaviest users.

An unexpected finding was that although Sun and Microsoft use very similar calendar applications, they are used in markedly different ways in the two organizations. A key distinction is the extent of access that one person in the organization allows as a default to anyone else in the organization. In Microsoft, the application default and standard practice is to allow a random person to see only what time is free and what time is taken or booked. (This default might be overridden for a particular individual, to provide them with greater or even full access.) In Sun, the application default and standard practice is to allow others to see all of the details of one's meetings (unless this default is overridden, as in the case of a private or sensitive meeting).

An intriguing aspect of this difference is that the two cultures have very different practices around scheduling meetings, practices that seem to be correlated with the different application defaults. At Sun, the scheduler typically has access to a lot of information about the person or persons to be scheduled, uses this to find a time likely to be convenient, and then often contacts people through email. At Microsoft the scheduler has much less information and simply schedules the meeting in available time, leaving the attendees to reject or accept the meeting. The latter more abrupt approach may be efficient if most meetings are accepted and if everyone has their calendars up to date. The former approach is gentler and may also be more efficient, if the scheduler makes good use of the others' calendar details.

The third organization, Boeing, was studied through interview and observation over a period of two months. At Boeing, far fewer individual contributors maintain personal calendars on-line. However, there are tens of thousands of managers at Boeing, and among then one can find large concentrations of calendar use. Seven competing calendar products are reported to have had user populations of 1000 or more. Several results emerging from the preliminary analysis of this data are contributing to the assessment of the survey data from the earlier studies. For example, we identified at least three distinct "cultures" of use at Boeing: the individual contributors who do use calendars, the middle management and their office administrators (secretaries), and executives and executive secretaries. Features and default settings that are valued differ across groups.

In the specific case of access privileges, both of the approaches described earlier have been widely used at Boeing, due to the nature of the applications acquired. The users of Digital's All-In-One are restricted to the more closed approach (only free/busy information available), whereas the most widely used calendar, OfficeVision/Profs, is almost invariably used to permit "read access" to the detailed calendar information of other users, except for meetings marked as private.

From this, we have surmised that in fact the choice of default settings can greatly shape the uses to which a technology is made, and that the choice of defaults is often made in a remarkably arbitrary fashion. This may not have been so consequential for large systems with few users, where changes were either prohibitively difficult (and thus not an issue) or quite easy to propagate. Now we have technologies where the choices of default settings are often simple to make, but can have substantial consequences due to the number of users. In addition, we see evidence of path dependence: behavior patterns established around a technology may be difficult to reverse.


NSF "NUGGETS"

Getting To Critical Mass. When everyone in a group -- everyone -- uses a technology, we may see changes in organizational functioning that far exceed those wrought by computer technologies of the past. However, it is rare that we find everyone using an application.

Technology to support groups often runs afoul of the problem that it would be helpful if enough people began using it, but it does not help the first few in the group, so they are likely to abandon it before it reaches the necessary "critical mass" of group members. Electronic calendars and meeting scheduling applications are an example. Meeting scheduling could be very helpful, but only if almost everyone keeps their calendars online. What can be done if a lot of people prefer paper calendars? Apparently very little -- one survey found that meeting scheduling was rated the most widely available but least useful "groupware."

Times are changing. In some major organizations, virtually everyone is keeping their calendar on-line and using it for scheduling, with great potential benefit. How is critical mass reached? It was suggested a decade ago that "management mandate" might be required to obtain the collective benefit. Instead, more in keeping with flatter orgaizations, we find that success has resulted when applications designers set aside the "group appeal" and focus on providing personal benefits for each group member.

Once a certain percentage of group members are using the application for the individual benefits, they realize the advantages that would result from univrsal use, and go about convincing their colleagues to "get with the program." In studies of calendar and scheduling use at Microsoft and Sun Microsystems, it was found that groupware adoption occurred in a "bottom-up" fashion. Some people felt that "peer pressure" was too strong a term -- but their peers felt pressured.

This has consequences for designers: They must focus on providing benefits for individual contributors, not only for the managers or others charged with recognizing collective benefits and encouraging group compliance. It has consequences for users, too: Training must emphasize the benefits of personal information management. And it has consequences for all of us, because we are seeing that indeed, when everyone in a group uses a technology, it can have profound effects on the ways they work and think and even talk. "When can we meet?" someone asks. The response at Sun is likely to be "browse me!" and at Microsoft "Schedule Plus me!" as the responding party hurries off. Let the computer find the compatible time.

Return to ITO Workshop Abstracts

Return to ITO Workshop Home Page