Tools for Distributed Facilitation

We have extensive experience with distributed facilitation of synchronous virtual meetings since 1992. More recently, we have gained experience in asynchronous virtual meetings and have learned lessons from which best practices have been distilled. Best practices for synchronous and asynchronous distributed facilitation give insight into requirements for toolsets. Our experience with toolsets, including (1) widely available unstructured toolsets, (2) virtual reality offices, and (3) persistent visualization, provide direction for further tool development. 1. Distributed Facilitation Virtual meetings, meetings between geographically dispersed participants, are being supported by more and more software systems, as catalogued by Woolley [10]. Yet documented experience with virtual meetings remains limited, as exemplified by Beise et al. [2], who found that only 15 of the 34 facilitators investigated had ever participated in virtual meetings. Facilitators in this study had expectations about virtual meetings, but lacked experience to inform differences between virtual meetings and face-to-face meetings. Without experience, these facilitators don’t know what, if any, new toolsets will be required for virtual meetings. Some early studies have classified virtual meetings as synchronous, where participants are geographically dispersed but interacting at the same time, and asynchronous, where participants may join and leave the meeting and return at different times and any two participants may not be involved at the same time, but nevertheless interact with each other. Another way to classify tools is by the communication channels: text alone or with the addition of audio or video. Table 1 shows six modes of virtual meetings and 0-7695-0493-0 how challenges accumulate (i.e., each cell inherits the problems of the cells above and to the left) from mode to mode as we add communication channels and relax synchronicity. Activities of virtual meetings can be grouped into two kinds, content and process. Content refers to the matter of the meeting, the issues, objectives, and organizational outcomes. Process refers to the techniques, such as brainstorming, converging, summarizing, and voting, used to achieve outcomes, as well as to the organization of those techniques. Both process and content must be discussed explicitly. The term “back channel” refers to a conduit for communication about other than the main content. Typically, “back channel” communications are communications about the processes used to address content and the mechanisms used to support processes. For instance, the content may be that the group needs to agree on (a) measurable goals for improvement. The process is to (b) brainstorm possible goals through (c) an electronic brainstorming tool, (d) converge on best goals through (e) multi-voting, and (f) summarizing specific measures for these goals through (g) group writing. In this example, (a) is content, (b), (d), and (f) are processes, and (c), (e), and (g) are process support mechanisms. Group facilitators typically need to be able to explicitly separate and discuss these issues. We have had extensive experience in synchronous virtual meetings since 1992. For instance, Mittleman et al. [7] reported on findings and lessons learned from a synchronous distributed group facilitation study. These findings illuminate needs arising in virtual meetings. How can these needs be met? Our experience with synchronous meetings has given us an advantage in asynchronous, but we have found difficulties carrying over expertise from one to the other. Mittleman et al. [6] developed a set of best practices from the prior study and a related study of asynchronous distributed group facilitation. The characteristics of these practices point the way to the requirements for tools to assist in facilitation of virtual meetings. /00 $10.00 (c) 2000 IEEE 1 Proceedings of the 33rd Hawaii International Conference on System Sciences 2000 −−−−−−−−−−−→ less synchronous Synchronous Asynchronous y more channels Text Focus and context: Which window are participants looking at? Why aren’t participants typing in text right now? Establishing timeframe: Will people return to the same window tomorrow? How can I contribute without immediate feedback? Audio Give and take of focus: I can only listen to one voice at a time. When should I stop talking if I can’t see whether participants understand me? Labeling, storing, retrieving, sorting, processing clips: How can I choose the most interesting, relevant clip to listen to? Video Framerate, connectivity: Am I understood? Where do I restart after a technical interruption? How do I know one site has lost video reception and another has lost audio transmission? Magnification of audio problems: Can I store as much video as participants want to transmit? How do video clips fit into the task completion? What’s the minimum video I can view to get the gist? Table 1: Six modes of virtual meetings 2. Best Practices Findings The best practices developed by Mittleman et al. [6] are summarized in Table 2. They offer some insight into research questions and tool development for distributed facilitation. The best practices can be grouped by synchronous and asynchronous virtual meetings. In synchronous meetings, participants are geographically dispersed, but meet at the same time. Asynchronous meetings, by contrast, involve participants dispersed by time as well as geography. Global work teams, for example, must be able to operate asynchronously because so many time zones may be involved. For synchronous meetings, some important patterns emerge from the given best practices. These patterns share a common basis: signals we take for granted are missing from virtual meetings. We can’t see puzzled faces, hand gestures, body language, ebbs and flows of physical energy and involvement. We have to be more explicit in the absence of feedback signalling understanding or puzzlement, recognition or bafflement. Consequently, the best practices have the following character. Many practices call for more advance planning for virtual meetings as opposed to faceto-face meetings. Many practices call for verbal interjection during virtual meetings. Some practices call for separate channels for content and process communication. It is typical of the practices described that they can be measured after the fact to assess the degree of success. One practice suggests a team dictionary to solve the problem of a shared vocabulary not evolving during the virtual meeting process. Asynchronous meetings suggest some additional patterns as well as reiterating those for synchronous meetings. One practice suggests a chat window as a back channel for communication about process rather than content. Another practice mandates a meeting scoreboard. Several practices emphasize the importance of explicit instructions. 0-7695-0493-0 One practice insists on reminding participants about who is present at any given time. Practices relating to focus suggest visual images of data as well as a wide view of participants. The above best practices led the authors to seven steps for asynchronous meetings, summarized in Table 3. 3. Impact of Best Practices Findings Implementation of many of the best practices mentioned above could be distilled to a variant of the ever-unpopular dictum: work harder, accomplish more! Worse still, many implementations could be distilled to the demoralizing dictum: exchange and digest more information! Simon [9] identifies (as in many of his works) attention as the scarce resource to manage in an information intensive environment. In Simon’s view, any addition to information exchange and digest must viewed as a challenge for the management of human attention. In this spirit, the best practices mentioned above may be viewed as challenges to develop implementations to permit participants to accomplish more without relying solely on working harder and challenges to manage attention to alleviate the burden of increased information exchange and digestion. Are there mechanisms to ease the pain implied by the above best practices? There may be, but further research is needed in this new area. Our aim here is to identify research questions and tools based on prior findings. The study by Mittleman et al. [7] rests squarely on a link between cognitive load for three kinds of processes and group productivity. The three kinds of processes are (1) communication, (2) deliberation, and (3) information access. The best practices seem to increase the load for (1) communication the most, followed by (3) information access. The theory, called Focus Theory of Group Productivity, is described by Briggs [3]. /00 $10.00 (c) 2000 IEEE 2 Proceedings of the 33rd Hawaii International Conference on System Sciences 2000 Lesson Best Practices 1. Converging from a distance entails different requirements: People can’t see restlessness, agreement, sub-group coalitions. • Explicitly structure the convergence process. • Hold frequent process checks. • Use ad-hoc teams to negotiate compromise solutions. 2. Different techniques build a team over a distance: Building loyalty across sites requires bonding across sites. Different time zones have different lunch hours. • Engage in distributed breaks. • Create multi-site sub-teams for breakout work. • Photos and bios in People window matter more. 3. Following a meeting from a distance requires different signals / planning: cues are missing to signal success, completion, transitions. Completion of a major task needs a gesture or tally to give participants the sense of victory. • Create a scoreboard • Focus transitions • Enunciate interim goals •Make explicit pre-meeting plan 4. Focus wanders easily when working over a distance. Some tasks require inherently more processing from some participants. Bottlenecks occur when skills of subsets are required. • Engage vested interest • Assign parallel tasks • Default video feed should prefer overview to to talking heads • Default video feed should include data affected by process 5. People forget who is at a distributed meeting. Participants sometimes defer discussion until relevant players can be brought together, forgetting relevant players are together! • Remind them every 10 minutes • Use the people window a lot • Reflect user names when facilitating 6. Remote users feel like second class citizens. If one site is perceived as “main” site, morale droops at other sites, participation crumbles. • Place facilitator at an independent site • Employ co-facilitators at multiple sites • Focus local people on remote users 7. Audio channels need different attention in a distributed meeting: Lack of gestures and eye contact makes source and target of message ambiguous. • Talk to a person, not the group • Engage in a dialogue with someone you know • Frequently check on audio quality (intelligibility) 8. Network connections are unpredictable: focus wanders, participation diminishes, and chances of success dwindle during long waiting periods due to disconnects. • Pre-establish protocols and test • Anticipate a learning curve • Establish a re-bootstrap mechanism • Consider bandwidth cost with multimedia benefit • Have on-call tech support at each site • Have a fallback plan 9. Communication channels need to be specified: How should I send a message about a tool malfunction? Which window should I look at to see if we’re switching to a new task? • Separate task and process channels • Use video to support process rather than document talking heads • Use process support tools to focus group attention on specific channels 10. New physical environment issues emerge in distributed sessions: Meeting rooms are often designed to promote face-to-face interaction, deterring tool use, leaving remote users at a disadvantage. • Design to support online focus within rooms • Support distributed breaks • Provide online attendance monitoring 11. Asynchronous meetings differ from synchronous meetings. • Follow the seven steps for asynchronous meetings (Table 3) Table 2: Problems to Solve and Best Practices