Instant Messaging for Collaboration:
A Case Study of a High-Tech Firm

Authors

  • Anabel Quan-Haase,

    Corresponding author
    1. A professor at the University of Western Ontario in the Faculty of Information and Media Studies and the Department of Sociology. Her research investigates social networks, community, communication technology, and social change. She has been a Fellow of the Knowledge Media Design Institute and the McLuhan Programme in Culture and Technology. Her articles on Internet-related social and organizational change have been published in journals such as American Behavioral Scientist, Journal of Computer Mediated Communication, and Analyse and Kritik.
    • Address: Sociology, Social Science Center, University of Western Ontario, London, Ontario 6A 5C2, Canada

      Address: BTC, 635 N. Fifth Avenue, Ann Arbor, MI 48104 USA

      Address: Centre for Urban & Community Studies, University of Toronto, 455 Spadina Avenue, Toronto M5S 2G8, Canada

    Search for more papers by this author
  • Joseph Cothrel,

    Corresponding author
    1. (M.A., University of Michigan) is the president of BTC, a consultancy specializing in collaborative technology. He has spoken at many conferences on the subject of computer-mediated communication, including Communities & Technologies, the International Conference on Virtual Communities, Organizational Learning and Knowledge Management, and Vircomm. His articles have appeared in such publications as MIT Sloan Management Review, IBM's Knowledge Directions, Strategy & Leadership, and the Journal of Knowledge Management.
    • Address: Sociology, Social Science Center, University of Western Ontario, London, Ontario 6A 5C2, Canada

      Address: BTC, 635 N. Fifth Avenue, Ann Arbor, MI 48104 USA

      Address: Centre for Urban & Community Studies, University of Toronto, 455 Spadina Avenue, Toronto M5S 2G8, Canada

    Search for more papers by this author
  • Barry Wellman

    Corresponding author
    1. Professor of Sociology and director of NetLab at the University of Toronto's Centre. He is author or (co-)editor of more than 200 articles and three books:Social Structures: A Network Approach, Networks in the Global Village, and The Internet in Everyday Life. Prof. Wellman is Chair of the Communications and Information Technologies section of the American Sociological Association, Chair Emeritus of the ASA's Community and Urban Sociology section, Founder Emeritus of the International Network for Social Network Analysis, and President of the Sociological Research Association. His current research looks at how the Internet is affecting the “connected lives” of people at work, at home and in the community.
    • Address: Sociology, Social Science Center, University of Western Ontario, London, Ontario 6A 5C2, Canada

      Address: BTC, 635 N. Fifth Avenue, Ann Arbor, MI 48104 USA

      Address: Centre for Urban & Community Studies, University of Toronto, 455 Spadina Avenue, Toronto M5S 2G8, Canada

    Search for more papers by this author

Abstract

This article examines uses of instant messaging (IM) in a high-tech firm to illustrate how knowledge workers use this new work tool to collaborate with co-workers. The objectives are 1) to identify the collaborative practices of individuals in mediated work environments by looking at uses of IM; 2) to discern what social processes are reflected in employees' use of IM; and 3) to investigate how three factors proposed by Erickson and Kellogg (2000) to support social processes—visibility, awareness and accountability—are used in an IM system. Questionnaire and interview data show that while IM leads to higher connectivity and new forms of collaboration, it also creates distance: employees use the mediated environment as a shield, distancing themselves from superiors. We use Erickson & Kellogg's ‘social translucence of technology’ framework to discuss the social consequences of working in a computer-mediated work environment.

The Spread of Instant Messaging at Work

Designing collaborative tools that will be effective for people in the workplace is an issue that lies at the heart of research in computer supported cooperative work (CSCW). While there have been significant economic investments made into developing effective work group systems, the results have been rather disappointing. Consequently, although a wide range of tools are available, many organizations continue to rely principally on email for communication in collaborative work, complemented by peer-to-peer communication and calendaring (Pew Internet and American Life Project, 2002; Sillince, Macdonald, Lefang, & Frost, 1998). At the same time, the nature of communication is changing. Managers and professionals often communicate in fluid, multiple social networks rather than being solely embedded in a single work group (Teigland, 2000). Hence, many Internet-based communication systems are open rather than confined to the work group. Moreover, it has become clear that CSCW cannot be developed or studied solely online. Successful development and deployment must take into account the social context of use, and must understand the situations in which users switch among different means of communication (Nardi & O'Day, 1999).

In many organizations, employees are now collaborating via instant messaging (IM), either as a complement to email or as its replacement (Handel & Herbsleb, 2002; Herbsleb, Atkins, Boyer, Handel, & Finholt, 2002; Poe, 2001). Often, the adoption of IM has been at the grassroots level, as workers carried over the habit from home or social settings. What makes IM popular among workers is that it adds speed and ease to workplace communication, and eliminates the time typically lost to “telephone tag” or wasted trips to the office of a coworker who is absent or otherwise occupied (Handel & Herbsleb, 2002; Nardi, Whittaker, & Bradner, 2000). Because IM requires only an Internet connection and a small application loaded on a computer, it has often been under the radar of technical staff and management.

IM differs from email primarily in that its focus is on the immediate delivery of messages. To accomplish this goal, IM applications include: 1) a “pop-up” mechanism to display messages the moment they are received; 2) a visible list (“buddy list”) of other users, compiled by the user; and 3) a method for indicating when “buddies” are online and available to receive a message. IM applications also allow users to change parameters in the system in order to provide a more detailed view of their availability (e.g., in a meeting, out of the office, at lunch, away from desk). Other users are made aware of this status via automated replies from the user (e.g., “I'm out of the office now”) or by indications visible on the buddy list. They can then decide whether to contact the person later or send an email, voicemail, or other message that the recipient can respond to later.

Perhaps as a result of its use both at work and home, IM has garnered considerable media attention in the past few years (Poe, 2001). However, few field studies of IM use exist. Not surprisingly, given the newness of the technology, several studies have focused on groups being introduced to IM for the first time (Herbsleb et al., 2002; Isaacs, Walendowski, Whittaker, Schiano, & Kamm, 2002; Muller, Raven, Kogan, Millen, & Carey, 2003). Following a theme common to other investigations of CMC, other studies have focused on distributed groups, assuming that their need for CMC-enabled communication would be greater than that of co-located workers (Handel & Herbsleb, 2002; Herbsleb et al., 2002; Nardi et al., 2000; O'Neill & Martin, 2003).

The technologies under study have varied. Some studies have focused on experimental or custom-built IM platforms (Handel & Herbsleb, 2002; Isaacs et al., 2002). Only a few have looked at groups using the commercial IM products that represent the majority of IM use both inside and outside corporations today (Muller et al., 2003; Nardi et al., 2000). Finally, existing studies have largely concerned themselves with the functions and character of IM interactions (Handel & Herbsleb, 2002; Isaacs et al., 2002).

The Research Approach of this Article

This article examines how workers in a small, high-tech organization manage their IM use for the purpose of information exchange and collaboration. We compare the use of instant messaging to email and face-to-face (FTF)/telephone interactions. This allows us to identify the collaborative practices of individuals in mediated work environments by looking at uses of IM. Further, we analyze what social processes are reflected in employees' use of IM. Specifically, we use the social translucence of technology (STT) theory proposed by Thomas Erickson and Wendy Kellogg (2000) to evaluate the uses of IM at work and to examine new forms of collaboration.

The characteristics of the case we present differ in several ways from extant research on collaborative practices in computer-mediated work environments. First, the company under examination has employed IM extensively throughout the organization for more than three years. Hence, our research focuses on a technology that has been entrenched and embedded in the social and work processes of the organization. Second, the technology used by the organization, AOL Instant Messenger (otherwise known as AIM), was the most widely-used IM tool inside and outside the workplace, with 195 million users worldwide in 2004. Finally, while our study is similar to extant research in that it examines collaboration in computer-mediated work environments, it is different in that the focus is on the relationship between IM usage and social norms, structures, and networks, rather than on IM functions and interactions.

Social Mediation of CMC Use

Many analysts see computer-mediated communication (CMC) as stimulating positive change in organizations because of its rapid diffusion, diminishing costs, and ease of use (Jones, 1998; Rheingold, 2000). CMC can foster new forms of work by providing the opportunity for people with common interests to connect, overcoming limitations of space and time.

Yet, the positive effects of working in a cue-reduced environment are not without a cost. Studies have found that lean media were not adequate for transferring complex knowledge, whereas rich media, especially FTF, were much more adequate (Daft, Lengel, & Trevino, 1987; Kiesler & Sproull, 1992). Similarly, Fish, Kraut, Root, and Rice (1992) found that media with low social presence were inadequate for solving complex problems or communicating socially difficult messages, where low social presence means diminished cues about the characteristics of a person (e.g., gender, age, power, and social status) and no information on a person's facial and bodily expressions. Features of the task at hand influence how communication technology is used. In addition, the cue-reduced characteristics of digital/textual communication intersect with the interpersonal and organizational salience of such attributes as power and status (Short, Williams, & Christie, 1976; Sproull & Kiesler, 1991). Thus, the characteristics of media lead to specific choices about the use of different media.

What analysts often neglect when designing and analyzing CSCW systems is the fact that computer networks are social networks connecting people who share complex relationships among one another. Although CSCW takes place in the ether, its use is socially grounded. Therefore, analysis of IM use needs to move beyond the laboratory and into the real contexts of actual users. This is no mere rant about one research method's superiority, for as we shall show, a key reason why IM has become popular is that it provides cues about the status of interactants and their behaviors over time.

Social Translucence of Technology (STT)

The social translucence of technology (STT) framework compares computer-mediated work environments with real work environments. STT argues that when people interact in the physical world, they use social information to guide their actions. For example, Hillier (1996) recognized that one type of social information, visibility, facilitates informal interactions. His comparison of high-tech labs shows that open spaces lead to more frequent and spontaneous conversations between colleagues working on different teams, thus promoting cross-pollination of ideas and innovation. Therefore, socially translucent communication systems will be more effective if they provide social information about users (Erickson & Kellogg, 2000).

According to Erickson and Kellogg (2000), socially translucent systems have three characteristics:

  • a.Visibility: Movements and changes are used to guide interactions and exchanges.
  • b.Awareness: Knowing that others are present or available for communication. This knowledge can be used for starting communication, and it affects our communicative behaviors.
  • c.Accountability: Knowing that others know that they are there. This knowledge makes their behavior accountable. This is a relevant distinction because accountability and awareness do not co-occur in CMC as they do in offline communication.

As noted by Erickson and Kellogg (2000), visibility is the characteristic that brings awareness and accountability into play. For example, BABBLE, a system developed at IBM, allows its users to see each other's availability and current status of activity, thus enabling and inviting spontaneous interactions (Bradner, Kellogg, & Erickson, 1998). Bradner, et. al.'s study of the BABBLE system shows that visibility of employees' work activities promotes conversations, leading to higher performance. The interface and its ability to depict the social world facilitate information exchange.

Similarly, Bregman and Haythornthwaite (2003) identified visibility as a key element in online interaction. Focusing on interactions in distance learning environments, they treated visibility primarily within the context of how individuals present themselves online, and the means, methods, and opportunities they have for doing so. They found that, for distance learning students, how they were perceived—that is, how they were visible to others—was of major concern. The first step in becoming visible is to place a representation of the self into view. Speakers can become visible in a number of ways; Bregman and Haythornthwaite call this the repertoire for visibility. Speakers' decisions about the type of media used, the form of expression, the style, etc., are important components of their repertoire for visibility.

Social Networks at Work

Current theory emphasizes the value of informal social interactions for exchanging information, collaborating, and initiating spontaneous interactions (Cross & Parker, 2004; Wellman, 1997). The STT framework suggests that systems that depict social cues facilitate information and social interactions, enhancing performance. While the STT framework makes an important contribution to the CSCW literature by examining social aspects of design, it also incorporates a number of implicit assumptions about communication in computer-mediated work environments that need examination: 1) Actors have the same roles; 2) Power relations do not influence communication; and 3) Social cues are beneficial in promoting exchanges. These assumptions are not unique to the STT framework but rather are characteristic of the CSCW literature in general.

In this article, we use a social network perspective to expand on the STT framework and examine the assumptions discussed above. Social network analysis focuses on patterns of relationships between actors, emphasizing the influence of relationships on communication (Scott, 1991; Wasserman & Faust, 1994; Wellman & Berkowitz, 1988). It provides a theoretical, methodological, and substantive basis for extending CSCW research and development beyond its current individualistic approach, which treats “people as if they were socially disembodied beings without positions in social systems” (Haythornthwaite & Wellman, 1998, p. 1101). It points out that different types of ties link people and that each type of tie requires them to communicate differently. This makes it relevant and timely to examine the types of ties linking people and how these ties influence information exchange and collaboration in computer-mediated work environments (Garton, Haythornthwaite, & Wellman, 1997).

Work at KME: A High-Tech Firm

Setting

The organization in this study, which we refer to as KME, is a software and services company in the high-technology industry. KME's products and services enable its customers to implement online collaboration among clients, employees, or partners. KME's customers include Fortune 500 companies. The company operates in a fast-changing area and thus needs to be constantly developing and fine-tuning products and services to meet customer needs and remain competitive. This is the case in both the product side of the business, where new software products need to be designed, and on the services side, where processes and practices are not yet well-established in the industry.

Data Collection

KME has 80 employees. We selected 28 employees for our sample, drawn from two groups: software development and client services. These two groups were selected because they had existed as functional groups for at least one year and the tasks accomplished within each group are interrelated and fairly homogeneous. Hence, we expected to find stable patterns of communication and technology use.

Of the 28 employees selected, 27 completed the questionnaire, yielding a 96% response rate. 16 employees from the client services group and 11 employees from the software development group participated in the study. In total, there were eight women (30%) and 19 men (70%). The duration of employment at the organization ranged from 5 to 43 months with a mean of 27 months. The education of respondents was distributed fairly evenly among high school graduates, university graduates and those with graduate degrees.

The numerical data were supplemented by in-depth interviews conducted by Quan-Haase with a sub-sample of 10 employees, and observations of daily work practices (see below). Participation in the interviews and observations was voluntary.

Each interview lasted about 45 minutes. Interviewees comprised employees from a range of positions and roles. Semi-structured interviews provided flexibility to follow important leads while covering the same set of questions in all interviews. Transcribed interviews were sent to interviewees for review and approval. To guarantee the confidentiality of interviewees, pseudonyms are used throughout our research reports.

Our analysis is principally based on questions in the interviews about the use of media. The purpose of these questions was to understand participants' personal media use and unique media profiles, including what media participants believe are appropriate for communicating with different types of communication partners and for communicating different types of messages. This approach allowed us to examine the social context of media use.

Specifically, participants were asked what type of media they use on a daily basis to communicate with colleagues inside and outside KME. They were also asked about each medium's relevance for their work, in terms of frequency of use and types of tasks performed. To obtain more detailed information on media use, we also asked participants what aspects of each medium they perceived as most useful, and why. In addition, participants were asked to report what type of medium they thought of as optimal for specific kinds of communication and information searches, and to discuss the characteristics of the medium that made it the best choice. Participants were also asked to report specific instances that were representative of their use of various media.

The interviews were tape-recorded, transcribed, and imported into NVIVO software, a text analysis package that is specifically designed for the analysis of interviews (Richards, 1999). We followed Anselm Strauss' grounded theory when using NVIVO to code the interviews, developing themes through coding (1987; also see Strauss & Corbin, 1998). We discuss here only the themes that relate to media use and the maintenance of community.

Quan-Haase also observed full-day work practices to see how people handle CMC and how they fit CMC into their relationships and communication. Over one week, she observed all of the 10 interviewed employees, spending one full day with each of these individuals.

The one-on-one observations started at 9:00 AM and finished when the employee left the office (at approximately 4:30 PM). All FTF and online interactions were observed and recorded, including email, instant messaging, FTF, and phone exchanges. The start and end time, duration, and content of interaction were recorded. Although participants were given the opportunity to conduct private exchanges, no one did.

The analytic framework employed for the observations was a combination of grounded theory and social network analysis. Notes were taken on a daily basis and behaviors were recorded in a time diary. The notes were then coded and themes were developed in the same grounded theory manner as with interviews. Social network analysis guided the observations by focusing our attention on social relationships and their influence on the choice and use of media (for more details, see Quan-Haase & Wellman, 2004, 2005).

Collaborative Practices In Computer-Mediated Work Environments

Collaborative Practices at KME

Collaboration abounds at KME. Employees share best practices and engage in joint problem-solving. Collaboration in this technology-intensive firm takes place both FTF and via CMC. We asked participants in the questionnaire to indicate how often they used each of three media (FTF/phone, email, and IM) to communicate at each of three distances (work group, intra-organization, and extra-organization). The scale used to measure frequencies ranged from 1=“never” to 7=“daily.” For the purpose of the present analysis, the original 7-point scale was transformed into days per year: “never”= 0; “a few times a year”= 5; 1/month = 12; “1/week”= 52; “several times a week”= 130; “1/day and several times a day”= 365. Table 2 shows the mean frequency of use (standardized into days per year) of FTF and telephone, email, and IM for within group, intra-organizational, and extra-organizational communication.

Table 2. Communication practices at KME: days per year and percentages of communications media
Work MilieuFTF & PhoneEmailIMTotal Milieu %Row %
Work Group240306306852100%
 26%36%36%54% 
Intra-Organization99213215527100%
 19%40%41%33% 
Extra-Organization2110372196100%
 11%53%37%12% 
Total Media %3606225931,575 
 23%39%38%100% 

Communication within the group relies more on email and IM than on FTF, accounting for nearly three-quarters of all intra-group communications. Nevertheless, because of physical propinquity, the highest frequency of contact within the work group is FTF/phone.

Intra-organizational communication relies even more on CMC: Email and IM are each used more than twice as often as FTF and the telephone. Email is preferred even more for extra-organizational communication, accounting for about half of all such communication, with IM accounting for about one-third. Both social distance—IM is a more informal medium—and incompatible IM systems limit its use in comparison to email. Where the daily use of email extra-organizationally is one-half of intra-organizational use, the daily use of IM extra-organizationally drops to one-third of intra-organizational use. Moreover, because KME's clientele and supply chain are national and international rather than local, there is little FTF extra-organizational communication.

While we often think of CMC as primarily aiding communication at a distance by bridging time and space constraints, the primary use of email and IM at KME is for contact within the work group. About half (52%) of all IM and email (49%) communications are within the group. This high use of CMC occurs even though the employees work in the same building. The two groups occupy relatively small, densely organized, open-plan spaces on two floors of a small, three-story office building. Each group is located together on a different floor. The size and configuration of their workspaces encourage FTF and telephone contact rather than online interaction. Nevertheless, employees in this co-located setting access human sources primarily via CMC.

Informed by our findings at KME, we introduce the term local virtualities1 to describe physically bounded places where people communicate via computer-mediated means, creating dense networks of collaboration. The fact that people are not interacting visibly in public spaces does not mean that they are isolated. They go online to send and receive emails, and they use IM to talk with colleagues. This finding is contrary to theories of network and virtual organizations that argue that CMC creates sparsely knit, boundary-spanning structures (Monge & Contractor, 1997, 2003). The fact that KME's CMC has created local networks of collaboration suggests that CMC can lead to new and unexpected forms of interaction and community that blend both local and distant communication.

Collaborative Practices in IM

In this section, we describe how employees at KME manage their IM systems on a daily basis, how IM promotes collaboration, and how new forms of collaboration emerge through the use of IM. Although email is used as much as IM for group and organizational communication at KME, in this article we focus on the less-frequently studied use of IM. Our findings are in accord with previous studies (Nardi et al., 2000) that have found that employees use IM to exchange information; pose quick questions and clarifications; arrange and coordinate FTF and telephone meetings, both formal and impromptu; conduct simultaneous conversations over multiple media; and solicit immediate responses. We also found that IM speeds up the exchange of information and creates new forms of collaboration among employees.

When employees arrive at their workstations in the morning, the first thing they do is log on to the IM system. The IM system then displays the online presence of the employee to all other members of the organization who are also logged on to the system. Although KME has no explicit norms around collaboration, employees implicitly understand that those who are physically present at KME are expected to be logged on.

Employees perceive being logged on to the IM system as providing them with opportunities for communication. Even when employees are not at their desks, but at a meeting or out for lunch, they continue to be logged on to the system. In those cases, the system then changes their status from available to idle. If the employee chooses, the system also replies to incoming IMs with a customized message, such as “out of the office,”“in a meeting,” or “right back.” This information is important because it gives other employees more detailed information about each other's availability, allowing them to estimate when colleagues will be available for consultation. Availability information is particularly crucial for KME in that most projects are collaborative and the input of other employees is often central to completing a task.

Most collaboration via IM consists of asking questions about specific aspects of work. The questions often involve clarification or codified knowledge. Most of these informal collaborations take place one-on-one rather than in groups. IM is used most commonly for one-on-one exchanges in which an individual contacts another individual to discuss needed information. It is also used, to a more limited extent, for in-depth problem solving, where one person helps another make sense of a problem and think through various strategies to find a solution. In other instances, help consists of referrals to others who can provide answers or help solve problems.

Through information about colleagues' availability, opportunities for interaction open up. In the framework of social translucence, information availability provides a social affordance that makes others aware of the possibility for collaboration. One employee expressed how the system allowed them to ask others for help: “I see that they are online, I need an answer now, I need to talk to them now, I will ping them.” Furthermore, at KME, IM not only provides opportunities to collaborate; it also plays a role in the collaborative process. As one programmer reported, employees perceive IM as a principal way of connecting with each other: “I just know that if you call or send an IM, you will get a faster response than email.”

IM fundamentally affects the nature of collaboration at KME. It allows employees to ask for advice on demand. When a problem is difficult or demands expertise in another area, employees reach out to their colleagues for help. This occurs on an ongoing basis, creating a culture where most problems are solved by a network of people drawn into the problem-solving process whenever necessary. This new form of collaboration facilitated by IM is in synch with the notion of an “active social network,” in which the only relationships active at any time are those useful for the particular task at hand. It saves time and effort on the part of other colleagues, who can work on their own projects and are only asked to participate when needed. Unlike traditional stable work groups, KME employees are forever reconstituting their active ties to deal with issues at hand. Of course, such reconstitution is not random: Over time, people develop comfortable working relationships through prior collaboration and socializing. One employee described how IM is used to obtain answers quickly, on demand:

I use IM a lot. IM is great if you have one question that you just need an answer to. When you need to explain something in detail—an outline, kind of a business case for doing something, or for getting somebody to take action—email is the best.

Implicit norms at KME dictate that IM takes priority over other media. It even takes precedence over FTF, despite the more abundant media cues present in FTF. Employees know that others rely on them to respond quickly. Our observations discovered that as soon as an IM message appeared on someone's screen, they would glance over and read the message. In almost every case, they immediately responded, either with a substantive reply or with a short reply indicating that they received the message and would provide a full response later. As part of the culture at KME, people feel compelled to reply, even if it is only to say that they are in a meeting and will respond later.

The precedence of IM over other forms of communication may be a result of IM's functionality, in which a pop-up window automatically appears on the screen when an IM message is received. In this respect, IM is different from email, which requires the user to launch the program in order to read or send messages. As a result, an IM cannot be ignored as easily as an email. At KME, this difference influences message-sending as well as message-responding. Senders recognize IM as the best medium when timeliness is crucial. Email is used when the message is less urgent.

Employees use FTF/phone interactions less often, and perhaps differently, because they perceive them to be more disruptive than IM to their colleagues' work process. Unlike a FTF/phone interaction, users can choose to reply substantively at a later time, or, in rare cases, not at all—choices that could be difficult or awkward in a FTF setting. IM thus occupies a useful niche within the range of communication media available to employees: less interruptive than a face-to-face interaction or phone call, but offering a greater potential than email for obtaining an immediate response.

Collaboration at KME occurs in an “always-on” mode (Gray et al., 2003; Quan-Haase & Wellman, 2000) because employees spent most of their time working at a networked computer. Communication occurs almost simultaneously through multiple media, rather than sequentially. For example, workers answer an IM while having a conversation. Employees do not switch between media and people to communicate, but use various media simultaneously to interact with different people.

In sum, the always-on mode allows employees to be aware of one another: 1) when someone logs on at the beginning of the day or logs out when they leave work; 2) when someone is away from their desk; and 3) when someone is busy. Communication practices at KME reflect a shift from the traditional way of working, where communication usually occurs outside the regular workflow. Because work at KME is conducted primarily online, communication occurs simultaneously with other tasks. Workers treat IM as simply another task within a multitasking work environment. However, as will be discussed below, this is not to suggest that adoption or adaptation is easy for all workers (for a more general discussion, see also Quan-Haase & Wellman, 2000).

Social Processes in the Use of IM

We noted earlier that IM is an important means of collaboration at KME. IM provides employees with an alternative form of asking for help, solving problems, and exchanging information. However, to understand fully the use of IM for collaboration at KME, it is important also to examine characteristics of the application and how these characteristics interact with social factors. We use a combination of the social translucence of technology (STT) framework (Erickson & Kellogg, 2000) and social network analysis to describe how social processes play a role in the use of IM.

Characteristics of Actors' Relationships

Through its ability to convey some awareness of a person's availability, IM creates a sense of connectedness (Nardi et al., 2000). Most participants at KME see visibility as a positive feature of the system; they feel that it is important to be able to see if colleagues and friends are online. However, many do not perceive visibility to be a positive feature in all contexts: Its value depends on the relationship, especially the closeness of the relationship, and the interdependency of tasks.

Close relationships.

If two individuals share a friendship relationship, they feel comfortable using availability information to contact each other. One KME worker describes such a relationship this way:

I enjoy communicating with him—I can say: “What is going on?” and “by the way I have this technical question.” Whereas other people I wouldn't want to IM because I don't have that kind of relationship with them. Unless it was really important, then I would, but I would be less likely to IM because I would not want to disturb them without a real excuse.

Thus, although awareness information is available to everyone on the system, it may have greater utility for those in a close relationship.

This characteristic is related to the limited capability of current IM systems to display more detailed information on availability. Although IM systems attempt to accommodate this need with icons or customized messages (e.g., “I'm on a conference call”), such messages cannot convey the many nuances of availability that exist in the workplace (e.g., “I'm available to others, but not to you”). As a result, people are not fully aware of whether it is convenient to contact another person. Clearly, eliminating one barrier (lack of awareness) does not eliminate other barriers (social, hierarchical) that may exist to the free flow of communication.

Task interdependency.

When two individuals are working on the same project, they feel comfortable asking each other questions via IM. In other words, there is a recognition that, when work tasks are interdependent, input is often required in order for individuals to be able to continue their work. IM in this case breaks down social barriers and allows individuals to ask questions and tap into each other's knowledge. Therefore, being able to monitor others' availability is useful for people working interdependently.

Sense of Community in Instant Messaging

IM supports one-to-one communication (user to user), one-to-many communication (user to multiple users), and many-to-many communication (where a user can initiate a session in which all invitees can interact with one another). Many-to-many communication can foster a sense of community among participants, since users become more aware of one another and their opinions, understandings, likes, and dislikes (Nardi et al., 2000; Rheingold, 2000). With features such as buddy lists and presence indication, IM extends this awareness to patterns of behavior, allowing users to become more conscious of—and potentially more involved with—the dynamics of their groups and social networks.

IM's community-reinforcing potential occurs within a user-centric context (Alvestrand, 2002). Only those individuals who are added to the buddy list become part of a user's community. At KME, employees typically download a buddy list consisting of all employees of the company and install that list on their computer. Hence, the system creates higher awareness of others' availability throughout the company. However, employees are free to add or delete names from their list, and to create categories to make close colleagues more visible and others less so. Our findings showed that users are usually aware only of the availability of their friends and colleagues who work on interdependent projects. Awareness is strongly influenced by social relationships among IM users, and data on presence or availability alone does not create a sense of community with others.

IM does not always create greater connectivity. KME employees also see IM as a useful tool because it creates distance between themselves and their superiors. When they have to deal with difficult decisions or discuss sensitive topics where they know that they may disagree with their superiors, they often prefer to communicate via IM rather than FTF. The reason is that IM allows them to reflect on their superior's opinions and provides them with time to think about their own reactions. IM in this case provides a barrier between communicators, albeit a barrier that has positive benefits for employees.

Organizational Control via IM

While awareness can make interactions easier and more efficient, it can also create the potential for control where users are not equal in status or power. Because of the KME norm to log on to the IM system upon arrival at the office, everyone knows who has arrived and is available for contact. People have flexible work hours, especially at higher levels in the organization, and work hours are not explicitly monitored or controlled. Nonetheless, while IM is not explicitly used for control at KME, the fact that the system provides the potential for accountability makes some employees uncomfortable.

One functional aspect of the IM system used by KME is particularly important. The user name appears in black if the person is actively typing on the keyboard, and becomes shaded once the user stops for more than a couple of minutes. (Users can configure the time interval or disable this feature altogether, although it is not clear that all employees know this or feel their superior would approve.) In this way, the system shows not only if the person is present at his or her desk, but also if they are using the keyboard. Some employees indicated that managers could use keyboard activity as an indicator of work being performed and expressed concern about this as an abridgement of their privacy or a mechanism for control. One of the managers interviewed said that the system allowed her to check who was in the office (and, by extension, who was out of the office). Accordingly, employees reported that they would log on specifically to make sure that others saw that they were at the office. Thus, the fact that the system creates visibility means that it also creates accountability and a means for monitoring.

Another form of accountability is that employees feel compelled to reply to messages because as receivers they know that the senders are aware that they have received the messages. Thus, the awareness of others' availability leads to expectations in the sender about how long it should take the recipient to reply. If a recipient does not reply promptly, he or she may be interpreted by the requestor as evidencing a lack of respect or concern for the requestor's needs.

To avoid misunderstandings, users who are unavailable to respond often activate the automatic reply feature within IM. This feature sends one of several standard messages, or a message created by the user, to indicate that the user is not available. This feature is often used to negotiate accountability:

If I know I am going to be busy or I need some time to concentrate on an issue, I turn one of the following messages on: “I am on the phone, get back right away,” or, “There is someone in the office, will email when free.” One of my colleagues uses “Of course I am not here now.”

People feel accountable to respond to the messages that they receive. In most cases, messages are replied to quickly: within one minute. Even during FTF meetings, workers monitor their computer screens and, if an important IM pops up, they reply while continuing their FTF conversation.

I just know that if you call or send an IM, you will get a faster response than email. Although sometimes I have to resend them, like “HELLO.” And they say: “I don't remember getting this.”

The downside of social translucence in IM is that a user may not always want to convey social information about him- or herself. We observed at least three situations in which workers did not want their social information to be displayed. First, employees working under time pressures often did not want their availability information displayed because this meant others could interrupt them. Second, employees often wanted to be available to a sub-group of employees, but not to all employees—at least not all the time. Third, if employees were working on a difficult problem, they did not want to be interrupted. Many problems require full attention and employees felt in these cases that IM distracted them from the task at hand.

When workers do not want their social information displayed, they sometimes engage in subversive strategies. Alice, for example, did not log on when she had a pressing deadline. Others leave their computers and IM application running 24 hours a day, to eliminate the logging on/off activity that could be used by others for control purposes. In the first case, a worker who is present appears to be absent; in the second, a worker who may be absent appears to be present. These subversive uses of technology are in accord with social constructivist views that emphasize the distinct uses of technology (Bijker, Hughes, & Pinch, 1999; Markus & Keil, 1994). Therefore, while the system sometimes provides more social information than users may want to provide, it also provides more flexibility than is available in real physical settings. In real physical settings, one cannot easily deceive others about one's presence or absence. The IM system provides occasion for both disclosure and concealment.

Visibility information in IM is limited to a person's online status. The system does not convey information about what the user is doing, with whom the user is currently engaged, etc. While future systems could convey these types of information, it is unclear what their social consequences are and how they would be received by users, especially considering the potential overrun of privacy boundaries (Gray et al., 2003).

Status Relationships and IM

A person's status within the hierarchy of the organization plays a key role in how their messages are replied to. For example, Leonard, a senior manager, received two IM requests during our interview. When a message popped up on the screen, Leonard glanced over quickly and read it. Both times he excused himself and initiated an exchange that lasted for 2-3 minutes. When asked what had happened, he replied: “I usually do not answer messages while I am engaged in a FTF meeting unless they are short questions or are urgent.” When asked who had sent the message, he said that it was from his superior. We noticed similar reactions from other interviewees. As soon as an IM message appeared on their screen, they would glance over and read the message. As part of the culture at KME, people feel compelled to reply, even if it is just to indicate that they are in a meeting and will get back to the request later. Clearly, IM often comes before FTF, despite the more abundant social cues present in FTF. With IM, the status of the communicator and the urgency of the message can be more compelling than the physical presence of someone FTF.

One worker reported that he preferred IM because it gave him social distance from his superior. This distance was afforded by the absence of social cues in IM. Although the lack of cues in CMC is often seen as inhibiting communication, here it is perceived as positive and used strategically.

Instant Messaging as a Dyadic Process

There are clear variations in how people communicate via IM, suggesting that IM use is in large part a dyadic process. Over time, people develop a good understanding of each other's media preferences and user habits. A programmer reports on her understanding of her boss's use of IM:

I know that my boss prefers to be reached over the phone because he finds IM disturbing. I prefer IM, but if I know that it is going to be a longer message I will pick up the phone and call him. With everyone else I just send an IM message. It is so easy and fast. I know they will get it right away.

People use the media that their communication partners prefer. Besides having a good understanding of individual preferences in media use, people also know the best way to reach particular people.

Conclusions

Local Virtualities

We have found that workers at KME rely heavily on CMC for communication both within and outside the organization. In accord with previous studies on the uses of IM at work, we found that IM is used extensively to exchange work-related messages, coordinate and arrange meetings, and inquire about colleagues' availability for discussion. This corroborates Handel and Herbsleb's (2002) finding that IM is largely used for work-related exchanges, while non-work-related exchanges are secondary.

By focusing on the use of IM in a physically co-located group, our study questions the assumption made in the literature that IM is exclusively used, and useful, for work in distributed teams. Further, we show that, at KME, IM is used as much for internal communication as for external communication. We coined the term local virtuality to describe these new forms of work, where co-located workers using IM for communication create dense networks of information exchange (Quan-Haase & Cothrel, 2003).

Furthermore, our results show that IM not only creates higher connectivity and a sense of community, but in some cases also functions as a barrier. Employees use IM in part as a way to create distance between them and their superiors. This is particularly useful when difficult decisions have to be made or sensitive topics discussed.

Social Translucence as Design Principle

This case study of IM in work environments supports the framework of STT and shows how visibility, awareness, and accountability can be used to evaluate the uses of IM at work. The study finds that people use availability information to coordinate tasks, exchange information, and create social links. Furthermore, people negotiate their availability through IM by manipulating their availability status on the system. The availability of employees depends on the complexity of tasks they are currently engaged in and on how busy they are. Yet, as shown in Figure 1, declarations of availability and responsiveness to IM messages also depend on the relative status and power of the communicators.

Figure 1.

Social design principles in IM systems

The availability information in IM also provides a sense of connectedness, bringing people together. This is in accord with other CSCW findings showing that it is not informal communication that seems to be responsible for the success of teams, but specifically opportunities for problem-oriented and unplanned, spontaneous interactions that allow people to take advantage of the collective knowledge available in the team (Hinds & Kiesler, 2002).

Far from being limited to communication within work groups, IM facilitates many “personal” online communities at KME. It is used selectively to communicate with certain peers, superiors, and subordinates. It is used to communicate with people elsewhere in the organization, and even used at times to communicate with people outside the organization. IM at KME supports what Wellman has called the turn to networked individualism: connected, shifting interactions in sparsely-knit, loosely-bounded social networks rather than focused interactions in densely-knit, tightly-bounded groups (Wellman, 2002).

Information about visibility, a distinguishing feature of IM tools in comparison to other CMC systems, enhances and supports social interactions. The uses and meanings that KME employees give to visibility information depend on a number of factors, such as the status relationships among users, their task interdependencies, their past history of collaboration, and their informal socializing. Therefore, visibility information is a potentially relevant feature of the IM system at KME when analyzed in its social context.

Awareness is a key affordance of the KME IM system, and is a byproduct of visibility information. If close ties link employees, then they use visibility information to stay aware of each other's online activities. However, if workers do not know each other or do not share strong ties, visibility does not result in awareness. This reflects previous research by Handel and Herbsleb (2002), who concluded that “the usual indication of presence … does not itself provide sufficient conversation-starting material” (p. 9). While these authors argue that further cues need to be built into IM systems to encourage spontaneous, informal exchanges, our findings suggest that social networks are the key, rather than the specific features of the IM system. Rather than adding more social cues, a more useful approach would be to use tools that provide visibility in combination with FTF meetings so that people can become acquainted with each other and establish the trust necessary to use IM to initiate conversations.

Accountability is a product of visibility information. It has variable impacts at KME. If relationships are characterized by large discrepancies between the status of the users, visibility is more likely to be used as a means for accountability. Such accountability is reflected in the behavior of both lower and higher status workers. Accountability in IM works differently for employees with lower and higher status. For lower status employees, accountability means that information requests need to be answered in a timely manner because higher status employees are aware of the fact that they are working at their desks. For higher status employees, accountability means that they can expect responses within a short time. Higher status employees can also monitor the online behavior of their employees by observing when they log on and out of the system.

Several lessons emerge from this study for the design of collaborative systems. This study has shown that visibility is a useful feature for promoting informal, ad hoc exchanges, a sense of community, and ease in collaboration. However, visibility primarily promotes these collaborative exchanges among people who share strong, trusting ties. This study has also revealed a pattern of networked individualism, consisting of shifting sets of communication partners rather than interactions among a densely knit group.

Finally, our research has shown that social translucence is a useful framework for examining existing communication systems and providing guidance in the design of new systems—in particular real-time communication systems such as IM. At the same time, the social translucence framework needs to be expanded to account for the influence of social factors such as power, social relationships, and norms within organizations.

Acknowledgments

Research underlying this article has been supported by CITO, the Centre for Urban and Community Studies, the IBM Institute of Knowledge-Based Organizations, Microsoft Research, Mitel Networks, and the Social Science and Humanities Research Council of Canada. The first author acknowledges assistance from the Alumni Research Awards Program, Faculty of Social Science, the University of Western Ontario. We thank the editors and an anonymous reviewer for useful comments. We especially want to thank all the employees at KME who completed the survey, and even more so, those employees who also participated in the interviews and observations.

Notes

  • 1

    We use the term “local virtualities” here to describe work settings where people are physically near each other (i.e., co-located) and yet use CMC to exchange information, share best practices and socialize. See Quan-Haase and Cothrel (2003) and Quan-Haase and Wellman (2004, 2005) for more extensive descriptions and discussions of local virtualities at KME. The term “local virtuality” has also been used in a study of email use in rural communities (Koskikallio, 2002).

Ancillary