SHAPING ELECTRONIC COMMUNICATION:

The Metastructuring of Technology in Use

Wanda J. Orlikowski, JoAnne Yates,

Kazuo Okamura, and Masayo Fujimoto


MIT Sloan School Working Paper #3611-93

Center for Coordination Science Techincal Report #155

April 1994





Wanda J. Orlikowski, JoAnne Yates, Kazuo Okamura, Masayo Fujimoto



Shaping Electronic Communication:

The Metastructuring of Technology in Use




Wanda J. Orlikowski and JoAnne Yates

Sloan School of Management

Massachusetts Institute of Technology

50 Memorial Drive

Cambridge, MA 02139, USA

wanda@mit.edu jyates@mit.edu

Kazuo Okamura

Matsushita Electric Industrial Co., Ltd.

1006 Kadoma, Osaka 571, Japan

okamura@isl.mei.co.jp

Masayo Fujimoto

The Sumitomo Research Institute, Inc.

27-2 Shinkawa 2-chome

Chuo-ku, Tokyo 104, Japan




April 1994



ACKNOWLEDGMENTS

The authors would like to thank Acorn participants who provided the data for this study. Thanks are also due to Tom Callaghan for his able research assistance. The research support of the Center for Coordination Science at the Massachusetts Institute of Technology is gratefully acknowledged.


Shaping Electronic Communication:

The Metastructuring of Technology in Use


Table of Contents




Abstract

In this paper we suggest that the use of computer-mediated communication technologies in new and fluid organizations can be facilitated by the explicit and ongoing adapting of those technologies to changing contexts of use. In an exploratory study on the use of a computer conferencing system in an R&D setting, we found that the new medium's effectiveness was significantly influenced by the intervention of a few individuals who took on a role we label technology-use mediation. These mediators shaped everyday use of the conferencing technology, modifying the technology as well as the context of use to promote effective electronic communication. Drawing on the insights of this empirical study, we develop a theoretical framework that views technology-use mediation as influencing how users structure their communication technologies, and hence as one form of metastructuring. We believe that the role of technology-use mediation constitutes a valuable mechanism for providing the ongoing attention and resources needed to contextualize what are often generic computer-mediated communication technologies to the shifting conditions of dynamic organizational forms.

A number of commentators have recently pointed to the profound changes -- economic, technological, and social -- that contemporary organizations are facing and implementing (Drucker, 1988; Lewin and Stephens, 1993; Malone and Rockart, 1991). For example, many organizations are attempting to transform their structures and processes through teamwork, global integration, and networking. Underpinning such changes are fast, accessible, and ubiquitous electronic networks that support computer-mediated communication technologies such as electronic mail, computer conferencing, and groupware. Such communication networks are expected to enable organizational members to work more flexibly, to span contexts and boundaries, and to collaborate more effectively. That expectation is widely shared, yet as Daft and Lewin (1993) point out, we still have much to learn about the interactions among computer-mediated communication technologies, new organizational forms, and changes in work and communication. In this paper we provide a framework for thinking theoretically about some of these interactions.

We suggest that the use of computer-mediated communication technologies in dynamic organizations can be facilitated by explicitly contextualizing those technologies and the social norms around their use to particular organizational circumstances over time. Conventional models of technology implementation, perhaps because they are premised on more stable forms of organization, generally assume that accommodation to the context of use occurs at the time of a technology's implementation, and that ongoing adaptation is unnecessary or adequately covered by routine maintenance. This assumption is problematic especially in the context of changing organizational forms. Such organizations are expected to be more fluid than their predecessors, changing frequently in response to changing circumstances. To keep up with such changes, the communication technologies supporting work and interaction in new organizations will also require frequent adaptation to ensure their continued effectiveness in changing contexts.

Further, many computer-mediated communication technologies are general purpose media that facilitate a range of possible interactions. This open-endedness offers benefits of flexibility but also creates the possibility that -- without adaptation -- the technology will not reflect local conditions or communication norms and hence will be underutilized or inappropriately utilized. The norms for interacting within computer-mediated communication technologies may not be shared a priori, and ambiguity or misaligned expectations may arise around how to use these technologies. In these cases, contextualization can effectively establish appropriate social protocols for communication in the new electronic media.

In this paper we identify a role -- technology-use mediation -- which we define as the deliberate, organizationally-sanctioned intervention within the context of use which helps to adapt a new communication technology to that context, modifies the context as appropriate to accommodate use of the technology, and facilitates the ongoing usefulness of the technology over time. The nature and importance of this role emerged from a study we conducted into the introduction and use of a asynchronous computer conferencing technology in one R&D project over an extended period. In this study, we found that the effectiveness of the conferencing technology was significantly influenced by the intervention of a few individuals who explicitly took on the role of technology-use mediation. Because technology-use mediation shapes how primary users structure their use of a technology, we suggest that this role may be interpreted more broadly as a process of metastructuring.

Our study and the metastructuring perspective we develop provide an empirically-grounded framing of the critical role played by individuals -- whom we refer to as mediators -- who deliberately adapt computer-mediated communication technologies to contexts of use. For research, the notion of metastructuring provides a theoretical lens for understanding the role and influence of activities that intervene in the process of technology use. In practice, the sustained effectiveness of computer-mediated communication technologies in flexible new organizational forms will depend in great part on ensuring that the technologies are relevant to specific contexts of use, and that they continue to be relevant as those contexts change. Thus, the role of technology-use mediation as an explicit, sanctioned form of metastructuring, constitutes a valuable mechanism for locating and providing the ongoing attention and resources needed to contextualize technologies to the shifting conditions and practices represented by dynamic organizational forms.

Prior Research

The literatures concerning interventions in technology use and technology structuring are relevant to this research.

Technology Intervention

Various studies have looked at the actions and influence of individuals who intervene in some aspects of users' use of technologies. For example, organizational and information systems researchers have studied the role of technology champions in facilitating the adoption of technological innovations (Beath, 1991; Howell and Higgins 1990; Maidique, 1980; Schon, 1963). This literature suggests that a technological innovation is more likely to succeed when it is vigorously promoted and endorsed by an influential individual, the champion. Training, typically conceived of as part of the initial implementation of systems is another form of intervention that has received attention (e.g., Davis and Olson, 1974). More recently people have recognized the need for continuous training for effective technology use (Johnson and Rice, 1987; Strassman, 1985). In practice, however, information technology training is often still restricted to the introduction phase. In a study of groupware use Bullen and Bennett (1990) found that training was typically offered at the point of initial implementation and focused on the mechanics of use, not the rationale behind such systems. Similarly, Orlikowski (1992a) found training around a new groupware product was likewise limited in content.

Study of ongoing intervention in information technology use has also identified the influence of other organizational actors. Culnan (1983) identifies surrogate users, labeled chauffeurs, who provide information directly to individuals. Bullen and Bennett (1990) observe that in organizations that recognized "the evolutionary nature of a person's use of software," designated support staff were organizationally authorized to provide ongoing guidance. In organizations without such sanctioned roles they found that expert users or local gurus emerged to fill the need. Bjorn-Andersen, Eason, and Robey (1986) note the emergence of two roles providing ongoing user support once technology is in place: systems staff (primarily for technical problems) and local experts (accessible and proficient users with knowledge of other users' needs). Similar to expert users or local experts are Mackay's (1990) translators, proficient users who shared their software customizations with less proficient colleagues. Administrative and social mechanisms for shaping use have also been identified. Sproull and Kiesler (1991) discuss ways of managing computer-mediated communication through the setting of policies and the development of self-policing social norms. In a detailed examination of the IBMPC computer conference over time, Foulger (1990) discussed the actions of conference administrators serving a community of thousands of IBM employees. These administrators established a limited number of electronic usage rules initially and over time as issues arose; they then reviewed messages to ensure conformance to these rules.

Our conceptualization of technology-use mediators -- as organizationally-sanctioned intervenors within the context of use who facilitate the establishment and ongoing use of communication technology over time -- has both similarities and differences with these existing understandings of champions, trainers, chauffeurs, designated support staff, translators or local gurus, and administrators. Champions and trainers typically intervene in the initial stages of technology implementation but rarely play an ongoing role. Chauffeurs do not mediate users' interactions with technology but relieve users of the need to interact directly with the technology. Translators and local experts or gurus are not organizationally recognized. Thus none of these meet our definition of mediators, though all engage in a broader process that we will identify below as metastructuring. Designated support staff and administrators, however, more clearly resemble our mediators, and their roles may more generally be understood as aspects of technology-use mediation. In addition to serving specific functions such as education, support, and administration, these roles also have a significant influence on how users appropriate their technologies and how those technologies are contextualized in certain conditions of use.

Technology-use mediation is an effective mechanism not only at the initial point of a technology's implementation but also and particularly during the ongoing use of that technology by users. [1]Our investigation suggests that the value of technology-use mediation in dynamic organizational forms lies precisely in its nature as a sanctioned and deliberate ongoing activity situated in the context of use, enabling more rapid adaptations of communication technology and technology use to accommodate changing circumstances, communication practices, and organizational forms.

Technology Structuring

As a role that influences the interaction of users with their technology, mediation can also be viewed from the framework of structuration theory (Giddens, 1984), and particularly in terms of how technologies are structured by users in their contexts of use (Barley, 1986; DeSanctis and Poole, 1994; Orlikowski, 1992b; Walsham, 1993; Weick, 1990). The structuring of technologies in use refers to the processes through which users manipulate their technologies to accomplish work, and the ways in which their actions draw on and reproduce (or sometimes change) the particular social contexts within which they work.

Technology structuring is influenced by users' interpretations of their work, the organization, and technology, their access to organizational and technological resources, and the normative rules that guide action in their social context. A general outline of the technology structuring process is depicted in Figure 1. [2]Users draw on existing institutional rules and resources (e.g., division of labor and work procedures; arrow 1) to use the technological capabilities available to them (arrow 2). In using their technology to accomplish some task, users appropriate their technology (arrow 3) and enact certain social practices, which reinforce, adjust or change the existing institutional realm (e.g., division of labor and work procedures; arrow 4). The influence of individuals' technology use on the institutional properties are often unintended and unnoticed, just as the influence of the institutional properties on technology use is often unnoticed and unacknowledged.

Recently, work using the technology structuring perspective has specifically examined the case of computer-mediated communication technologies (e.g., Contractor and Eisenberg, 1990; Poole and DeSanctis, 1990, 1992; Yates and Orlikowski, 1992). For example, Poole and DeSanctis (1990) have developed adaptive structuration theory to characterize the role of technologies such as group support systems in group functioning. Using this theory, Poole and DeSanctis (1992) conducted detailed micro-level analysis of technology use to determine how groups appropriate decision support technologies in support of their tasks and how such appropriation produces intended and unintended structural and decision outcomes. In another example, Yates and Orlikowski (1992) suggest that users of electronic communication technologies structure the genres or types of communication they use, sometimes drawing on and reinforcing the social rules defining the institutionalized genres (e.g., rules as to the purpose, substance, and form of such communication), and sometimes intentionally or unintentionally violating some of the established genre rules.

The research on the structuring of technologies in use -- whether general or specific to communication media -- tends to focus primarily on the activities of users who shape their technology as they use it in particular contexts. The research we report below identifies another set of structuring activities that, although carried out by users, were not activities of use. Rather, they involved the shaping of other users' activities of use -- a process which we designate metastructuring. This metastructuring perspective allows us to see that such interventions occur frequently, by all of the various intervenors discussed above. We have focused here on a particular type of metastructuring that is carried out by deliberate and organizationally sanctioned technology-use mediators. As we found, this type of metastructuring can be particularly powerful because it significantly structures users' interactions with their communication technology by influencing their communication practices, changing the institutional context of use, and modifying the technology itself. After presenting the findings of our empirical research, we will revisit the structuring process depicted in Figure 1, and extend it to accommodate our insights about metastructuring.

Research Setting and Methods

Research Setting

We studied the use of computer conferencing in an R&D project group within a large Japanese manufacturing firm. The system had been installed to support communication among the members of a newly formed project group developing a computer product, Acorn (a pseudonym). Because Acorn was expected to be an innovative product important to the company's competitive position, a new organizational and technical infrastructure was set up specifically to support the Acorn product development group. The group was composed of about 150 employees, primarily members of the R&D lab supplemented by new employees and external contract programmers as needed. The project group was divided into six teams: an administration team, a hardware development team, and four software development teams. All Acorn project members were experienced computer users, and were provided with powerful networked workstations readily supporting electronic communication.

The history of electronic media in the R&D lab prior to the Acorn project is essential background to the introduction and use of electronic conferencing within the project group. Electronic mail (e-mail) had been introduced into the R&D lab about two years before the Acorn project began, by four young software engineers who were interested in maintaining contact with colleagues and friends outside of the firm. Use of e-mail spread slowly but steadily within the lab, so that about half of the Acorn members were e-mail users by the start of the project. The four software engineers were also interested in the world-wide USENET news-system (a computer conferencing system publicly available and widely used on a number of electronic networks) and particularly in the version adapted for the Japanese language that had been created to link several Japanese universities. Having used this communication technology in college and seeing its potential for internal use, these engineers decided to create a local news-system within their R&D lab a year after introducing electronic mail.

The news-system software is organized hierarchically into topics known as newsgroups, but in setting up this local news-system, the engineers initially created only three parallel newsgroups: general, the software's default category, and miscellaneous and recreational, two of the other standard USENET categories. [3]Users could read and respond to messages posted on the newsgroups as well as post their own messages. The software allowed them to include part or all of a previous message embedded in a new message. Normally, messages accumulated within newsgroups until they were automatically purged after three months. This news-system was initially used only by a very small set of individuals (the engineers and a few friends). As one of these engineers recalled, "At that time we never thought those newsgroups would ever become official communication media in the lab. We were just playing with the software for fun."

A new organizational unit, drawing members from three different labs within the R&D Division, was formed to develop the Acorn product in late September of 1989. The four software engineers were selected to be part of this new project group. Since they all enjoyed using the news-system, they decided to set up a similar news-system specifically for the new project. This news-system was in operation and accessible by all Acorn members for the full 17 months of the project (from the end of September 1989 to February 1991), although it was only an official communication medium for the project from February 1990. Initially, the engineers' news-system activities continued to be unofficial and casual, and usage remained restricted in volume and user base in the early days of the Acorn project. The engineers soon realized, however, that a large project group such as Acorn would require network administration to maintain linkages among the powerful workstations for communication and data transfer. They explained this requirement to project managers and volunteered to play this administrative role.

As a result, the Network Administration Group for Acorn (NAGA), the focus of this study, was created. Its nine members included the original group of four software engineers plus additional members recruited to ensure that NAGA represented each of the six project teams within the Acorn project. Within NAGA, decisions were made primarily by consensus in regular face-to-face meetings (the minutes of which were distributed to all NAGA members via e-mail); e-mail messages were also frequently exchanged among NAGA members to supplement their meetings.

Data Collection and Analysis

As part of a larger study, we examined the policies and process through which the news-system was managed over time by the administrative group, NAGA. [4]Two types of data were collected: extensive textual data supplemented by retrospective interview data. The interview data came from a series of in-depth and unstructured discussions with a key informant, conducted in Japanese, but with later more focused follow-up interviews in English. The interview data revealed important background and contextual information about the firm, the Acorn project, and NAGA, which helped in the interpretation of the textual data. [5]

The textual data consisted of computerized records containing e-mail messages exchanged among the NAGA members and newsgroup messages that were posted on the news-system by Acorn participants (all in Japanese) during the project. We employed qualitative data analysis methods (Eisenhardt, 1989; Miles and Huberman, 1984), classifying the e-mail messages by common topics and actions, and carefully examining the newsgroup messages around the time of NAGA's administrative actions. This allowed us to trace the process by which decisions were made and implemented over time.

During the 12-month period beginning with NAGA's first discussions of the news-system's role and ending after the last major reconfiguration of the news-system (see below), 438 e-mail messages were exchanged by the NAGA members, of which 223 dealt with the news-system. [6]These messages included members' ongoing dialogues about administering the network, along with the minutes of their regular face-to-face meetings. We initially classified the contents of the messages into 97 specific topics about the news-system (e.g., the addition of a particular newsgroup to the system, or the differences between the e-mail and news-system media), often with more than one topic per message. We then categorized the topics into four general subjects: definition of the news-system (NAGA considered the role and purpose of the news-system); promotion of the news-system (NAGA discussed publicizing and educating users about the news-system and about specific newsgroups); revision of newsgroups (NAGA considered amending definitions and usage rules for existing newsgroups); and creation of new newsgroups (NAGA discussed creating new newsgroups suggested by users or felt by NAGA members to be appropriate). Figure 2 shows the distribution of these message types during the period analyzed.

Over the 15 months of the Acorn project for which news-system archives remain, 9302 messages were posted on the news-system. Figure 3 shows the distribution of these messages over time. We examined two types of news-system messages: messages that were posted by NAGA members in the execution of their administrative activities, and messages that were posted by participants around the time of NAGA's actions. Messages posted by NAGA members as project participants were not considered part of NAGA's administrative messages. NAGA members' administrative messages were mostly announcements of news-system changes or solicitations to discuss aspects of the news-system. These messages were useful indications of NAGA's administrative activities and background policies. Messages posted by participants reflected both the influence of NAGA administrative actions and participants' responses to these NAGA actions.

We analyzed all of these sources of data together and iteratively to get as complete a picture as possible of NAGA's activities. The topics of NAGA e-mail messages could be correlated with NAGA messages posted on the relevant newsgroup (e.g., announcing changes in rules or the addition of new newsgroups). Both could be correlated with the dates of first postings on new newsgroups, yielding a timeline of NAGA's changes to the news-system, shown in Figure 4. The key informant interviews were used to elaborate and interpret our findings. Together, these different sources of data allow us to characterize NAGA's activities throughout the project.

Results

The data revealed four types of mediating activities (summarized below and developed in subsequent sections) performed by NAGA members, involving both technical and social elements, in their role as mediators for Acorn's news-system:

Establishment

Establishing its role within the Acorn project was the first essential task faced by NAGA. Once NAGA members realized the scope of the technical and administrative functions they had taken on, they prepared a document for their managers that argued for the importance of having ongoing support of the project's technical infrastructure. In it they proposed that their NAGA activities be recognized as a part of their regular job duties, and that they be allowed to hire two contract programmers to do the necessary programming work. Once managers accepted this rationale, NAGA communicated its role to the rest of the project in announcements made by the six team leaders to make at team meetings, distributed through the project's internal bi-weekly newsletter, and posted in a news-system message which stated, "Our goal is not only to achieve trouble-free use of the network, but also to increase the productivity of the project by improving communication among members." At this point, NAGA was authorized by Acorn managers and recognized by fellow members of the Acorn project group as the committee responsible for administering the Acorn project's network.

Once initial network infrastructure matters had been addressed, NAGA began to deliberate about the role the news-system might play in Acorn. As users of it themselves, they felt that it could aid communication in their large project group and began adding a few new newsgroups to encourage such use (e.g., chat and questions.). They also considered whether the news-system should be used for official communication, such as messages posted on the official bulletin board or announced in the daily lunchtime meetings, or for unofficial communication such as that conveyed via conversations and e-mail. While NAGA members discussed this issue in their meetings, they also solicited the opinions of project members, as evident in the following message which triggered participation and discussion among project members who were already using the news-system. [7]

-----------------------------------------------------------------------------

Subject: New newsgroups

Date: 6 Dec 89 10:05:48 GMT

Newsgroups: misc

We need to discuss the following issues.

(1) The officiality of this news-system is unclear.

Must everyone read the general newsgroup? Will it be just like another bulletin board?

=> Should the news-system be used simply according to a person's taste?

The next topic clarifies the relationship between the news-system and other media such as meetings, bulletin boards, and e-mail.

(2) The relationship of this news-system to bulletin boards organized within each team, and any e-mail mailing lists is unclear.

As long as we use only these closed media, we cannot share useful information among all project members.

=> Should we not gain the advantage of many members working together?

Please continue this discussion by e-mail or the news-system.

-----------------------------------------------------------------------------

This message reveals a great deal about NAGA's consensus-building methods at this stage. It invited participation of other project members in deciding whether the news-system should be official, at the same time that it presented NAGA's view that the news-system would offer "the advantage of many members working together." Several participants posted responses supporting an official role for the news-system

Because most project members were not yet using the news-system, NAGA members also talked individually with many team members and with project managers whose consent was critical if NAGA was to make the news-system an official medium for the project. NAGA's inclusive and encouraging but not directive approach enabled it to build general support among other project members, most of whom gradually adopted the view that the news-system should be an official communication medium for the project group. The fact that they were developing new technology motivated some of the users to incur the costs of learning to use a new technology, as demonstrated by one Acorn member's message: "Because we are developing new technology, we should adopt a positive attitude to introducing this new communication technology."

Having gained considerable agreement that the news-system should be made an official communication medium for the Acorn project, NAGA deliberated internally about how best to position the new medium in relation to the other communication mechanisms commonly used in the firm: daily lunch-time meetings (required by company policy in each division or group, and including official announcements as well as ceremonial elements such as the company song), routing slips (used by administrators to circulate information to all members of a group), bulletin boards (used for posting paper-based announcements), and e-mail (used by an increasing number of project members, for one-to-one communication as well as with mailing lists). After NAGA members evaluated the current uses of these media, they discussed how these should change when the news-system was added to the array of available media. Their decisions were documented in the following meeting minutes:

-----------------------------------------------------------------------------

To: NAGA@xxx

Subject: Meeting report from December 26, 1989

...

Communication media within the division

<principle>

Currently, daily lunch-time meetings, routing-slips, bulletin boards, e-mail, and the news-system are used as communication media. Our purpose is to encourage the use of the news-system.

<daily lunch-time meetings>

The purpose of daily lunch-time meeting should be restricted to confirmation of information that has been announced in the news-system beforehand. This meeting should not be an official announcement tool.

<routing-slips>

The routing-slip will be terminated. Printed information should be posted on bulletin boards and we should direct project members to use the news-system as much as possible.

<bulletin boards>

Bulletin boards can be used only for printed information. However, all information should be provided through the news-system and at least summaries have to be posted on the news-system.

<e-mail>

Precise e-mail use guidelines should be set. E-mail must be used only for urgent or confidential information.

The definition of 'urgent' information: information that should be shared within a day.

-----------------------------------------------------------------------------

As the minutes make clear, NAGA's explicit objective was to encourage news-system use by allocating most of the information exchanged within the Acorn project, including all of its official information, to the news-system. They also persuaded the Acorn team leaders to require their team members to use the news-system, essential support in making the news-system an official communication medium. NAGA also believed that making the new medium official would legitimate its use and, as our informant noted, "This would increase usage as people wouldn't feel guilty using it." During the daily lunch meeting on January 30, 1990, one of the project managers announced the new policy requiring project members to use the news-system and, in particular, to access two newsgroups -- general and announce -- each day. [8]

This authorization of the conferencing system as the official communication medium of the project was an important departure from traditional norms and procedures in the firm and in the R&D lab. NAGA's efforts created a new set of institutional guidelines around what is considered to be official and legitimate communication in the organization. It diminished the role and centrality of the daily lunch-time meetings that had previously constituted the official communication medium, and it made the entire project dependent on the computer conferencing system, requiring all members to become somewhat proficient with this new technology so as to access it daily.

On the day that the manager's announcement made the news-system official, NAGA posted on it a usage guide to differentiate and establish guidelines for using e-mail and the news-system:

-----------------------------------------------------------------------------

Newsgroups: announce

Subject: Guideline for the usage of mail & news

Date: 30 Jan 90 11:03:28 GMT

...

(1) Use e-mail and the news-system effectively!

If you want to send information to some specific person, e-mail may be useful. However, when you send it to a group of people or to all members in the division, please use the news-system as much as possible.

If you use e-mail all the time, we will receive a huge number of messages and have to read all of them. Remember the difference between a traditional bulletin board posting and a letter. Think again when you send an e-mail message to any mailing-list (like all@xxx) You may be able to provide useful information to other project members by using the news-system.

The news-system will function as an official tool starting in February 1990.

(2) Obligation to access both e-mail and the news-system

NAGA has established a facility allowing all members to use both Mail and News

We will support it to make sure that everyone is comfortable using these systems.

All members must access Mail/News:

e-mail: Twice a day both in the morning and afternoon

News: At least once in the morning to specific newsgroups--

the general and announce newsgroups

*If you want us to send messages on the general and announce newsgroups by e-mail, NAGA will provide a service to send messages by e-mail automatically.

*You also can post a message by using the e-mail function.

-----------------------------------------------------------------------------

These guidelines promoted usage by assigning to the news-system specific functions previously handled by other media and communicating very concretely its vision of how the news-system could be used effectively as part of the Acorn project's internal communication. Thus NAGA attempted to remove any uncertainty that might restrict usage.

At the same time, NAGA took other actions designed to increase usage. First, it created local newsgroups for each of the six teams within the project, to provide an alternative to e-mail distribution lists. These newsgroups, according to our informant, were also intended to reduce people's apprehension about writing a message to the entire project organization. This tactic seems to have had an effect, since almost half of all messages on the news-system in the early weeks of its official use were on the local newsgroups; later, that proportion would drop to about one third as more discussion shifted to the non-local newsgroups. NAGA did not, however, restrict access to the local newsgroups because they believed that sharing information across as well as within local teams would ultimately benefit the project as a whole. As they explained in a message posted to announce and define these local newsgroups, "...it is fine if members of other teams also access the local newsgroups. In fact, these newsgroups can even be used as places for public debate." During the entire project, about 33% of the messages in the local newsgroups were written by members of other teams. Cross-fertilization between the groups occurred via this mechanism, as when someone posted a message on another team's local newsgroup saying, "I am an outsider, but I would like to explain my idea about [a certain topic]."

NAGA also manipulated the news-system software to make it easier to use. For example, NAGA used a technical fix to deal with resistance anticipated from some project members who were less comfortable with the new conferencing medium (including some of the project managers). As indicated at the end of the usage guide message shown above, NAGA established a facility allowing members to send and receive news-system items by e-mail, a medium more widely accepted than the news-system in the Acorn project at this time, though still not universally used. This technical solution to a social problem allowed NAGA members to avoid possible challenges to their objective of establishing the news-system as the primary and official communication medium within the project.

As shown in Figure 3, NAGA's activities in establishing the news-system generated a dramatic increase in the number of messages posted on the news-system. Before the announcement of the new official role for the news-system, about 20 messages a week were posted on the news-system; after that, the number rose to around 200 messages per week (excluding holiday periods) until the end of the project was announced in February of 1991. Thus NAGA's campaign to establish the news-system as an official Acorn communication medium and to get project members to use it was clearly successful. Indeed, by the end of the project, approximately 95% of the members had posted messages, and while a few individuals did not post messages on the system, they read items posted on the mandatory newsgroups via their e-mail.

Reinforcement

NAGA members were not content simply to establish the news-system in the Acorn project; they were conscious of an ongoing need to promote usage so that the news-system would be sustained as an essential part of the project's communication system. In addition to maintaining the technical operation of the system, NAGA members also wanted to reinforce their vision of appropriate usage of the communication medium.

E-mail messages among NAGA members demonstrated a low but fairly consistent level of discussion concerning user education to reinforce and promote use of the news-system during the eight months following its establishment (see Figure 2). An item in the minutes of one NAGA meeting, for example, stated, "We will hold a lecture to educate new recruits about the usage of the news-system. We will use the company-wide network usage guide as a text book and especially explain details about unique issues originating in our news-system." NAGA members also posted messages (as users) on new newsgroups to encourage and guide others' use. Such lectures and posted messages allowed NAGA to promote as well as to shape use of the news-system. In addition to reinforcing system use, NAGA maintained the news-system and network technically. They established new user accounts when newcomers joined the project, backed up the system regularly, managed disk space, and performed or oversaw other aspects of technical maintenance necessary to keep the system operational.

After a high level of usage had been sustained for several months, NAGA shifted its attention from supporting use of the medium to promoting what it saw as its effective and appropriate use. As both users and system administrators, they could easily monitor use. When a project member posted a message on the wrong newsgroup, NAGA first responded with a private explanation. If several participants made the same mistake, NAGA would post a general reminder explanation on the news-system. For example, NAGA felt that the mandatory newsgroups should be kept manageable in size. The general newsgroup was initially defined as containing "important announcements to all members," and the announce newsgroup as containing "information for all project members such as meeting and event schedules." With such broad definitions, many participants were confused about which of these newsgroups, if either, was appropriate for a message. Unimportant messages and discussions sometimes appeared on these newsgroups, especially in announce. Thus NAGA kept trying to clarify--both generally and with regard to specific messages--what belonged in the general and announce newsgroups and what should be posted in the discretionary newsgroups. Eventually, reinforcement was deemed insufficient, and NAGA attempted adjustments in the definitions and usage rules (discussed below).

Given high usage levels, NAGA was also concerned that people might waste too much time reading and posting messages. As a result, it posted the following general guidelines:

-----------------------------------------------------------------------------

Newsgroups: general

Subject: About messages

Date: 5 Oct 90 10:42:26 GMT

Cautions about news-system usage

(1) Long embedded messages

There are still many messages that include long embedded parts.

As readers can refer to the original message through the automatically embedded 'message-ID', please shorten the embedded part as much as possible.

(2) Choose the proper newsgroup

What do you do if you cannot find a proper newsgroup for your message? There have been many messages that noted, "This does not fit to this newsgroup..."

Please ask NAGA members if you have any question or request. We will respond to all participants' feedback.

Move to a proper newsgroup when discussion content changes by using the follow-up-to command.

(3) Cross-post

Please use the cross-post command when you send the same message to several newsgroups. If you use this command, readers need not read the same messages several times.

NAGA

-----------------------------------------------------------------------------

This message both reminded users of communication rules and suggested use of resources to conform to NAGA's envisioned usage of the medium. Such ongoing educational efforts reinforced the type of usage institutionalized in the news-system structure and in NAGA's rules for its use. NAGA, however, passed up at least one opportunity to ensure that users followed one of its guidelines. Responding to managers' concerns about news-system misuse, NAGA had issued a policy that the rec newsgroup was to be used only outside of normal work hours. When NAGA members observed messages posted on this newsgroup during normal work hours, they sent e-mail reminders of this policy to the transgressors. Although NAGA was not able to eliminate this practice, it chose not to institute a technical fix that would have prevented any violation of the policy, but focused primarily on attempting to change behavior through education.

In their ongoing monitoring of the level and nature of news-system usage, NAGA members paid close attention to direct and indirect user feedback, attempting to identify and remedy errors, confusion, or inappropriate usage, and to reinforce appropriate usage. They also maintained technical aspects of the system to avoid deterioration of usage caused by system problems. In these reinforcement activities, NAGA members ensured that the communication technology they had established continued to be used at a fairly high level and usually in ways that fit their vision of the new medium's role in the project.

Adjustment

NAGA's objective was not, however, simply reinforcement of the new medium's structure and rules; it also sought to adjust and enhance the news-system and its use--both technically and socially--in response to user feedback. In doing so, it modified newsgroup definitions and usage rules, changed the system's software, and created new newsgroups while maintaining its overall vision of the system's role in supporting communication within the project.

Revisions to newsgroup definitions and rules were made in response to direct user feedback in the form of questions and statements of problems and indirect user feedback from visible errors in use, as well as statements that suggested confusion about the use of certain newsgroups. Such feedback triggered revisions to the definitions and rules for various newsgroups, such as the general and announce newsgroups discussed above. When reinforcement did not elicit what NAGA considered appropriate use of these mandatory newsgroups, NAGA members had a series of discussions over several months about how better to define these two newsgroups. They experimented with a number of revisions to the definition and usage rules, but the problem of how to make best use of the mandatory newsgroups was not solved through adjustments; ultimately it was resolved during one of the change episodes discussed below.

NAGA also enhanced the software to make the news-system easier to use. The news-system as imported into the project allowed Japanese characters in the body of the messages but required subject lines to be written in English or in Romanized Japanese (Japanese characters represented phonetically in Roman characters). Although such representation was not too difficult for technical language, which was often based on English, it posed problems for administrative messages posted by secretaries and members of the administrative group. Soon after the news-system was made mandatory, NAGA modified the news-system software to allow Japanese subject lines.

Another form of system enhancement was the addition of new newsgroups, made in response to specific requests or on the basis of NAGA observations. In general, NAGA preferred to cluster the addition of newsgroups and changes in news-system configuration together, as discussed in the next section. However, NAGA also wanted to be responsive to users' requests in its ongoing adjustment of the system, so it added a few newsgroups at other times. Such technical adjustments included subdividing the computer newsgroup into various hardware and software categories and creating newsgroups such as headlines, union, and release. For example, users requested a newsgroup for announcing and discussing company union matters, and union was created in response. In another instance, a participant posted the following message on announce:

-----------------------------------------------------------------------------

Newsgroups: announce

Subject: Subscription to magazine xxx

Date: 13 Feb 90 02:40:16 GMT

Distribution: Acorn

Subscriptions of two magazines were started last month.

They are in xxx bookcase. Please read them at your convenience.

I would like to send reference messages about new books each month

because building B is separated from the main building, so this will

be useful for people who are working in building B.

May I post these messages to this newsgroup? (the announce newsgroup)

Or is there any other more proper newsgroup?

-----------------------------------------------------------------------------

This message initiated a dialogue among NAGA members. After discussion with other project members, they formed the headlines newsgroup for the announcement of newly arrived books and magazines.

Thus in its role as ongoing administrative group in charge of the Acorn news-system, NAGA attempted to respond to both direct and indirect feedback and suggestions from other project members, while keeping in mind its own objectives for the medium. In doing so, it modified and enhanced the news-system socially and technically by modifying definitions and usage rules and by adding new newsgroups. NAGA thus shaped both users' understandings of the technology and the technology itself to keep news-system usage more effective and relevant. Some such refinements were reactive, executed in response to users' requests or on the evidence of user confusion, while some were proactive, executed opportunistically by NAGA when they saw a way to improve the system. Both kinds of adjustments, however, were designed to improve and thus reinforce the existing conditions of use, and were not intended to create radical departures from the status quo.

Episodic Change

The adjustments attempted to solve problems as they arose, but sometimes NAGA either failed to solve the problem with such actions (as with the confusion about the announce newsgroup) or postponed a full solution. Periodically, NAGA addressed clusters of such problems and the larger issue of system configuration. During the news-system's official use in the Acorn project, NAGA conducted two episodes of change involving significant additions to the news-system functionality and reconfigurations of the structure of newsgroups within it.

The first change episode, in April 1990, introduced three new newsgroups (see Figure 4). Although one of these, Acorn-troubles (reporting certain types of problems with the Acorn product), was similar to existing newsgroups, they other two mail-lists and guide, expanded the functionality of the news-system as well as NAGA's vision of the medium in new directions. NAGA had originally envisioned the medium as a dynamic mechanism for internal communicating across the six teams making up the Acorn project. The two new sets of newsgroups significantly expanded this vision by providing access to external information and long-term archiving of certain types of reference information.

The mail-lists newsgroups were designed to allow Acorn members to access up-to date technical information available by contract from an outside public network organization, techinfo (a pseudonym). By the terms of the contract, however, this external information had to be kept confidential to the Acorn project group itself. Therefore, NAGA created a different social and technical structure for this set of newsgroups. It named a moderator to monitor and answer questions about these newsgroups. Moreover, it prohibited participants from posting messages externally, and introduced the mail-lists/techinfo/discuss newsgroup for internal discussions about related topics. NAGA ensured compliance by providing an additional technical feature for automatically transferring to this internal newsgroup any follow-up messages users inadvertently tried to send to the external mail lists. NAGA announced these usage rules on the news-system:

The other major modification to the news-system in this change episode was the addition of the guide newsgroups as archival, rather than ephemeral, sources of shared information. The news-system automatically deleted a message from a newsgroup after it had been posted for three months. But by April of 1990 NAGA had seen the need for convenient storage and retrieval of reference information with a longer life span -- an organizational memory for the project. Thus by modifying the automatic deletion feature they created a number of moderated guide newsgroups to function as data-bases of reference documents (e.g., addresses lists, administrative procedures, maps, etc.). Whenever participants posted messages on these newsgroups, these messages were automatically sent to the moderators, who checked whether or not each message was appropriate for that newsgroup. If it was, the moderator posted the message on the newsgroup. If not, the moderator sent an e-mail message to the participant explaining the reasons why it was not appropriate. NAGA also created a specific format and usage rules for all guide newsgroups, explained in a guideline document which each moderator posted on each guide newsgroup.

In the second change episode, NAGA attempted to rationalize and restructure the news-system and to solve some long-standing problems with mandatory newsgroups. NAGA began planning this episode in July of 1990. In anticipation of it, NAGA members posted a message soliciting from participants requests for new newsgroups and ideas about the structure of the news-system. NAGA announced the schedule and details of the change in a very long message posted to the announce newsgroup on October 17 of 1990, the beginning of which is reproduced below:

-----------------------------------------------------------------------------

Newsgroups: announce

Subject: A modification of the news-system

Date: 17 Oct 90 00:49:42 GMT

Information on a modification of the news-system

*******************************************************************

*******************************************************************

We introduced the news-system one year ago and now it has become an indispensable medium for our division.

As the number of messages has increased, several problems have come up. For example, some newsgroups are used improperly, some newsgroups are not used at all.

Therefore, NAGA has changed the configuration of the newsgroup structure.

This message provides information on the new news-system.

...

-----------------------------------------------------------------------------

One key set of changes in this episode was the creation of the official newsgroups to solve the ongoing problems with the mandatory general and announce newsgroups. These latter groups were eliminated and replaced by two moderated newsgroups: official/secretary, for "important announcements of meetings and events," and official/general for "other important announcements." While the definitions were specified slightly more carefully than those for general and announce, NAGA's real solution for restraining the amount and nature of material posted to the mandatory newsgroups was to put a moderator in charge of each of them to screen messages before posting them.

Two other major sets of changes involved clustering newsgroups, in one case by priority and in one case by subject. First, several individual newsgroups were organized into a second tier of info newsgroups (e.g., info/look-for, for lost and found announcements; and info/release, for information related to software releases) that were recommended but not required daily reading. Second, to help project members find information about Acorn previously scattered among several newsgroups, NAGA subdivided the Acorn newsgroup, creating new categories and clustering existing Acorn-related newsgroups under it. This second modification thus tried to reconfigure the news-system to improve its clarity, coherence, efficiency, and usefulness as a communication medium supporting the Acorn project.

At the end of the long message announcing the second set of changes, NAGA stated its philosophy about shaping the news-system: "NAGA thinks that the news-system schema should be changed depending on the situation. We would like to manage flexibly. This modification is one of those actions." As this statement makes clear, by this time NAGA members viewed the structure of the news-system not as fixed and permanent but as context-specific. As the situation and the project's needs changed, and as NAGA refined its own concept of what was valuable to the project, it changed the system structure and its rules. Yet NAGA tried to balance the evolving communication requirements of the project organization with a certain amount of stability in the communication medium by undertaking substantial changes only episodically. This approach allowed NAGA to contemplate the changes over a period of time, assessing how they would reshape their vision of the news-system, then to institute them together for maximum visibility and impact. Moreover, NAGA took steps to ease the transition by retaining all of the terminated or restructured newsgroups in read-only form, posting a final message on each telling users where messages should now be posted.

Technology-Use Mediation and the Metastructuring Process

The research study of NAGA described above has provided insights into the actions of organizational actors who significantly and deliberately shaped users' initial and ongoing use of their new communication technology. The above description provides a detailed account of the objectives, activities, and influence of NAGA's role. While intervention in users' technological use happens in many ways -- both intended and unintended -- we have focused here on a kind of intervention (technology-use mediation) that is deliberate, ongoing, and organizationally sanctioned. Based on our empirical research we argue that such intervention may serve as a powerful mechanism for contextualizing technologies in use, an issue of increasing importance as more general purpose communication technologies are used to support dynamic organizational forms and practices. By foregrounding and framing this mediation role, we may gain insights into how to organize it so as to support the ongoing usefulness of electronic media in conditions of fluidity and change.

In order to provide a theoretical interpretation of the technology-use mediation role, we draw on what we already know about how users structure their technologies in use (Barley, 1986; DeSanctis and Poole, 1994; Orlikowski, 1992b). As we saw earlier, analyses of technology structuring have tended to focus on users as the key actors who shape and are shaped by their technological and institutional contexts during use. Yet our study shows that in some situations, others -- those we have referred to as mediators -- routinely and deliberately intervene in users' structuring activities by influencing users' understandings, altering the technologies to ease use, modifying institutional rules to promote particular kinds of communication, and facilitating access to and operation of the technologies. Such intervening actions are not captured by our current understanding of technology use as a process of structuring (see Figure 1), and an extension of this understanding is thus required.

Figure 5 depicts the structuring activities of mediators as they intervene in users' structuring of their technology. To allow for the possibility that mediators may also be users (as in the case of NAGA) we have distinguished between the role of use and the role of mediation. Like users, mediators are influenced by the institutional context within which they work, and which furnishes them with certain resources, authority, job responsibilities, and knowledge (arrow 5). Technology-use mediation is also constrained (and enabled) by the existing technological configuration within the organization (arrow 6). In taking action, mediators create policies, procedures, guidelines, templates, access mechanisms, applications, and physical configurations. These change the institutional properties of the organization (arrow 7), and they directly affect the technology itself (arrow 8). In addition, technology-use mediation influences users' interpretations (and hence their actions) by providing them with understandings, images, concepts, knowledge, and heuristics about their specific technology (arrow 9). In doing so, mediators influence users' work habits around use of the technology, their technological frames (Orlikowski and Gash, 1993), and their perceptions of their technology (Fulk, Schmitz, and Steinfield, 1990). [9]Such action by the mediators, in turn, is influenced by users' existing assumptions, expectations, and knowledge, their level of experience, and their current requirements (arrow 10).

Technology-use mediation can thus be seen to structure the institutional, interpretive, and technological rules and resources that users draw on when they use their technology (arrows 1 and 2). In their appropriation by users, these images, concepts, policies, templates, configurations, etc., are accepted, legitimized, and reinforced (arrows 3 and 4), becoming institutionalized over time. Management's official and/or implicit sanctioning of such rules and resources further serves to authorize them as appropriate ways for users to interact with their technology.

In Figure 5, we have depicted technology-use mediation as a process of structuring resembling that engaged in by users when they structure their technologies in use. Yet, the structuring involved in technology-use mediation shapes users' own structuring of their technologies. Hence, it is a form of second-order structuring, which we refer to as the metastructuring of technologies in use. The right-hand side of Figure 5 depicts this process of metastructuring engaged in by mediators such as NAGA. More generally, this process can also characterize the less deliberate interventions of individuals taking on roles such as champions, trainers, and local experts.

By engaging in a deliberate and sanctioned process of metastructuring, mediators are involved in shaping and reinforcing particular forms of meaning, power, and norms. By developing and defining categories of use, disseminating and enforcing rules for use, and allocating resources to certain kinds of contextualizing activities, mediators exert considerable influence over how a particular communication technology will be established and used in an organization, and thus over communication norms associated with it. Whether such influence will be experienced as coercive and controlling or helpful and supportive undoubtedly depends on various contextual factors, including existing norms around help-seeking and help-giving behavior, and how the mediators conduct themselves vis-a-vis their users -- as experts dictating particular usage or as facilitators working with users to enable a range of ongoing uses. Further research is needed to determine under which conditions such influence will be experienced as constraining and under which as enabling.

We can now use the findings of the NAGA study to elaborate this metastructuring framework. In particular, we have suggested that technology-use mediation involves at least four types of activities with which mediators contextualize technologies in use: establishment, reinforcement, adjustment, and episodic change over time. This categorization of mediating activities may also be seen as a preliminary classification of metastructuring moves, although metastructuring can take less comprehensive and perhaps quite different forms, as well. However, we believe that this categorization serves as a useful starting point to begin thinking about the technology-use mediation role and the process of metastructuring more systematically and generally, as we have attempted to do below. Table 1 summarizes these mediating activities, as well as providing illustrative examples from the NAGA story recounted above. Figure 6 shows how these four mediating activities might be sequenced in an ongoing process of metastructuring.

Establishment

During establishment of a communication technology, mediators set up the technology itself (its physical parameters, features, etc.), they articulate the ways in which the technology will be initially adopted by the users in their communication, and they propose modifications to the institutional context and the technology that will facilitate assimilation of the technology into standard operating procedures. Some work has suggested that the initial understandings, routines, and procedures established around a new technology congeal fairly rapidly after its implementation (Tyre and Orlikowski, 1994), as users reach a plateau of competence (Bullen and Bennett, 1990) and are impatient to get on with using the technology to accomplish productive work. Thus, establishing the technology physically in the organization, and establishing it socially in the communicative habits of its users is a particularly influential activity for mediators. NAGA apparently recognized this, for its members--informed by their vision of how the news-system could serve as a communication system for the whole project--spent almost two months discussing ideas, soliciting feedback, persuading management to proclaim the news-system the official and required medium of the project, and facilitating a comfortable transition for the project members to this new medium.

Establishment has the effect of creating a particular set of interpretive, institutional, and technological conditions of use. Depending on whether the intention behind the implementation of the new communication technology is to improve the existing way of doing business or to change it fundamentally, the establishment process will reinforce, adjust, or replace ways of working and communicating. In NAGA's case, both minor improvements and major changes were apparent with the establishment of the news-system. For example, NAGA's replacement of e-mail lists with local newsgroups represented a minor change in their communication within project teams. On the other hand, when NAGA persuaded project managers to designate the news-system as the official medium for the project announcements -- in place of the long-standing and traditional daily lunch-time meeting -- they accomplished a major shift in the institutional rules governing the meaning, privilege, and legitimacy of particular communicative practices within the project.

Reinforcement and Adjustment

Once a technology has been established both technically and socially, mediators shift their attention and actions to offering ongoing assistance and consultation, as well as various levels of encouragement and support. Shortly after establishment, mediators help users to incorporate the new technology into their work, providing substantial amounts of advice, hands-on demonstration, and guidance. In addition, mediators can correct inappropriate or inefficient usage as they encounter it, either through their interaction with users or through formal or informal monitoring of technology use. Appropriate use is then promoted through various training and communication sessions. NAGA members had a certain vision of the role of the news-system in the project, and they advocated this through newsletters, news-system messages, lectures, and meeting announcements. In addition, mediators maintain the technology itself to reinforce use.

In addition to reinforcing the use of an established technology, mediators also extend or refine particular uses of the technology. By adjusting the technology itself and/or its surrounding social norms, mediators may accomplish incremental changes to a technology's use. Such activities include changing usage rules to address problems that have emerged, or making technical extensions or alterations (e.g., allowing Japanese subject lines) that modify the nature of the technology or its use, but do not change it fundamentally. As we saw in the Acorn project, mediators may make such adjustments in response to a number of triggers, including user requests, detection of a problem with the existing pattern of use, or a perceived opportunity to enhance use. Such adjustments result in minor modifications to the technology, its use, and users' understanding of the technology. Ultimately, however, they have the effect of reproducing the institutional and technological conditions of use, not of challenging them. While the refinements may improve and enhance the technology and its use, the net effect is to increase the effectiveness of the existing system, not to transform it.

Episodic Change

The fourth type of activity performed by mediators -- episodic change -- involves a significant reassessment and restructuring of the technology and its use. Unlike adjustments, episodic changes are attempts to create major improvements in the coherence and performance of a communication technology, its use, users' understandings, and the institutional context of use. Because change episodes have greater potential to disrupt the established norms and ongoing flow of communication in the group, they tend to be scheduled periodically and require extensive justification to counter user resistance. In our study, such episodes were proactive, initiated by the mediators themselves, although they may take into account user requests that have been accumulating for some time. With episodic change, there is an opportunity to step out of the normal routine of daily use and examine the bigger picture of the whole technology and its relationship to the context of use. For example, NAGA did this when they reconfigured the news-system to establish a more logical relationship among the newsgroups. We can also imagine such episodic change resulting from external shocks (Mackay 1990).

Episodic change resembles Poole and DeSanctis' (1992) notion of "junctural structuring events" which they describe as "episodes during which major determinations are made about which structural features of the [technology] to appropriate, how they will be appropriated, and whether and how they will be reproduced" (p. 28). In their case, the junctural structuring events are precipitated by the primary users of the communication technology (a group decision support system). In our case, however, it is the mediators, not the primary users, who initiate the structuring event, and their action appears much more deliberative and reflective than that of the primary users whose action is often (but not always) unplanned and emergent. Like junctural structuring events, episodic change events create the possibility for institutional transformation. While NAGA's change episodes did not appear to result in significant institutional changes, the potential for such changes was always present in its activities.

Conclusions

In this paper we have identified another level of technology structuring -- metastructuring-- that influences the structuring activities of communication technology users within organizations. We have outlined the form and general character of the metastructuring process, and suggested that such a process occurs with or without conscious planning. Our interest, however, has been to articulate a particular kind of metastructuring -- technology-use mediation -- represented by deliberate, ongoing, and organizationally sanctioned intervention in the use of a communication technology. We have articulated and categorized the various actions of these mediators over time and shown how they shaped and influenced the use of a new electronic medium in one setting. We now discuss the significance of the mediator role in creating effective electronic support for communication in dynamic organizations by analyzing the events related to the attempted transfer of the news-system after the Acorn project ended. After analyzing this episode, we will explore some of the implications of our study for research and practice.

Context Specificity and the Transfer of the News-System

When the late February 1991 termination of the Acorn project was announced in January, system usage began to decline precipitously (see Figure 3).[10] At the same time NAGA began to make plans to transfer the news-system into one reorganized lab where many of the Acorn project members, including the founding members of NAGA, had come from and would return. (This lab still had its old news-system created before the Acorn project began, but usage was quite low.) Based on our informant interviews, we can trace the transfer attempt and examine the reasons for its relative lack of success.

NAGA members decided to introduce the structure of the successful Acorn news-system into the lab's news-system with only two variations: changing the local newsgroups from reflecting the six Acorn teams to matching separate projects in the lab, and deleting the set of newsgroups related specifically to the Acorn product. NAGA assumed that the Acorn news-system schema would fit this lab with only these minor adjustments because two-thirds of its members were former members of the Acorn project. While NAGA members obtained managers' consent for this action and announced definitions and usage rules for the new local newsgroups before introducing them, they did not obtain consent to continue operating as an official newsgroup administration group in the lab. Each NAGA member belonged to one of the different product development groups, and news-system administration for the lab as a whole was not considered by their various project managers to constitute part of their jobs. [11]Under these circumstances, NAGA's further activities were discouraged and hence were relatively ineffective. Their establishment activities were limited in comparison to those performed in the Acorn project; moreover, they could not pursue the news-system's ongoing reinforcement and adjustment, nor could they initiate change episodes.

Without ongoing contextualization, usage gradually declined on the recently transferred news-system. Only the new local newsgroups were popular. Some of the reasons for the decline were rooted in the lab's structure and its differences from the structure of the Acorn project group. Because administrators in the lab were not familiar with the news-system and because they used a different type of computer that was not designed to support the news-system, none of them posted important official or administrative messages on the news-system. In the Acorn project, the administrative group was part of the entire project group and administrators used or had access to the same workstations as other project members. If NAGA had retained its status and resources, its members might have been able to solve this problem as they had solved the problem of reluctant users by setting up forwarding of the mandatory newsgroups in e-mail. If that had been accomplished as part of system establishment, and if the administrators had been convinced to participate, the news-system could have retained an official role in administrative matters within the lab. Instead, the news-system lost its officiality and usage dropped.

There was a related reason for the decline in usage. The members of this lab were developing several different products, in contrast to the Acorn project members, who were developing a single product. Thus information sharing across groups was less useful here than in the Acorn project. The new local newsgroups NAGA created for these product development groups continued to be used for sharing information within project groups. But in this case, rather than serving as a migration path towards other, lab-wide newsgroups that allowed wider sharing of information, the local newsgroups simply replaced e-mail lists as the locus for within-project communication. With the exclusion of the lab's administrators from the system (and thus of lab-wide administrative matters), the lab as a whole lacked the motivation to share information. Thus non-local newsgroups received very little usage.

Although NAGA succeeded in establishing an effective news-system for the Acorn project (as measured by both volume of messages posted and percentage of project members using the system), their attempt to transfer this system to a different organizational unit seems to have been relatively unsuccessful. This outcome underlines the context-specificity of NAGA's activities. In interaction with users, NAGA had contextualized the communication technology within a given context -- the Acorn project. The resulting system, while highly attuned to the Acorn context, did not match the different context of the new lab even with minor adjustments. Because organizational and technical factors precluded technology-use mediation in the lab context, the communication technology supported only limited use within a new context. The mediator role enabled NAGA to establish and support the communication technology in one context and organizational form (the Acorn project), but lack of institutional support constrained its functioning as that organizational form dissolved and the context of use shifted. This episode highlights the importance of contextualizing activities in facilitating effective communication in dynamic organizations.

Implications for Research

The identification and articulation of the metastructuring process and of technology-use mediation have a number of implications for research and practice. On the research side, the notion of metastructuring draws attention to the fact that there may be multiple levels of action and interaction in organizations, and that a process of technology structuring at one level may itself be structured at another level. This suggests that our understanding of the use of new computer-mediated communication technologies should consider the role of these additional levels of structuring, for the intervention of others in technology use exerts a significant influence on the nature and effectiveness of organizational communication via new electronic media. Understanding the effect of different types of metastructuring, from the deliberate and sanctioned technology-use mediation to the relatively inadvertent metastructuring of expert users, are important goals for future research.

With respect to technology-use mediation, further research is clearly needed to examine the various aspects of technology-use mediation and their consequences in different organizational contexts and with different communication technologies. For example, NAGA's problematic attempt to transfer the Acorn technological configuration into the firm's lab suggests that the more extensively a technology has been involved in technology-use mediation, the more context-specific and useful it is in that context, but the more difficult it will be to transfer the technology directly or with only minor adjustments to other contexts of use without a similar contextualizing process. Empirical studies should examine the importing and exporting of technologies into and out of different organizational settings to assess the validity of this proposition.

Technology-use mediation probably varies with the type of technology and mediators involved. Communication technologies that are more specialized in their purpose (e.g., group decision support systems) may need less contextualizing than more general purpose technologies. Although in the case studied here the NAGA members also happened to be users, we do not believe that this is a requirement and suggest that the technology-use mediation role can be played by a range of different organizational actors. Further studies could illuminate how effectively different actors fill this role, based on how close they are to the context of use, how well they understand users' communication practices and interaction norms, how much credibility they have with the users, and how knowledgeable they are about users' technical abilities and the communication technologies at their disposal.

Our outline of technology-use mediation is grounded in a single case study in a Japanese R&D lab. In particular, the Japanese cultural setting of this study undoubtedly influenced the use of this conferencing technology. We cannot assume that the Acorn project and its organizational context are typically Japanese. Straub (1994) believes that Japan is at least five years behind the U.S. in use of information technologies, including electronic communication technologies such as e-mail. He claims that very few Japanese firms use e-mail, and presumably still fewer have implemented conferencing systems such as the Acorn project's news-system. Because we studied an R&D project developing an information technology product, this case may not be typical of other Japanese enterprises.

While this study cannot be said to represent Japanese use of electronic media, the Japanese setting undoubtedly shaped aspects of NAGA's mediation role. For example, the careful consensus-building efforts initiated by NAGA during establishment of the communication technology seem to reflect the process of nemawashi, described by Ulfhielm (1987: 128) as "a tactful (and indirect) sounding out of each group member ... about his or her thoughts and feelings concerning a certain issue; it also functions as an instrument for subtle persuasion." The process followed by NAGA thus seems to reflect an existing, culturally embedded group process, suggesting one reason that project members responded so rapidly (as reflected in Figure 3) to the pronouncement by management and NAGA that the new electronic medium was now to be considered official and mandatory. A similar process might be less effective in a comparable American R&D group. Although the details of the mediation role probably differ in different cultures -- both national and corporate -- some technology-use mediation probably has value in most contexts where users are attempting to appropriate sophisticated communication technologies into their work practices and interaction patterns. Future studies should examine technology-use mediation and metastructuring in general under different cultural, organizational, technological, and mediation conditions, and validate or elaborate the process as well as the institutional and interpretive influences articulated here.

Implications for Practice

This paper also has implications for organizations implementing networks and computer-mediated communication technologies. If such electronic media are to be viewed as the backbone of new and fluid organizational forms, the ongoing contextualization performed by mediators adds value by keeping technology usage aligned with user conditions and organizational circumstances. In the particular case study related here, the intervention was initially undertaken casually and without explicit authorization. As NAGA's activities became more deliberate and as its members sought and obtained organizational authorization and resources, thus taking on the role of technology-use mediators, the effects of their activities became more significant and apparently more effective. The transfer episode discussed above highlights the importance of adequate institutional support -- social, monetary, and technical -- for the mediation role to make it as effective as possible. When the mediators lost that support, they could not effectively maintain their role. We argue that the effectiveness of technology-use mediation in practice depends considerably on the extent to which such activities are officially recognized, sanctioned, supported, and rewarded by the organization.

Our study of NAGA's activities suggests that organizations may wish to be more deliberate about the process of metastructuring and institute an explicit role of technology-use mediation. Metastructuring occurs, with or without careful deliberation and management. Trainers, local experts, and system administrators all intervene in some way in users' interactions with their communication technologies, and hence engage in a type of metastructuring. With reflection and effective distribution of resources, organizations might use this process to advance particular kinds of communication, both initially and over time. For example, organizations wishing to promote innovative use of new electronic media could empower a group of mediators to provide images, procedures, and use guidelines different from those comprising the technology frames, habits, and genres of communication already established in the organization. Organizations could also use technology-use mediation to occasion and support episodes of experimentation, reflection, and change in technologies and their use, so as to allow for the evolution of technological frames, work habits, and communication routines in conditions of change. Indeed, research by Tyre and Orlikowski (1993) suggests that the effective adaptation of process technologies to their contexts of use requires conscious management of opportunities and resources for improving technology use.

Computer-mediated communication technologies are, by and large, general purpose tools that help individuals communicate, share information, and make decisions in a broad range of settings. Because these settings are specific and the tools are usually general, customization of the technologies and their use is typically needed to make them relevant to contexts of use. In the new and fluid organizational forms now emerging in the face of rapidly changing environments, contextualization of technologies will be a critical mechanism for helping communication norms and work practices adapt. A process of metastructuring shaped by organizationally sanctioned technology-use mediators may thus be a significant element in organizational adaptation to change.

References

Barley, S.R. "Technology as an Occasion for Structuring: Evidence from Observation of CT Scanners and the Social Order of Radiology Departments," Administrative Science Quarterly, 31, 1986: 78-108.

Beath, C.M. "Supporting the Information Technology Champion," MIS Quarterly, September 1991: 355-371.

Bjorn-Andersen, N., Eason, K. and Robey, D. Managing Computer Impact, Norwood, NJ: Ablex Publishing: 1986.

Bullen, Christine V. and Bennett, John L. "Groupware in Practice: An Interpretation of Work Experience," Proceedings of the Conference on Computer Supported Cooperative Work, Los Angeles CA; October, 1990.

Contractor, N.S. & Eisenberg, E.M. "Communication Networks and New Media in Organizations." In Fulk, J. & Steinfield, C.W. (Eds.), Organizations and Communication Technology: 143-172. Newbury Park CA: Sage, 1990.

Culnan, Mary J. "Chauffeured Versus End-User Access to Commercial Databases: The Effects of Tasks and Individual Differences," MIS Quarterly, 7:1; March 1983: 56-67.

Daft, Richard L. and Lewin, Arie, Y. "Where are the Theories for the "New" Organizational Forms? An Editorial Essay," Organization Science, 4, 4: 1993: i - vi.

Davis, G.B. and Olson, M.H. Management Information Systems (2nd Ed.), New York, NY: McGraw Hill, 1974.

DeSanctis, G. and Poole, M.S. "Capturing the Complexity in Advanced Technology Use: Adaptive Structuration Theory," forthcoming in Organization Science, 5, 1: 1994.

Drucker, P. F. "The Coming of the New Organization." Harvard Business Review, January-February 1988: 45-53.

Eisenhardt, K.M. "Building Theories from Case Study Research," Academy of Management Review, 14, 4, 1989: 532-550.

Friedman, A.L. Computer Systems Development: History, Organization, and Development. Chichester, UK: John Wiley, 1989.

Foulger, D.A. "Medium as Process: The Structure, Use, and Practice of IBM's IBMPC Computer Conferencing Facility." Unpublished Ph.D. Thesis, Temple University: Philadelphia, 1990.

Fujimoto, Masayo "Electronic Communication Technology in R&D Settings: Policy Implications for Public and Private Sectors," Unpublished S.M. Thesis, Cambridge MA: Massachusetts Institute of Technology, 1993.

Fulk, J., Schmitz, J. and Steinfield, C.W. "A Social Influence Model of Technology Use," in Fulk, J. and Steinfield, C.W. (eds.), Organizations and Communication Technology, Newbury Park CA: Sage, 1990: 117-140.

Giddens, A. The Constitution of Society: Outline of the Theory of Structure. Berkeley CA: U. of California Press, 1984.

Grudin, J. "Interactive Systems: Bridging the Gaps between Developers and Users," IEEE Computer, April, 1991: 59-69.

Howell, Jane M. and Higgins, Christopher A. "Champions of Technological Innovation," Administrative Science Quarterly, 35:2, 1990: 317-341.

Johnson, Bonnie M. and Rice, Ronald E. Managing Organizational Innovation: The Evolution from Word Processing to Office Information Systems New York: Columbia University Press: 1987.

Lewin, Arie Y. and Stephens, Carroll U. "Designing Postindustrial Organizations: Combining Theory and Practice," in George P. Huber and William H. Glick (Eds.) Organizational Change and Redesign, New York, NY: Oxford University Press, 1993: 393-409.

Mackay, Wendy E. "Users and Customizable Software: A Co-adaptive Phenomenon," Unpublished Ph.D. Dissertation, Cambridge MA: Sloan School of Management, Massachusetts Institute of Technology, 1990.

Maidique, M.A. "Entrepreneurs, Champions, and Technological Innovation," Sloan Management Review, 21:2; Winter 1980: 59-76.

Malone, T.M. and Rockart, J.F. "Computers, Networks, and the Corporation," Scientific American, September 1991: 111-117.

Miles, M.B. and Huberman, A.M. Qualitative Data Analysis: A Sourcebook of New Methods, Sage Publications, Newbury Park, CA, 1984.

Orlikowski, Wanda J. "Learning from Notes: Organizational Issues in Groupware Implementation," Proceedings of the Conference on Computer Supported Cooperative Work, Toronto, 1992: 362-369.

Orlikowski, Wanda J. "The Duality of Technology: Rethinking the Concept of Technology in Organizations," Organization Science, 3, 3, 1992b: 398-427.

Orlikowski, Wanda J. and Gash, Debra "Technological Frames: Making Sense of Information Technology in Organizations," ACM Transactions on Information Systems, forthcoming.

Poole, M.S. and DeSanctis, G. "Understanding the Use of Group Decision Support Systems: The Theory of Adaptive Structuration." In Fulk, J. & Steinfield, C.W. (Eds.), Organizations and Communication Technology, Newbury Park CA: Sage 1990: 173-193.

Poole, Marshall Scott and DeSanctis, Gerardine "Microlevel Structuration in Computer-Supported Group Decision Making," Human Communication Research, 19, 1, September 1992: 5-49.

Schon, D.A. "Champions for Radical New Inventions," Harvard Business Review, 41:2; March-April 1963: 77-86.

Strassman, Paul A. Information Payoff: The Transformation of Work in the Electronic Age, New York, NY: The Free Press, 1985.

Straub, D. W. The Effect of Culture on IT Diffusion: E-Mail and FAX in Japan and the U.S. Information Systems Research, 5, 1, 1994: 23-47.

Sproull, L. and Kiesler, S. Connections: New Ways of Working in the Networked Organization. Cambridge, MA: The MIT Press, 1991.

Tyre, M.J. and Orlikowski, W.J. 1993. "Exploiting Opportunities for Technological Improvement in Organizations." Sloan Management Review, 35, 1, 1993: 13-26.

Tyre, M.J. and Orlikowski, W.J. 1994. "Windows of Opportunity: Temporal Patterns of Technological Adaptation in Organizations." Organization Science, 5, 1 (forthcoming).

Ulfhielm, F. Spoken Japanese: Linguistic Influence on Work Group Dynamics, Leadership, and Decision-Making. In Organization -- Communication: Emerging Perspectives II. Ed. L Thayer. Norwood, NJ: Ablex Publishing Corporation, 1987.

Walsham, Geoff Interpreting Information Systems in Organizations, New York, NY: John Wiley and Sons, 1993.

Weick, K.E. "Technology as Equivoque: Sensemaking in New Technologies." In P.S. Goodman, L.S. Sproull and Associates (Eds.) Technology and Organizations. San Francisco, CA: Jossey-Bass, 1990: 1-44.

Yates, JoAnne and Orlikowski, Wanda J. "Genres of Organizational Communication: A Structurational Approach to Studying Communication and Media," The Academy of Management Review, 17, 2, April 1992: 299-326.

Tables and Figures

Figures 1, 2, 3, 5 and 6 are not available.



Table 1: Activities constituting Technology-Use Mediation

Mediating ActivityActionse.g., From Naga
EstablishmentSet up physical parameters and features of the technology.Transferred news-system from lab and installed it on project network.
Added new newsgroups, e.g., local, announce
Modified news-system to accept and send messages via electronic mail.
Modify institutional properties of the organization to facilitate technology assimilation.Persuaded management to allow use of news-system and to sanction technology mediation activities.
Obtained management's authorization to declare the news-system the official medium of the project.
Articulate the cognitive and behavioral routines through which the technology may be appropriated by users. Defined role of news-system vis-a-vis other media in use in the lab.
Proposed guidelines for news-system use.
Communicated the functions, role, and sanctioned use of the news-system.
Provided education around use of the news-system.
ReinforcementMaintain the operational fidelity of the technology.Maintained the operations of the news-system, e.g., new user accounts, backups, network management.
Help users adopt and use appropriate cognitive and behavioral routines to use the technology.Promoted use of the news-system.
Engaged in communication of and education in use of the news-system.
AdjustmentAdjust technical features of the technology to promote use. Added new newsgroups, e.g., union, headlines.
Modified news-system to accept Japanese characters in the subject line.
Alter usage rules and procedures to facilitate the use of the technology. Modified definition of newsgroups to clarify their purpose and distinguish them from each other.
Modified usage guidelines to ease use.
Episodic ChangeRedesign the technical functions and features of the technology.Extended functionality of news-system through archival newsgroups (e.g., guide), and external news-groups (e.g., techinfo).
Reconfigured the news-system to clarify relations among news-groups, to aggregate shared newsgroups, and to create greater coherence of the whole.
Modify institutional properties of the organization to facilitate change in technology use.Established moderated news-groups to monitor and control appropriate use, e.g., guide, official
Redefine cognitive and behavioral routines to facilitate change in users' appropriation of the technology.Redefined usage guidelines and routines for use to accommodate changes in news-system configuration.
Communicated the redefined functions, role, and use of the news-system.
Provided education for redefined news-system.


Footnotes

[1] Both Friedman (1989) and Grudin (1991) have discussed the role of third party players as mediators in systems development, but both position this role as an intervention in development. We use the term technology-use mediation to refer to activities employed within the context of use after development is complete.

[2] See Orlikowski (1992b) for a more detailed discussion of this process.

[3] Names of newsgroups are shown in italics. Newsgroup/* indicates a hierarchical nesting of related newsgroups.

[4] Fujimoto (1993) documents the examination of NAGA's policies and process in detail.

[5] While access to additional NAGA members and to users (including managers) would have added to our understanding of the situation, the problem of using retrospective data from a single informant is mitigated by the extensive news-system and NAGA e-mail data analyzed. The message texts, not the interviews, were our primary data.

[6] The other 215 e-mail messages dealt with the physical computer network.

[7] All messages cited in this paper have been translated from the Japanese.

[8] The announcement of this policy also appeared in a bi-weekly project newsletter distributed to all Acorn members.

[9] In Figure 5 we have dotted arrows 9 and 10 to indicate that the interactions as shown are always mediated through the institutional properties of the organization. We have shown a direct relationship for expository convenience only.

[10] The project was terminated for reasons unrelated to the news-system or the internal functioning of the project group.

[11] The lab already had technical support for the physical network.