In the last month, the entire world has moved online. Teams are using collaboration tools like Miro, Figma, and Slack more than ever. User researchers are using virtual facilitation tools such as Validately to conduct research that would typically happen in face-to-face labs. Zoom has become a household name due to its use in online classrooms for displaced students as well as meetings and happy hours for now-distant team members. Many UX-related issues have been encountered as we’ve made this shift. This online trend will continue well after COVID-19 isolation ends and raises the questions of how will we, as UX practitioners, make these software experiences better?

Why are there so many issues with collaboration software online? Possibly because the product teams who built them tried to apply the methods we’d use for single-user systems to groupware.

Design researchers in the field of CSCW (Computer Supported Cooperative Work and Social Computing) have been considering the elements of good software design for groups for over 35 years. CSCW is a human-computer interaction special interest group that focuses on “the design and use of technologies that affect groups, organizations, communities, and networks,” [1] and it is not necessarily limited to work nor to cooperation. One way in which CSCW research and design have contributed to HCI more broadly is through work investigating how technology can support teams, especially teams that cannot work face-to-face.

Researchers and developers consider the task of understanding group collaboration a challenging problem because there are many social factors that can affect the experience of using these platforms. Those are harder to test using traditional UX methods. Consequently, traditional experimental and laboratory methods that remove the software from its context of use may obtain simplistic results that do not generalize well to real-world situations [2]. Traditional ethnography can help, but it works best at the beginning of design or at the end as evaluation, and it’s not suited for the way most of us work today in rapid iterative design teams.

UX Design for CSCW Technology

In order for software to effectively support team interactions, it must support taskwork (can the task be completed with this tool) and the constituent parts of teamwork (both its social and affective elements as well as the mechanics of collaboration) [2]. 

Researchers suggest that some groupware usability problems are not strongly tied to social or organizational issues, but rather are caused by insufficient or mismatched support for the basic activities of collaboration. These activities can be thought of as the mechanics of collaboration, the small-scale actions and interactions that group members must carry out in order to get a shared task done.

Heuristics for Groupware 

Here are some best practices when designing for groupware [3]:

Provide the means for intentional and appropriate verbal and gestural communication

Chatting, texting, FaceTiming, or Zooming: these are ways that we are able to intentionally, or directly, communicate with one another. Each of these modalities affords different kinds of communication; for example, deictic references (e.g. pointing and saying “that tissue box”) are best supported by videoconferencing. These references can be supported by other kinds of software through careful design attention to cues.

Provide the means for consequential communication in the workspace

Consequential communication is received as information is “given off” from objects in the shared workspace, or the shared environment in which work is taking place. Words on a shared document communicate the work processes that have already unfolded, and the moving cursor shows where someone is currently typing. These digital cues approximate the body movements and moving pencils (or typing fingers) that communicate action in face-to-face workspaces.

Support people with the coordination of their actions

Coordinated action is supported in text chat rooms, for example, by indicating when someone is typing to keep users from effectively “talking over one another.” In videoconferencing, users often speak over one another due to a combination of the lack of such cues and audio delays; however, in some instances users may raise a hand or a finger to indicate their intention to speak. Some software also includes digital solutions to this problem, such as the “raise hand” button in Zoom.

Allow people to adapt their plans and create new ones in the workspace

A lot of the planning process comes before collaboration even starts, but some planning still is important to support during collaboration, itself. For example, there is often on-going negotiation around who will do which tasks, and where certain content should be located. This kind of communication needs computer support.

Manage the transitions between tightly and loosely-coupled collaboration

Coupling refers to the degree (tightly vs loosely) to which people are working together. It is important to maintain an awareness of who is in the workspace, where they are working, and what they are doing in order to offer assistance or alter plans to better support the group task. For example, this information can be gathered from the colored cursors which represent specific users in collaborative tools and supplemented by chat programs. From these, it is possible to know where group members in a shared workspace are working and to decide what level of coupling is necessary to others or oneself. This switch necessitates an understanding of what someone is doing, and when the switch is initiated to offer assistance through tight coupling, it is important to know what someone has been doing. Ideally, both assistance that is specifically requested and that which is opportunistic should be possible.

Provide protection

Versioning, Forking, and Locking are perhaps familiar methods of protection in digital tools used by multiple people. These protect the work that is done by one person from the actions of another person, whether that be by accidental or intentional behavior. 

Next Steps for Designers

For designers, these heuristics can guide the rapid creation of prototypes that explore the different heuristics, which were first introduced by [3]. For example, these can be used to ensure that a designer’s prototypes support tasks that require explicit communication or monitoring of information. If a prototype focuses on the mechanics of single-user interfaces, which require only an understanding of how that user’s actions interface with the software, the prototype will not support research and evaluation that leads to more group-friendly software, or groupware.

UX Research for CSCW Technology

Cooperation and collaboration need specific UX support that differs from other UX pursuits in key ways. First, it is important to recognize that cooperation and collaboration are, in themselves, different things. The technologies that support these activities must account for the differences between cooperation and collaboration, and investigating these software requires different skills from UX professionals. 

When evaluating groupware, the typical one participant/one researcher modality no longer fits. When evaluating software created for coordinated action, it is important to uncover its impact on the work of two or more people. This shift requires additional considerations:

• When conducting research with additional participants, how will their interactions be recorded? It may be useful to train additional researchers to observe and moderate the study.
• If the software is to be used remotely, multiple moderators become necessary because participants should be separated to facilitate observation of (pseudo-)remote work.
• It may be helpful to recruit participants who are already coworkers, since coordinating the arrival of multiple participants with potentially disparate work schedules can be difficult. Also, people who are familiar will better be able to overcome the limitations of the software.
• To get the best view of how useful a particular groupware is to team processes, it would be best to study two (or more) individuals who have never before met. One way to accomplish this is to train a team member to act as a confederate. Research confederates are a common practice in psychology, where it is important to keep the human-human interactions constant across study sessions. In UX research, this could be additionally helpful in maintaining the focus of the study, and the confederate could skillfully lead the participant-user to the parts of the software that are specifically under review.

This article focuses on collaborative considerations, but cooperation necessitates some of the same mechanics. For example, it is still important to understand what others are doing and to communicate directly with cooperative partners. However, the management of coupling and availability of consequential information is less important because group members work on specific tasks that are decided before the work begins. 

Next Steps for Researchers

By leveraging the considerations presented here, researchers can structure their studies to account for the unique ways users will interact with the software. By scheduling multiple participants, the researcher will be able to observe how people negotiate changes in coupling in order to offer assistance, for example. By training a confederate, researchers can script the interactions of participants with the system and with their partner, ensuring each study session covers all of the tasks for which users will need software support.

References

[1] Beitz, M., Wiggins, A. (2020). CSCW 2020 – 23rd ACM Conference on Computer-Supported Cooperative Work and Social Computing. ACM Press. https://cscw.acm.org/2020/

[2] Gutwin, C., & Greenberg, S. (2000). The mechanics of collaboration: developing low-cost usability evaluation methods for shared workspaces. Proceedings IEEE 9th International Workshops on Enabling Technologies: Infrastructure for Collaborative Enterprises (WET ICE 2000), 98–103. https://doi.org/10.1109/ENABL.2000.883711

[3] Baker, K., Gutwin, C., & Greenberg, S. (2002). Empirical development of a heuristic evaluation methodology for shared workspace groupware. Proceedings of the 2002 ACM Conference on Computer Supported Cooperative Work – CSCW’02, 96–105. ACM Press. https://doi.org/10.1145/587078.587093