The Role of Offline Places for Communication and Social Interaction in Online and Virtual Spaces in the Multinational Workplace

To answer this research question, the present study uses findings from a qualitative and longitudinal study (2010–2013) of a multinational organisation. The theory of structuration (Giddens 1979; 1984) and Giddens's (1990; 1991) later work on modern society will be consulted. Insights from this study will be useful for organisational research on virtual work, outsourcing, cross-cultural communications and collaboration.

An anthropologist's most important research tool is ethnographic field work (Eriksen 2001). Ethnography is the close study of groups’ and peoples’ everyday lives in their social settings (Emerson et al. 2011), and typically involves the development of close connections between the ethnographer and the subject or situations being studied (Hammersley & Atkinson 1995). Ethnographic field studies were conducted in offices in Norway and Morocco (3 weeks each in June and July 2011), as well as participatory observations in offices in Norway (1.5 days in May 2011), Denmark (3 days in May 2011) and the UK (1 day in September 2011). The field studies in Norway and Morocco were repeated one year later (2012); this return was perceived by employees as a respectful gesture, and information that was deemed too sensitive in 2011 moved to the forefront in 2012 (e.g. conflict in work settings, discrimination, power, management styles).

Although TBC was a partner in a larger research project (of which the present study was a part), the researcher of this study still had to gain acceptance by TBC employees to be let into their social groups. The entity's community manager posted on the enterprise platform about the researcher being present at the workplace as part of a study of the social enterprise platform. In contextualised settings, the researcher worked on a laptop in the company of others in the shared main office space. By adopting a vague persona and following the participants’ dress code, the researcher neither symbolised management nor appeared as an outsider. In time, TBC professionals seemed to forget the researcher's role, and the consultants in the field responded to the researcher as a trusted colleague and friend with whom they could discuss work. The researcher arrived at the workplace early and went home late. Coffee and lunch breaks were a particularly important space for informal conversations and for getting TBC professionals to share their insights and thoughts. Several informal meetings and talks with consultants, managers and middle managers were held during the workday to obtain the best understanding possible. These informal conversations were not recorded (as the 27 interviews to be described below were), but were noted in the researcher's field diary. In the field, the researcher took on a student role in addition to using an existing background as a consultant – a field access strategy recommended by de Jong, Kamsteeg & Ybema (2013).

Approaching the field of study through an anthropological lens brought a valuable, in-depth understanding of the object being studied. With this approach, the researcher could observe how employees worked, who they interacted with at work, who they spoke with during lunch, met with at the coffee machine, sat next to during working hours, and more.

Twenty-seven open-ended, in-depth interviews were conducted with TBC employees from six offices in four countries (Norway, Denmark, the UK and Morocco) from May to September 2011. All but two of the interviews took place in the participants’ contextual settings, and were carried out by two researchers (see Table 1). In June and July of 2012, four participants from Norway and four from Morocco were interviewed again. All 27+8 interviews were recorded, transcribed verbatim and anonymised. Each interview lasted approximately one hour, and the participants received a gift card as an incentive. The sample was made up of 9 women and 18 men aged 22–59 years. The gender difference was representative in that the majority of ICT consultants are men. All the main services offered by TBC to clients were covered in the interviews with the following participants:

Social network data concerning whom the 27 participants approached for help and advice during the working day, and which co-workers approached them with work-related problems, were gathered at the end of each interview. Participants received and completed “colleague maps” with the photographs and names of employees in their departments. Blank spaces were left open so the participant could add any individuals not listed (e.g. from other entities). This created a roster network design (Wasserman & Faust 1994). This social network analysis (SNA) data was further coded in UCInet. The 27 participants listed in total 391 TBC colleagues as important to them. However, when asked to mark colleagues on the map who were important to their work, several participants commented that such individuals were not included in the map because they were clients or contract consultants rather than employees of TBC.

It was important to recruit participants from different departments and countries. Participants in this study were recruited using the snowball method. Starting at one entity (Oslo 1), the first participant was selected from the middle of the employee list. However, because the participants seldom listed colleagues working at entities other than their own, the snowballing process had to start over again several times. A limitation of the snowball methodology is its risk of leaving key persons out of the sample (Hanneman & Riddle 2005). However, because the snowballing had to start over again several times, and because it used a mixed methodology and included various entities and service areas in the sample, the main tendencies are believed to have been captured.

An exploratory case study, inductive and descriptive in its form, was chosen for this study. The aim of this study is analytical, rather than statistical generalization, a distinction suggested by Yin (2012). The analysis was inductive and holistic, starting when the researcher entered the physical settings to begin the data-gathering process. The analysis was characterised by moving in and out and back and forth between field notes and interview transcripts, analysis and theory in a holistic circle (Hastrup 2010; Wadel et al. 2014).

The data analysis process involved five phases. First, data was collected. This involved moving back and forth between transcripts and field notes, adding new dimensions along the way and sometimes contacting the participant at a later stage with follow-up questions. Second, the recordings were listened to repeatedly and the field diary read several times so as to identify themes, take notes and develop analytic categories and constructs. Third, the interviews were reviewed several times to identify overall themes and findings. These were then coded and analysed using NVivo 8 software to look for key patterns, similarities and differences. More than 240 nodes were coded and sorted into main topics (e.g. power, work model, pay model, user patterns, language, workplace, problem solving). However, this coding in a standardised computer system did not allow for a “dynamic” analysis, so it was used instead as an “assistant tool” to the field notes. The main findings were then discussed in depth by two researchers. Fourth, the researcher categorised the 27 participants in an Excel sheet, where each one was coded according to the main categories identified during the analysis (e.g. which entity they perceived as most similar or different to themselves; number of speak-to colleagues). Finally, the findings were presented to the TBC professionals involved as insiders (Brannen & Thomas 2010) to validate the findings, which they did.

The social network data regarding who TBC professionals address when seeking work help – and who seeks them – suggests there is little cross-unit collaboration across the 20+ entities in TBC, even when located within the same country or city. The analysis reveals that TBC consultants perform work in six different contexts: (1) from the main office (a category of workers labelled “In-housers” and “Overlappers” to denote the group of consultants in slack (between projects), (2) with clients (“Out-housers”), (3) with clients’ clients (“Fixed-site Teleworkers”), (4) from home (“Teleworkers”), (5) from a different country (“Distant Workers”, solving tasks that have been outsourced to them), and (6) on the go, unrestricted to a place and time (“Nomads”).

In TBC's work landscape, most employees across all entities work in the same physical place day after day, close to colleagues relevant to what they are working on. When TBC professionals need assistance in their work, they approach colleagues sitting physically close to them. As explained by one consultant from the UK, “it tends to be a locational thing. The people who literally sit near me.”. In the organisational work context, TBC employees choose communication spaces that are closed, small (e.g. the coffee machine, during lunch breaks) and personal (direct conversations, telephone or email), and that provide a good overview of who is part of the conversation and who is not. This tendency is expanded to closed groups in the enterprise platform, which provides a good overview of who the other members are. With smaller, semi-private online spaces comes a trust-building overview of the group members who are watching or participating in the conversations. The open nature of the platform, though, does not provide insights of who the others are, who can be trusted, or the boundaries in the conversation. To speak publicly in front of everyone in TBC's enterprise platform differs from speaking to a smaller group, in line with the theories of Giddens (1984) and Goffman (1959). Moreover, how one communicates in groups in the online platform at work depends on the audience, one consultant explains:

It depends on the group. We have a closed group for us here at the office, and we have a group for those working with [given topic] in Europe. They are very different settings. The office group has a funny name, and it is something totally different when I’m going to speak with people I sort of do not know at all. One puts on a seriousness filter in some of the online spaces.

Observations during field work in Norway revealed that in-house consultants were located close to other relevant colleagues, with whom they interchangeably communicated about work (“Did you get the report I just sent you?”) and life (“How did things go at the car repair yesterday/at the farewell party at school?”). In Denmark, a group of consultants were seated close together and discussed work loudly. Despite a smaller office space in London, employees were still able to roll their chairs to nearby colleagues to look at their computer screens or answer or ask questions. In the offices in Morocco, loud discussions on work and life matters among employees sitting close to each other took place. Distant workers sat together with other consultants to carry out work that had been outsourced to them, despite the fact that most of their tasks were completed alone or in the company of another outsourcer. Sitting close to important colleagues was observed among out-housers. These patterns demonstrate that sitting physically close to colleagues helps with work completion, as this consultant from Norway explained when asked where he seeks the information he needs to work:

I get much of what I need through the people I work with because this is knowledge the client and their two suppliers that we work with have, and now [name of another consultancy firm] are also working on it.

TBC professionals mainly consulted individuals they already knew and worked with. In this way, physical placement significantly influenced whom the employees approached when they needed work-related assistance. When sitting close, the employee also controlled conversation boundaries and maintained a good overview of the conversation partners (Giddens 1984; Goffman 1959). If close colleagues in the shared working context could not provide the needed assistance, the inquiring employee often asked these same colleagues for a personal recommendation instead of reaching out on TBC's social enterprise platform where over 5000 colleagues might help. When asked why they do not do so, several explain that they feel that asking out in the open where one does not know the audience is unpleasant, and that they put on a “seriousness filter” depending on the context (e.g. online group with management as members discussing TBC-related matters, or social groups planning the summer party).

Interestingly, many employees called, emailed or otherwise addressed non-TBC professionals with whom they are connected on external social sites (such as Facebook or LinkedIn) when in need of work-related assistance. These social sites allow employees to reach out to their contacts outside the company, as their contacts and friends often work on similar tasks and domains. This also follows the pattern of connecting only with those the employee already knows and works closely with (Pettersen 2016; Steinfield et al. 2009); employees communicate with those they already know rather than approaching new TBC colleagues.

Interestingly, the analysis of the social enterprise platform reveals a consistent pattern that communication and social interaction practices in the employees’ organisational context is expanded, or “lifted in” to use Giddens’ terminology, to the online platform. Those with whom the TBC professionals work with on a regular basis are mainly the ones they follow in the online enterprise space (the following feature), and other members in the platform's group are those whom the employee works with or shares other work characteristics with (e.g. belongs to the same department). When a consultant in his forties from Norway was asked who he has added as contacts in the enterprise platform, he replies; “Oh, almost everyone, I think. At least the Norwegian colleagues, and at an international level are those I have some kind of relation with”. Having a joint history plays a key role in developing trust (Giddens 1984), and shows the importance of building a social relationship the conventional way before expanding it in online settings, similar to Nandhakumar's (2002) study of virtual teamwork.

Moreover, the enterprise platform makes it possible to join groups regardless of geographical and organisational belonging in TBC, and strangers entering a group originally created for a specific local use pose a new situation for the group's members to interpret. A consultant shows me on her computer a group she created in the enterprise platform:

I have created a group so we can have a place to have documents. And everyone in the team should be here … [She pauses and studies closely the group members.] Well, there are more members here than the team. Him, for example, I have no idea of who he is or where he is from. [She clicks on him and reads out loud his name] from [work topic name]. I have no idea of who this is. Here it says he is from [another entity]. Oh dear, how exciting. But I created it originally because I thought it should be our collaborative space.

However, groups typically are created to serve a team's offline needs, and when others who are not part of the group offline enter it, it presents employees with a new situation different from how they work in practice. As Goffman (1959) would have put it: the front-stage suddenly occurs in the employee's more trustful, back-stage settings.

With the exception of the nomad group, work at TBC was organised in a certain work time or workday rhythm across all contexts, beginning in the morning and ending in the afternoon. Daily morning greeting rituals took place to symbolise the beginning of a new workday. Female employees in Rabat greeted each other verbally while kissing each other's cheeks. In Oslo, employees welcomed their colleagues with informal verbal gestures, often around the coffee machine, to have a social chat before they begin the workday. When leaving the workplace in the afternoon, personal symbols were often left at employee desks if the office had a landscape model (e.g. a yellow note or a family picture). Interestingly, although the teleworkers worked from their homes, they also organised their work as a typical workday, beginning in the morning and ending in the afternoon. One teleworker from the UK described a typical workday at home as follows:

I tend to start work at about half past seven in the morning. I’ll read my emails, work out if there's anything urgent that needs doing immediately. And I manage most of my to-do list in a mixture of Outlook and a spread sheet. So I sort of consult those and decide what I need to do urgently. I tend to be most productive from about half past seven in the morning up to lunch time. So I try to get the most important things that require a lot of thinking, and then I can focus on maybe other things in the afternoon. I work until sometimes seven in the evening.

The teleworkers also structured their work time in meaningful sequences, despite being physically separated from their teams. This is very different from the nomad worker, who typically picked up and continued work tasks (e.g. emails, completing documents, work tasks related to daily work and clock time) as needed. Nomad work was also done on days not typically regarded as work days, such as weekends and holidays. Nomad work was observed be done on the go, in cafés, from airports, from home, while travelling, etc., lending a temporal quality to the work.

Working with nearby colleagues makes fast replies on work matters in real time, in co-presence, very easy. Conversations were observed in the employees’ respective native languages (Arabic, Danish, etc.). However, working closely did not only influence getting work done; it also had an important social dimension. Social interaction was found to be a core dimension of TBC professionals’ working practices, where conversations among employees during the workday influenced their work, similar to Orr's (1996) study of copy technicians. During the workday, social and work aspects overlapped in conversations among employees. Social interaction mainly manifested through language and communication (Gumperz & Cook-Gumperz 2008), and a key factor in interpersonal communication was the social aspect (Brinkley 2009; Goffman 1959). Working together in the same geographical place, employees were socially integrated in their shared social working structures, and thus learned the implicit rules and practices of work tasks and social working relationships. When first established, social relationships among TBC employees can exist out of time and space as structuration processes due to mechanical integration (Giddens 1984). This illustrates how geographically removed teleworkers were still able to work in co-presence with their teams.

Every day at 9:30, we have a call for half an hour where we catch up and, you know, anyone that's got anything. Often, someone needs to discuss something with a particular person, and if they can’t resolve it in a couple of minutes, then we’ll take it and have a call afterwards. We have a continuous Skype joint chat running, so if anyone wants to ask anything, we’re always on Skype. But we keep this daily chat as a separate area so we can effectively say what we want. And also, it does get some social chat, non-work chat, on it as well

The quoted consultant illustrated the social aspect of work communication. TBC teleworkers communicated online throughout the day in co-presence. However, it is important to note that these relationships had been established previously when they workers were in the same physical workplace. These work relationships were deemed important by TBC employees; a consultant from Morocco described how she cared about her colleagues at the TBC office:

I like my work very much. It's a lot of work, but I really like it. To work with these colleagues is great – everyone is young and has the same mentality! We send a lot of emails to know the next steps to work on, or if there is a problem or something. We exchange emails during the weekends, and we phone daily. […] It is human to help someone that asks for something, and if they don’t help you, the day they will ask for help, it is normal that you say, “Oh, I don’t know” [she laughs]. It is human nature. It will help me to help you.

The consultant continued on to explain that her team is like a family that trusts each other. She also described assisting others in need of help beyond her team members with reciprocal gains. Trust, reciprocity and shared practices were routinely confirmed on a daily basis in TBC by employees working closely together, and described to influence social integration among employees in their office social structures. Reciprocity could also be a key reason for why employees were able to do nomad work.

A joint history facilitates the development of trust (Giddens 1984), exemplifying the importance of building conventional social relationships before expanding them to online settings, similar to what was explained in Nandhakumar's (2002) study of virtual teamwork. Being located together in a shared place – at least until employees get to know each other's personalities, social structures and practices – seems to be essential for work. As such, co-presence is crucial to the development of personal trust. When the employee becomes socially integrated, this state can be lifted out of time and space via mechanical integration (Giddens 1984). Their social involvement can then exist in a virtual space even when the employees move to different geographical places. The distant workers observed in this study, however, were not only geographically far from the work problem they were outsourced to solve (e.g. fix a bug), they were also socially far from their colleagues and the social structures in which the outsourced work problem originated. Unlike the teleworkers who worked from home, the distant workers seldom met those for whom they were fixing problems. Keeping in mind that social interaction is a core dimension in much knowledge work (Ettlinger 2003; Trygg 2014), the distant workers lacked important dimensions when solving work tasks, such as knowing their colleagues and other implicit aspects related to social structure. Simply put, the distant workers had no common ground or co-presence with the people they were helping.

A young consultant in his twenties from Morocco explained that the TBC entities he assisted had specialised personnel on the specific topic he supported. The IT problems the distant workers worked on were communicated from the client or via a TBC entity located in Europe via a service desk platform.

They are providing us documentation, but they are the ones that develop the application, and they will therefore know more than we do, so we need additional information because they have information we don’t have. They are very knowledgeable, and they are always providing additional information that we could not know of since they are the ones that develop the applications. On the surface level, we know the application, but then there is a deeper level in which we need some documentation so we can understand the problem.

The distant worker received only a brief or “shallow” description of the work problem he was supposed to solve, communicated in written text via the service desk tool. He never learned the implicit rules or practices at play in the social structure from which the work problem came; he did not know the key individuals in that context, nor what their jobs were (Nicolini 2011; Orlikowski 2002). He also lacked the social relationships that would allow him to discuss and better solve the work problem. The interviewed TBC professionals working with outsourcing tasks had no social relationships with the individuals involved in their work problems and were never part of their social structure. Some expressed that they wished they knew the people they assist more personally. A distant worker in his twenties from Morocco explained as follows:

It would be interesting to collaborate together because we don’t really interact. By collaboration, I mean collaboration as get there with the team, have trainings together, to be more human in the collaboration. Technology can do the job to a certain point. We can do it through Skype, the social enterprise platform, or for any means of communication. We can communicate about everything, but we do not know the person on the other side of the computers, so I don’t know his profile, his personality, and how they interpret what I am saying.

It is important to meet physically, he explained, because “you get much more comfortable when you know the person you are communicating with”. Geographical location provides a space where employees learn to know each other's practices and foster trustful and predictable social relationships. As such, geographical separation corresponds to the separation of social structures. The interviewed distant workers were not “lifted in” to the client's social structure, to use Giddens's (1990) terminology; instead, the problems they were tasked to solve were “lifted out” from the contexts in which they were established. This was not without difficulties; misunderstandings and conflicts occurred easily due to the absence of a common ground or co-presence.