Remote Persons Are Closer Than They Appear: Home, Team and a Lockdown

Numerous researchers have recently investigated how mandatory WFH impacts individuals’ wellbeing and/or health [1, 24, 42, 49, 69, 72, 78, 86, 92]. A prominent concern is stress stemming from extensive ICT use, which leads to detachment and communication overload [60], and was coined “technostress” [8, 24, 42, 78]. Interestingly, while ICT use increased due to WFH, ICT use for personal purposes also increased, resulting in even more technology exhaustion [52], a decrease in wellbeing [78], and greater fatigue [8, 28]. As most communication abruptly switched to technologically-mediated forms, this fatigue was often expressed towards video conferencing (VC), e.g., as “VC Fatigue”, or “Zoom Fatigue” [8, 28, 44]. Infringements on the work-life boundary (cf. Section 2.1.2) further led to increased emotional exhaustion and decreased happiness [72]. Consequently, work results and productivity might be at risk [59]. Counter-measures taken by organizations to keep employees’ engagement high include increasing individual autonomy and involving employees in discussions about future corporate strategy [71].

Other research reports negative effects on mental health without looking closely at their causes, including: stress, dwindling motivation, depression and feelings of isolation, a lack of social interaction [69], and feelings of anxiety, anger, annoyance, tolerance, frustration, or uncertainty [86]. Fear of isolation, both socially and professionally, was mentioned as a factor hindering the voluntary adoption of WFH [56]. Besides psychological outcomes, physical health (e.g., sitting long hours, physical pain) was also considered [69, 72, 86]. E.g., a lack of ergonomic furniture in home-working environments increased “musculoskeletal disorders” [86], and changes in lifestyle (e.g., sleep, exercise, and nutrition)—both negative and positive—were described [72]. Further positive effects, such as individuals’ positive emotions, or increases in job satisfaction and organizational commitment levels [24] or better time management [69] were also reported. A systematic review [86] came to similar conclusions around motivation (also see [55]), enhanced perceived autonomy, and improved work-family-life balance.

Both voluntary and mandatory WFH presented challenges for boundary management and work-life balance [3, 6, 7, 12, 27, 42, 51, 72, 82, 86, 89, 92]. Foremost, abrupt transfer of workplaces to workers’ homes blurred work-non-work boundaries [6]. Pandemic-induced school closedowns further forced parents to put extra effort into care and/or education during traditional working hours[92]. “[W]omen, and mothers in particular” were most restricted in their boundary management autonomy [7]. Employees’ strategies to compensate for lost productivity included working in the evening or weekends, further blurring boundaries and contributing to negative psychological consequences [3, 12, 72, 82, 86] (also cf. wellbeing, Section 2.1.1). Besides personal circumstances, individual preferences for either segmentation or integration of work-non-work spheres also plays a role in how WFH is perceived [3, 6, 27]. While physical separation of workplace and home traditionally allowed for simple segmentation, this was largely lost through mandatory WFH. Besides well-known strategies (based on [53]), e.g., declaring certain rooms in the home a dedicated home office or keeping strict work hours, new approaches for re-establishing segmentation were identified [6], e.g.: “emulating office routine” (e.g., dressing the same way as when going to work), “purposefully disconnecting” (i.e., being intentionally unreachable after working hours), and “reducing work and home role overlap” (i.e., working when others are gone or asleep [3, 6]).

Beneficial outcomes of (usually voluntary) WFH have also been reported, e.g., connected to employees’ perceived autonomy, freedom, and gratitude for the chance to self-organize their work arrangements [7] (also cf. work conditions, Section 2.1.4). “Workplace flexibility, environmental conditions of home offices and organizational supports are positively associated with productivity, satisfaction with working from home and work-life balance” [89]. However, these beneficial outcomes might be lacking when WFH is mandatory [7].

Mandatory WFH and remote collaboration might also affect work performance. Related effects include productivity [1, 43, 54, 64, 86, 88, 92, 93], effectiveness [5, 44, 48, 55], efficiency [2, 8] and creativity [35, 58]. There seems to be a lot of evidence that (mandatory) WFH can lead to a decrease of productivity at work [43, 55, 64, 88], e.g., due to a low task-technology fit (TTF) [1] (i.e., how well technology is suited for the current tasks, considering abilities [37]), general technical issues [88], specific technical issues (e.g., caused by background noise [44]), motivational aspects [43] and current life events (e.g., health and childcare responsibilities [64]). Additionally, problems of traditional meetings such as the behaviour of participants (late arrival) or the organisation of meetings (too long, lack of agenda) also have a negative impact on effectiveness in WFH situations [44]. Further, an analysis of remote teams’ creativity suggests that they are “less likely to integrate existing knowledge and produce new, disruptive ideas” [58].

However, literature indicates that a more differentiated view is necessary. The efficiency of WFH seems to depend on the nature of tasks, e.g., researchers were less efficient in sharing thoughts, staying in touch with the team or collecting data, but more efficient in analyzing data, working on manuscripts, or reading literature [2]. Further, creativity of virtual teams can be improved by creative and adaptive virtual environments [35], and a greater TTF facilitates enhanced performance (and reduced feelings of loneliness, cf. wellbeing) [1]. One study additionally suggests differences between agile and traditional teams [54]: while the overall quality of team collaboration and productivity increased in agile teams, it decreased in traditional ones. There is also evidence for explicitly positive effects of WFH on productivity or effectiveness, e.g., due to increased inclusiveness and empathy for teammates, team bonding or flexibility [64], use of more knowledge and skills, and multi-tasking [55]. With a clearly defined and well-organised workspace that provides easy access to resources [93], remote work is seen as conducive to productivity [92]. Additionally, a pre-pandemic study on flexible work found that remote workers often did more when they worked from home (“intensivation of work” [48]). From companies’ perspective, the productivity and international competitiveness could increase, as remote working would make it easier for employees to work in different time zones and thus 24 hours per day [86].

Existing studies agree that the pandemic has drastically changed how and under what conditions organizations and employees work [2, 11, 16, 21, 57, 78, 79, 88, 92]. A generally positive development is greater flexibility [2, 7, 70, 92], often accompanied by increased autonomy and control over the organization of one’s work [2, 56, 71]. Work conditions in a collaborative setting are further strongly impacted by relationships and roles within a team. Relationships have been reported to be of utmost importance during mandatory WFH [11, 56, 93] e.g., “[t]he pre-pandemic relationships of colleagues [...] serve to ameliorate or exacerbate the challenges of such [social talk] conversations” [11]. Further, they are described as the “common glue to determine the opportunities and etiquette around social talk in remote work” [11]. Positive relationships with colleagues can be a “motivational factor for boosting productivity, and decreasing the probability of voluntary turnover” [56] and “socialization is one essential component of teamwork that virtual teams may often miss out on, which can result in a lack of team spirit and cohesion” [93]. This is not only true for colleagues that have not known each other prior to mandatory WFH; existing relationships might also be more difficult to retain [93]. Many negative effects of mandatory WFH are related to workload and distraction or interruptions. Most related literature agrees that the workload increased during mandatory WFH [3, 4, 21, 42, 88] (also supported by pre-pandemic WFH research [48]). Distractions and interruptions might be directly imposed by the home environment (e.g., by family members or pets, arrival of parcels, incoming phone calls [21, 56, 57]) but also might come from coworkers reaching out more often [21]. In addition to the work conditions, mandatory WFH has been shown to also affect individuals’ working habits, e.g., increasing the frequency and scope of their communications (an effect that might continue into the weeks after lockdown) [26]. Further, multitasking behavior seems to have changed during remote work and, according to [23], depends on meeting characteristics (e.g., size, length, time and type). While multi-tasking can help boost productivity, it also leads to loss of attention/engagement and greater mental fatigue, and can be seen as disrespectful [23].

Importantly, literature on work conditions suggests a discrepancy between voluntary flexible work [48, 77] and mandatory WFH [16, 70, 71]. A pre-pandemic study found that job satisfaction and organizational commitment were higher and stress levels slightly lower for flexible (remote) employees than for traditional ones, although the flexible workers also experienced “work intensification” [48] (cf. also performance, Section 2.1.3). Intra-pandemic research, on the other hand, reports stress as a negative result of mandatory WFH [3, 8, 21, 60, 78] (also cf. wellbeing, Section 2.1.1), with a (perceived) increase in autonomy or control on the positive side. However, a pre-pandemic study interestingly came to the conclusion that perceptions of autonomy had actually been undermined for remote working employees [77] through the need to be permanently available to colleagues (and probably also reinforced by a lack of social legitimacy of telework at that time). These discrepancies suggest that the evolvement of work conditions through the pandemic has had a lasting impact on the perception of WFH. While we expect a strong mitigation of the negative factors directly induced by the pandemic (e.g., stress through home schooling responsibilities) in a post-pandemic world, positive ones will probably remain (e.g., increased flexibility). This also specifically motivated us to identify work and collaboration practices worth being retained in the future (cf. Section 6).

Much research on collaborative interaction during WFH found changes in communication behavior [19, 22, 29, 61, 64, 70, 90, 93]. Naturally, face-to-face communication significantly decreased during WFH while VC and chat significantly increased [61, 64]. Meetings got shorter but were held more frequently [70] (a possible main contributor to meeting fatigue and technostress, see wellbeing, Section 2.1.1). At the same time, email, text messaging, and phone usage did not significantly change [61]. Studies indicate a general increase in collaboration [90] but also a strong shift towards asynchronous communication [19, 90] (or decentralized work [43]) and an overall decrease in communication, e.g., resulting from less initiative to proactively seek remote communication [19]. These observations are partly contradicted by research on preconditions for successful remote collaboration during mandatory WFH [21], based on Olson and Olson’s distance framework [67]. The related findings suggest that, first, the collaboration readiness of technology and involved persons has to be high, and actors need to achieve common ground, but second, while previously, loosely coupled work was deemed preferable for remote collaboration [67], other forms of coupling (e.g., tight coupling in synchronous meetings) have also proved beneficial recently [21]. We thus expect that, fostered by technological improvements, tightly coupled work will ultimately be co-favored.

Further research suggests that employees adopted new communication strategies, improved their communication skills, learned how to use new technologies, diversified their communication channels, and generally had to learn to collaborate more effectively [19]. They had to accept that they could not interact with others as easily as before [64] (e.g., with customers in on-site visits [19], or simply with colleagues [19] and classmates [22]), and had to “rely more on [their] independence” [19]. Informal communication was found to be beneficial [22], but it was often hard to establish because social events were perceived as causing role conflict, issues with boundary management, and meeting fatigue [61, 87]. Consequently, many teams were unsuccessful in replacing the spontaneity and informality of face-to-face meetings with computer-mediated communication. This led to detrimental effects on rapport [61] and individuals missing social interactions [64]. At the same time, fostering synchronous participation in meetings and a certain amount of accountability for individual contributions was identified as a motivating factor [29].

Regarding coordination and organization of work and knowledge in virtual teams, it was found that digital artefacts play a decisive role and serve as vehicles for the transfer (i.e., mere sharing of artefacts without changes), translation (i.e., adding to artefacts while preserving their original state), transformation (i.e., fundamentally changing artefacts), and holding (i.e., restricting others’ access to artefacts) of knowledge [30].

Pre-pandemic research on social space already revealed that norms associated with traditional workplaces, e.g., visibility, presence, trust, and availability, have to be rethought for teleworking [77]. During the pandemic, intra-organizational networks became more “static” and “siloed” (fewer new contacts were established, old ones were discarded, and less time was spent with newly acquired ones [90]). Fewer connections and interactions between groups were observed, and the awareness of what was going on in other groups decreased [90]. However, on the positive side, remote connections that already existed pre-pandemic were intensified, e.g., remote young scholars’ career opportunities improved through “expanded and enriched professional and social relationships” [29].

Also related to closeness, a public sentiment study identified a strong need for social interaction [92]. One way through which sociability can be established is small talk [11], which usually has a considerable share of overall speech (in adults up to one third) and can help to “transition[.] between roles and activities throughout [our] workday” [62], but is also important for collegiality and productivity [36]. During WFH, a substantial decrease in informal talk was reported [64] (cf. Section 2.2.1), resulting in “less socio-emotional interaction” and consequently less motivation [43]. To counteract this, researchers suggested a tool that allows online meeting participants to engage in small talk during the initial and concluding minutes of a scheduled meeting (and also throughout the meeting) [36].

Issues with connectedness were not only reported among coworkers, but also for friends and family. Social distancing led to fewer contacts and attempts to compensate for productivity losses (also see Section 2.1.2) limited quality time with loved ones.

Knowledge sharing or exchange are of utmost importance in collaborative work and might be decisive for the quality of outcomes. Both are affected by mandatory WFH [47, 66, 85], by motivational (e.g., information value, ease of use, organizational and technical support) as well as barrier factors (e.g., low content quality, loss of knowledge, i.e., intellectual property, emotional anxiety, lack of time, or insufficient organizational support) [47]. Similarly, implications of the pandemic for knowledge exchange might be positive (e.g., reduced barriers for participation, opportunities to assemble larger and more diverse groups) or negative (e.g., absence of face-to-face interactions might hinder building personal relationships and trust) [66]. Additionally, research on digital knowledge sharing during WFH indicates that internal and external digital knowledge sharing is a significant predictor for a team’s creative performance and that the latter can be increased by use of digital platforms [85]. These findings show that organizations can promote knowledge sharing in virtual teams via personnel development (e.g., improving user engagement by training and education), but also by optimizing the (virtual) collaboration environments (cf. Section 2.3).

While numerous tools and platforms support remote work and collaboration, having too many different ones to use and integrate hinders productive collaborative work [20, 78]. In summary, VC seems to be the most important supporting technology in virtual collaboration environments. VC is used across different work domains and tasks, and supports communication and collaboration. While existing research on technology use in the context of WFH agrees on the importance and value of VC, it also reveals numerous related challenges, e.g., general technical issues [44, 57] or (lack of) socializing [1, 11, 20, 36, 74, 79], small talk [36] and emotional expression [20], lack of perceived proximity and tele-presence (“distance effect” [50]), lack of surveillance [20], lack of spontaneity [20], blurring between on- and off-hours [8, 20], privacy issues [11, 44, 57, 79] (e.g., videos of home environments [11, 57] or while eating [44]), and lack of specific functionalities such as idea generation [20] or collaborative sketching support [32]. Additionally, research has addressed the phenomenon of VC fatigue [8, 28] (cf. wellbeing, Section 2.1.1). Such effects can be triggered or intensified by personal (e.g., socio-demographic variables, cognitive traits, a user’s social network), organizational (e.g., number and duration of VC sessions, timing, anticipated outcome of VC, activity during VC), technological (e.g., camera type, lighting conditions, video resolution, background noise) or environmental factors (e.g., work context at home, distractions) [28]. Further, rapid changes in VC systems or use complexity—or platforms forcing users to change their behavioral patterns—can induce VC fatigue [8]. Some challenges can be relatively easily overcome, e.g., by using external tools if functionality is missing, or background effects for privacy. The number and complexity of different technologies a user has to work with should be minimized [78], and a high TTF should be ensured [1, 32]. Other challenges are more difficult to tackle (e.g., lack of socialization, permanent availability). Here, some tailored solution approaches already exist, e.g., a “virtual commute” feature preventing a user from being contacted outside working hours [63], or a “conversational transition space” that enables social time [36]. However, many existing approaches are not yet available for the broad public.

As users’ physical work environments during mandatory WFH are usually their private homes, strong overlap exists with the boundary management aspect [3, 7, 51, 57] (cf. Section 2.1.2) and is kept brief here. Related research describes physical components contributing to a user’s work environment at home while at the same time functioning as “boundary objects” (“material ‘things’ [...] constitutive of interactions in a particular social domain [...] particularly visible when two communities of practice negotiate relations as their boundaries ” [51]), especially technology such as a laptop or phone (enabling WFH [51]) and (ergonomic) furniture such as a sofa, desk or chair (accommodating WFH [51, 56, 89]). During the pandemic, many remote workers’ physical environments functioned as office and home for several people (accompanied by pets) [57], which again relates to distraction and interruptions [21, 56, 57] (cf. Section 2.1.4). Establishing a suitable workplace at home is a crucial factor [56], e.g., comprising a dedicated room that allows for work-non-work separation [3, 56, 60, 89], and infrastructure such as a high-speed internet connection [21, 32, 42, 56, 89].

Each team was asked to enter a protocol of their meeting and task activities into a spreadsheet in a standardized format comprising start and end times, activity type (e.g., discussion or brainstorming) and description, people involved, tools used to communicate and to work on the task, and an optional comment. The teams recorded between 11 (Team A) and 262² (Team D) logs, which gave us an initial overview of how the teams worked, the frequency and duration of their meetings, and their typical division of labor for tasks [65]. Please note that self-reports may have some drawbacks in terms of data quality, e.g., due to differences in the resolution of data points (some teams reported exact minutes while others rounded to the nearest 5 or 10 minutes). We tried to minimize such issues via clarifying discussions in the group interviews. Furthermore, the additional effort of log keeping during meetings in order to reduce a-posteriori reconstruction was deemed acceptable by the participants for the limited study period (e.g., the additional effort for Team C was rather low, since employees were already contractually obliged to keep records of their working schedule and this could be used as a basis for the manageable amount of extra information required in the logbooks). For reporting purposes, we rely on data provided by participants and report it as is (e.g., in Table 2 and Figure 3 through Figure 6).

Two co-interviewers (identical for all teams) conducted semi-structured interviews, speaking to each team as soon as possible after completion of the observed task. We took inspiration from contextual inquiries [39], interviewing participants in their “natural environment”, i.e., the same collaboration environment the task-related meetings were held in, establishing a good rapport and equitable conversational situations, using one interviewer as the main interlocutor while the other was responsible for the protocol (but could intervene in the discussion if needed), and trying to refer to concrete past situations to avoid too general answers. Ten participants had attended an introductory course in CSCW and were familiar with related terms, concepts and frameworks, which allowed us to have deeper discussions about their perceptions of what was happening and made these situations more comparable to expert interviews. We included questions around each team’s general impression of the collaboration (what worked well, problems that occurred, tools that were used or needed, how work was organized), territoriality [75, 84] (whether and how different territories, i.e., personal, group and storage territory, were used), Buxton’s media spaces [17, 18, 81] (how important the person, task and reference space were and whether available tools were sufficient in this context), roles and responsibilities [13, 68] (whether certain roles established implicitly or explicitly), closeness [34, 43, 65] (whether the setting had an effect on the team’s closeness), effort (whether the setting affected the effort involved with the task), and retention of practices (whether the team’s experiences might lead to changes in future work and collaboration for comparable tasks). For participants not sufficiently familiar with the CSCW models and frameworks we addressed, we provided a brief introduction prior to their group interview. All interviews were held after several months of mandatory WFH (for the student teams towards the end of the semester) and we encouraged participants to bring in perspectives from outside the scope of the observed task to complement their accounts (e.g., by relating to experiences from other projects). This was particularly helpful in capturing some effects of mandatory WFH that may take longer to develop, such as challenges for well-being or boundary management.

After the group interviews, we asked participants to complete a questionnaire. This let respondents give an individual account of their perceptions and thus allowed for more critical answers that might have been hindered by the group setting or personal disclosure preferences. The questionnaire comprised self-developed items about participants’ demographic background and their perception of collaboration, including their overall satisfaction with the task result and positive or negative consequences they felt accrued to the distributed setting. Additionally, participants completed the 24-item “Questionnaire on Working in Teams” (German: “Fragebogen zur Arbeit im Team”; F-A-T [45, 46]), which measured four dimensions of the team’s quality of collaboration (see Section 5.2.4).

We received 13 meeting recordings with an overall duration of 14 hours (mean=1:04, SD=0:32, min=0:17, max=2:17). Each team provided us material lasting between 3:17 (Team A) and 3:45 (Team B) hours. Participants were not required to activate their camera—however, except for Team C, where less than one third of the participants per meeting turned on their camera, all others had high shares of activated cameras, usually above 75%. Additionally, Teams A and B provided us with 10-20 chat messages taken from the VC software. The meeting recordings were selectively skimmed to find situations that either supported or contradicted what was captured through the other sources (i.e., group interviews, questionnaires, and logbooks).

Concerning the work environment and hardware, all participants except one member of Team A (who used a desktop PC), worked with private or institution-provided laptops. Arnold (Team A) once used a graphics tablet to sketch while another team member watched via screenshare. Others stated that they desired some form of touch and pen input, especially for sketching or brainstorming. Furthermore, especially for reading notifications and quick texting, (private) smartphones were used. Infrastructure-related problems were reported by Bella and Brooke (who occasionally had bad network quality), complaining that facial expressions and gestures were difficult to perceive or completely absent if no video was used in a meeting. In Team C, according to Chris, the lack of an external webcam (due to supply bottlenecks) was problematic, since using the laptop-integrated camera in combination with an external monitor resulted in an inappropriate view angle. A similar situation arose due to the lack of headsets, where, according to Chris, “a feedback effect is the first moment you want a physical counterpart again”.

All participants worked from their homes during the time of the study and had a sufficiently quiet environment (living either alone or having a dedicated room) and furniture suitable for office work.

Learnings: First, our results confirm the rather obvious expectation that a stable and high quality internet connection is crucial. This comes as no surprise, but still raises the question: Why is it that in the early 2020s, we still have to discuss bandwidth? Current internet speeds are often faster than LAN speeds just two decades ago. Nevertheless, the comments of our participants show that considerations of how to handle such sub-optimal-connectivity situations—and the inclusion of disadvantaged groups who do not have access to stable internet service—are necessary. Even in a relatively wealthy country (all participants resided in a country within the top 20 in GDP ranking), broadband connections were not always available in participants’ homes in such a way that rich video conferencing tools ran smoothly. However, issues with connectivity were rather the exception than the rule. Further, we found that the home environment (furniture, etc.) is appropriate in a WFH setting if it is generally suitable for work. Additionally, we identified a challenge which probably affects all WFH settings: employees may often only have “standard hardware” (e.g., a laptop) available during compulsory WFH, which leads to problems that do not occur when working on-site. Adequate equipment with (external) cameras and headsets is a basic requirement for successful collaboration. For special tasks, specialised hardware (e.g., laptop with touch input) should also be provided, a requirement which often remains unmet by organisations due to financial limitations.

MS Teams was the primary software used by all teams for VC and screen sharing (see Table 3). Barbara stated it is often better to not have too many different tools to keep track of:

However, although MS Teams includes chat functionality, Team A and B made heavy use of WhatsApp groups that existed before the study. Brooke added that this allowed them to use their smartphones. We assume that WhatsApp had a higher visibility for the students as compared to MS Teams (i.e., the university’s default solution). A common task for Teams A to C was the creation of sketches and wireframes for low-level prototyping (using Balsamiq and Miro). Team D preferred E-mail over chat for asynchronous communication, probably due to the hierarchy in the team (the student at first hesitated to directly contact the professor via chat). Cloud services for sharing files were used by Teams A and B (Google Drive), and Team C (Nextcloud). Additionally, Team C used the version control system git to collaboratively work on source code. Finally, several tools for operational tasks were utilized (cf. Table 3).

Concerning the TTF, screen sharing in MS Teams was used by every team and was considered useful. Daisy explained that it is better than having a physical laptop positioned between student and supervisor: “Then it is not necessary that two people jostle for space but I have the entire screen for myself, which is a good thing!” Dora added that often a combination of viewing the shared screen and navigating the shared workspace itself (e.g., a Miro board) is important in order to have a separate view with individual control:

A negative aspect mentioned was confusion in finding channels in MS Teams: e.g., a search by Arnold for “Business Process Management” resulted in a list of 70 channels to sift through. In Team C, the organization’s policy to restrict creation of “teams” in MS Teams resulted in all communication being outsourced to chats. This led to lack of documentation of individual tasks, lack of categorisation, and thus more difficult retrieval of relevant information.

Concerning sketching and prototyping, different strategies were employed. Team A did not collaboratively sketch on a shared whiteboard but coordinated individual tasks verbally over a continual audio connection while working on individual and shared (e.g. Google Drive) documents in parallel. Despite previous experience, the initial enthusiasm about working together using Google Drive decreased as this turned into a “big mess”, e.g., through duplicate slides or documents. In Team B, everyone did individual scribble sketching before the team united the best ideas. While Beatrice used pen and paper for this, Bella and Brooke met in person³ to do the sketching in Balsamiq. In Team C, Chris and Craig (who worked from close offices before) identified a gap in tool support for drawing quick sketches. Before WFH, they usually drew during their shared in-office time by simply using pen and paper. Although collaborative sketching tools were known, they were not used due to poor integration into the available tool set and lack of pen input. Craig told us that he sometimes used tools like ScreenToGif to create short animations through screen recording (e.g., of solving processes) or Greenshot for enriched screenshots as a method of “collaborative debugging”. Chris added that these were nice tools to prepare material for meetings because they could record situations that might not be easily reproducible. Resulting short animations were posted in the MS Teams chat for others to immediately check or refer to during a later meeting.

Learnings: In accordance with previous work, our study shows that video conferencing tools are essential. Similarly, it shows that a well-organized work and collaboration environment is crucial and that the number of different tools that need to be used and coordinated should be minimised. Our study additionally reveals that users might be expected to use smart phone tools even though they are reluctant to install these on their private smart phones, leading either to the overriding of reasonable privacy concerns or to a gap between what team members “should” be using to collaborate, and what they are actually accessing. Furthermore, it shows that creative processes are oftentimes more difficult to perform remotely, partly due to the aforementioned lack of suitable hardware. Remote work might be easier to keep track of (e.g., through versioning) than copresent collaboration. However, the use of cloud services and file sharing systems requires compliance with individual rules and processes (e.g., who creates documents when and where, who is allowed to delete them), as well as organizational support (e.g., external access via VPN), in order to ensure an efficient exchange of documents.

According to the group interviews (and logbooks, cf. Figure 3), Team A might have been particularly prone to VC fatigue (see Sections 2.1.1 and 2.3.1), which could explain why they worked on the task as quickly as possible. However, they were not completely satisfied with their results (everyone gave the second best score on the questionnaire). Although according to Abigail, the task did not take more time “because you do not have to leave the house [which eliminates commuting]” she said that it was a big challenge to sit “all day in front of the laptop”, which in the long run is much more exhausting than “being at the [university] and going from room to room”. Thus, Team A tried to complete tasks during the available time slots (according to the university’s timetable), so that they did not have to invest even more time online. As negative consequences, all teams identified a lack of social contact, less interaction within and, importantly, also between groups, leading to perceived extra effort. Furthermore, working was perceived as a less creative process by student teams (Teams A & B; however, Team C also reported less contact with other groups), because they reported not hearing from other groups, whereas in an in-classroom setting they would have been able to hear what the other groups were discussing and thus engage in a multi-directional, inspiring group process. Finally, constantly working remotely and having no meetings in person quickly led to a feeling of missing social contact. Abigail, for example, stated:

According to Arnold, this was because students had to be attentive all the time, without small breaks including social chatter. He complained about sitting in front of his computer all day long and—even during breaks—not getting up and going anywhere (e.g., his balcony). Abigail agreed that, when she was “alone” (i.e., currently not in a meeting with others), she often remained seated. All in all, the group described the situation as challenging for personal wellbeing. Learnings: We conclude that remote work saves time, which is positive (efficiency), but at the same time this is partly achieved via reduced interaction with others, which is rather negative. Exclusively remote work and collaboration tends to be more exhausting and the lack of social contact causes discouragement and reduced creativity. We estimate that some of the negative effects were further exacerbated by the fact that next to WFH, regulations reducing private contacts were also in place at the same time.

Beatrice complained that it is more difficult to coordinate activities when there are no fixed time slots reserved for classes and that usually (while being in-class) the team would have simply done some work in longer breaks. She thought private appointments could have been made more flexibly throughout the day because there was no fixed time reserved for classes. However, in actual fact, it seemed more difficult to find time for an additional meeting where everyone was free than before WFH.

As a negative consequence, Cameron felt that the social dimension was set aside a bit. Chris bemoaned the lack of coffee breaks and lack of social chatter (which is relevant also for wellbeing), as well as an attitude of “working around the clock” that made boundary management between work and leisure time more difficult. Bella added that “in general, the weekend wasn’t quite the weekend it used to be. There is no real boundary anymore because you are always available and online.” Interestingly, no team reported problems related to privacy, e.g., concerns about sharing pictures of their home environment. We hypothesize that this can be attributed to the teammates knowing each other well.

Learnings: Organizations can support remote collaboration by reserving time slots (even if no physical rooms are needed). Furthermore, WFH easily leads to longer working hours and blurred boundaries between private and work space (actively restricting availability as suggested in Section 2.3.1 could help). When team members know each other well, they have fewer privacy concerns (if not, team building measures are advisable prior to intensive remote collaboration). In our results, there was a discrepancy between how student teams A & B tried to cope with infringements on their work-life boundaries in comparison to the paid-work employees in Team C. The former (particularly Team A) tried to get the work done as quickly as possible in order to stay within the schedule, while Team C (paid workers) reported more blurring of boundaries (i.e., separation vs. integration of boundaries). One explanation for this might be the different motivational structure and/or performance requirements in an educational versus professional work setting. Another aspect which we hypothesize to play a decisive role in this regard are participants’ current personal living circumstances. Some participants lived alone, some with their parents, some with a partner, some in shared flats, and some had childcare or other family responsibilities. For example, when a child’s immediate needs have to be dealt with during working hours, some work necessarily shifts into later portions of the day or night. All of these factors are known to contribute to personal preferences and strategies on how to manage boundaries (see Section 2.1.2).

We discussed the effort necessary to complete the current task in comparison to similar tasks completed on-site in the past. As positive consequences of collaborating remotely, Teams A and B agreed they worked more efficiently and effectively. These teams’ impression was that they did not increase the time spent (although their working habits changed, see “working around the clock”, Section 5.2.2), since they tried hard to accomplish the tasks within their schedules’ time slots to prevent additional meetings (Team A) or meetings on weekends (Team B), and because commuting was not necessary (Teams A and B, Daisy of Team D). In Team C, efficiency was also felt to have increased due to quick responses in the group chat. According to Dora, another advantage of the online tools (e.g., Miro), was that the latest version was always accessible, which prevented efficiency losses from working on outdated versions (a problem when reviewing the master’s thesis, which was sent in PDF form at certain milestones instead of being shared via collaborative editor). In line with what was reported in Sections 5.2.1 and 5.2.2, efficiency was probably increased at the cost of eliminating social chatter. For example Arnold stated:

Team A also stated that the loss of social contact makes working much more exhausting in the long run, which might then lead to a decrease in productivity. Thus, measuring the actual working hours invested, the teams might have been more efficient while the perceived effort seemed higher. Learnings: Our teams were more efficient and more effective with their tasks, working and collaborating remotely (partly due to lack of social chatter). Work was also perceived as more exhausting (cf. wellbeing). Paradoxically, the perceived effort was higher and the satisfaction partly lower (which confirms existing evidence) despite being actually more efficient (a novel finding which we term the “efficiency paradox”).

Regarding working habits, Teams A and C tried to split the tasks into subtasks and distribute them evenly to work on them autonomously. Exactly when, where, and how was up to the flexibility of each member. None of the teams reported that any members had failed to participate well in their tasks. As already mentioned in Section 3.1, roles and hierarchies in the teams were highly diverse. Nevertheless, the division of labor was mainly done from a functional perspective, i.e., people with the respective knowledge or preferences took on the respective tasks. Additionally, a certain level of expertise could also lead to greater implicit leading responsibility. Bella of Team B described how Beatrice had taken charge of the group because she was writing her bachelor thesis in that particular subject area and was therefore seen as more competent. However, according to Chris, the lack of clear formal roles (with the exception of being the product owner himself) in Team C resulted in confusion about whom to address because he lost track of responsibilities. He switched to addressing not a specific person but the “location” (i.e., the organizational unit which is further derived from the city the phyiscal campus is located in) in the collaboration environment. Interestingly, due to the ambiguity of functional responsibilities, the (former) physical locations were used as proxies for message addressing (although actually irrelevant in this remote-only setting) and communication was thus directed less to individual persons and more to the “location” based also on the functional and organizational division. In this team, therefore, there was a perception of two sub-teams of two people each. The communication impetus generally came more from a specific location side (represented by Chris and Craig) owing to asymmetric informational needs. In Team D, the hierarchy was stricter and there were clear roles, with Daisy as a student and Dora as a professor. Over time, the easy accessibility of the other location (Team C) or the professor (Team D) made it common practice for Teams C and D to quickly obtain help by convening online meetings, which again made work more flexible (“shorter paths”). To get an overall impression of the work conditions in the teams, we examined the average level of goal orientation (in German original: “Zielorientierung”), task achievement (“Aufgabenbewältigung”), team cohesion (“Zusammenhalt”), and mutual accountability (“Verantwortungsübernahme”) perceived by each team’s members, measured using the F-A-T questionnaire [46]. Further, we asked individuals to rate their overall satisfaction with the results of the collaborative task and averaged these results for each team (cf. Table 4).

Results show quite high ratings on all subscales, though Team A reported slightly lower values for the performance-related factors of the F-A-T (goal orientation and task achievement) than the other three teams. This finding was also echoed in Team A’s (relatively) lower overall satisfaction with their task result. Looking at the person-oriented factors (team cohesion and mutual accountability) reveals a slightly different pattern. All teams showed team cohesion near the maximum rating of 6, which may indicate that our data derive from an atypically harmonious sample of teams. In contrast, the two teams (A and B) consisting of a group of students show somewhat lower mutual accountability than either the professional research team or the advisor-student dyad—which may be explained by their shorter-term and less formal collaboration context. However, given the small sample size and the exploratory nature of the study, none of these descriptive differences were formally tested.

Learnings: All teams tried to organize themselves, distribute tasks and define responsibilities. While team members were flexible in organizing their tasks, they did not work fully autonomously due to the tight schedule (and pre-arranged time slots for meetings, which, however were perceived as positive, see Section 5.2.2). The remote setting did not lead to any new establishment of roles or hierarchy; existing ones remained but were softened (which is a novel finding that contributes to the existing still small body of evidence on roles and hierarchies). The teams identified themselves as teams, worked on a shared goal, had shared priorities and knew what to do. Potential problems (e.g., through uneven distribution of workload) were not explicitly addressed. The results of the questionnaire suggest that lack of mutual accountability was a slightly more relevant issue for student groups working on a class assignment than for teams working in longer-term, closer collaboration.

All teams stated that the use of communication tools had a very positive impact on cross-site collaboration. To gain further insights, we asked them about their impression of the existence and support of Buxton’s different communication channels (“media spaces”) [17, 18, 81]. The task space was predominant during collaboration and perceived as most important by all teams. Further, the teams agreed on the existence of the person space, typically implemented by a list of participants and their video pictures. Smaller video pictures during screen sharing were not perceived as insufficient, as they at least allowed for peripheral awareness. Often, the video transmission was even deactivated or participants ignored it, concentrating on the area of the screen where the team was currently working (cf. task space). Regarding the reference space, the teams’ impressions were divergent. In Team A, referencing basically worked well. Potential referencing problems were solved by sharing the screen, highlighting, or verbal communication (e.g., giving slide numbers). Team B reported that questions such as “is my mouse pointer visible” occurred often. Beatrice added that during the semester, students even learned new vocabulary in relation to online meetings “because you have to be much more precise so that others understand what you are thinking.” In Team C, where referencing in a 3D model was a crucial factor, significantly more problems occurred. Chris, for instance, stated:

Craig noted the problem that a person might suddenly start to scroll while another is still looking at something. This indicates a requirement for more explicit coordination when discussing an item. Craig also pointed to the request-control-feature in MS Teams, which the team did not use at all. In Team D, Dora explained that the reference space is mainly restricted to the area where one makes mouse gestures. She stated that she frequently used mouse gestures when working with Daisy, besides also resorting to them during her remote teaching activities. Dora also reported that one does this “almost automatically” when the screenshare is enabled. Moreover, Team D told us that the acoustic channel was usually combined for referencing (e.g., “Look at the concept XY!”), but according to Dora, if the jointly explored space is too large (e.g., in a big diagram), it can easily happen that not all participants look at the same area at the same time. Daisy suggested that it might be helpful if the other person got a mouse pointer via MS Teams as well, in order to be able to reference specific areas of the shared space independently (or do further operations depending on permissions) similar to Doug Engelbart’s “The Mother of All Demos”⁴. Both participants in Team D agreed that the reference space is implemented poorly or unnaturally, and Daisy added that seeing gestures depends a lot on camera angle and placement.

Learnings: Communication is crucial for team collaboration across different task domains. Complementing existing research, we utilized Buxton’s media spaces as both analytical lens as well as research subject. Our related findings can be summarized as follows: For teams that know each other well and basically agree on what they are doing, the person space seems to be least important (the opposite is likely true for teams that do not know each other beforehand). The task space is most important across all task domains, while the importance of and requirements for the reference space vary. Referencing was least well supported by the tools available to participants in our study.

As already indicated by related studies summarized in Section 2.2.2, closeness can be viewed from different perspectives that focus either on social and personal closeness, or on professional closeness. Social and personal closeness overlaps with wellbeing (especially in regard to social interaction), as described in Section 5.2.1. Participants largely agreed that private social contact was reduced during the mandatory WFH period, posing challenges for the maintenance of personal social closeness. However, in this section, our focus lies on professional closeness, i.e., task-related closeness of team members based on their work-related interactions. We asked our teams about connectedness, i.e., how closely coupled they worked, whether there was a lot of independent work, and how often they met (this overlaps with working habits, see Section 5.2.4). We were also interested in how they felt they would have collaborated differently on-site. In Team C (and similarly for Teams A and B), the cooperation across the different locations seemed closer and more intensive than before. There were more frequent meetings (“more often it happened that we discussed five times a day”), but this was not perceived negatively, as the video calls took place in a well-considered and targeted manner, and helped to solve problems efficiently. Chris added that the team would not have advanced so far in the task without this form of cooperation and technological support (cf. also performance). At the same time, the collaboration among previously co-located colleagues became less tightly-coupled, including reduced meeting frequency. In Team D, the cooperation (due to different tasks and roles) was mostly loose. The joint meetings focused on strategic issues (e.g., how to structure expert interviews). Both members agreed that the frequency and length of the meetings changed during the course of the task. Daisy stated that at first she hesitated in asking for appointments. This inhibition quickly decreased, as meetings took very little time and effort, and she realised that this was not a problem (also see the findings related to hierarchy and roles in Section 5.2.4). Daisy mentioned traveling distance as the main factor that led to more frequent, shorter meetings, while she said that otherwise she would prefer fewer, longer meetings. Short questions were usually dealt with via email. Overall, this led to a feeling of increased connectedness and closer collaboration, also emphasized by Dora:

Related to territoriality [75, 84], we asked what participants thought could be regarded as personal, group, or storage territories during their collaboration, and whether any of these were missing or poorly supported. Teams A and B built their collaboration around a shared storage territory on Google Drive and worked together in the sense of a group territory in Google Docs or Google Slides (cf. Table 3). Team A illustrated the trend away from personal territories towards group territories, e.g., via Abigail’s statement:

However, teams reported frequent alternation between tightly (e.g., creating a presentation together) and loosely coupled tasks (e.g., research activities). Bella used a personal area by drawing sketches on her Apple iPad that she initially saved locally and only shared with others once she was finished. In Team C, there was a rigid separation between the private area (local repository) and the group area (code pushed to the server) (through the use of git). Nextcloud, which was provided by the organization, served as the storage area (in addition to the git server). Unlike Google Drive, however, Nextcloud was not considered a group area because “changes are not so easily noticed”. The heavily-used MS Teams group chat was seen as a kind of shared group area. Yet, according to Chris, this was complicated and a separate team would have been better suited as group area (e.g., tasks could have been managed better there).

In Team D, both members had their own local and personal territories and created group territories mainly during meetings, e.g., everything Daisy did concerning writing the thesis, evaluating data etc., was in her personal space. When artefacts were viewed via screensharing, there was a transition from the personal to a shared space, thus more similar to group territory; however, while sharing the screen, only one party had control. Daisy explained that they used verbal statements when one of them wanted to manipulate a shared UI “Uhm, could you please go back to...” (which also relates to the reference space, see Section 5.3.1). Matters around presence and social space were not addressed in the context of closeness by our teams. We attribute the former to the fact that teams were small (perception of members’ presence is relevant especially in large groups with loose collaboration), members knew each other well, and everyone was required to contribute.

Learnings: Some of our findings on closeness are in line with those of others, including non-work settings or pre-COVID-19 evidence (e.g., [25, 73]). For example, fewer connections and interactions between groups were observed, and the awareness of what was going on in other groups decreased. However, we additionally observed that the connectedness within a team paradoxically seems to have increased over distance, relatively independently of the nature of the task. The latter can, however, be a decisive factor for the importance of different territories, which has not been considered in detail in related work so far: the more interaction a task requires and the closer a collaboration gets, the more important group territory support becomes. Support for presence is not as important for small groups where roles and responsibilities are clearly defined. The increased professional closeness between team members was particularly strongly apparent in the team with the most distinct roles and the clearest hierarchy, but generally evident in all of our teams.

After switching to WFH, some common behaviors and strategies for sharing and exchanging knowledge naturally had to be changed. As outlined in Section 2.2.3, virtual environments, administrative aspects, and personal factors (e.g., motivation) are decisive for a successful transfer to WFH. With suitable environments in place, some knowledge-intensive tasks were even easier to fulfill than before, e.g., jointly reviewing documents (Team D reported advantages of looking at a remote screenshare instead of competing for the limited space around a computer display on-site). Further, equitable access to shared resources was conducive to a free flow of knowledge. According to Team C, how well knowledge could flow was highly dependent on tool support. As mentioned above, Craig sometimes used a specialized tool (ScreenToGif), but the created artefacts—though helpful in themselves—did not integrate well into the shared knowledge base. Overall, Team C reported that they struggled to achieve a persistent knowledge base and many resources of mutual interest were simply posted to a shared MS Teams chat. Finding, structuring, and reacting to content was error-prone there; a Wiki, project management software, or similar tool might have been a better place to store the knowledge items created (including bug reports and short video figures). Craig argued:

Chris identified a further disadvantage of using MS Teams chats for this purpose:

Cameron explained that in a MS Teams team they would have had text threads instead of just individual messages that were chronologically ordered but not actually categorized. Finally, Teams A and B reported that a separation between the teams was introduced through WFH which was not there in prior classroom settings. Without peripheral attention to what other teams were achieving (sometimes referred to as ‘porosity’), creativity was curbed, as knowledge stayed strictly within the teams, creating information silos (cf. discussions on wellbeing and interaction).

Learnings: Our study confirmed that knowledge sharing is important for the quality of outcomes in collaborative work. Additionally, our study showed that equitable access to digitally shared artefacts can create a useful stage for knowledge sharing and exchange. While a small number of tools capable of supporting these activities is preferable, it should be possible to integrate specialized tools for minor miscellaneous tasks. Individuals should be made aware from the start about certain tools’ suitability as a maintainable, permanent knowledge base.