Conceptualizing social well-being in activity-based offices

Purpose –The purpose of this article is to aid conceptualization of social well-being at work by identifying its components in a contemporary office context, so adequate measures can be developed to monitor social well-being and to assess the impact of interventions in the workplace. Design/methodology/approach – This study used existing interview data from recent post-occupancy evaluations of two large activity-based flexible offices in the Dutch public sector. Data-driven concept mapping of 182 different employees’ statements on social aspects of well-being was used to find communalities in their perceptions. Findings – From the data 14 key concepts emerged referring to employees’ social needs, reactions to (anti-) social behaviour of others and perceived social affordances of the work environment. Contrary to established theory, social well-being appeared to be a context-bound phenomenon, including components of both short-term hedonic and long-term eudaimonic well-being. Research limitations/implications – The findings serve as an inductive source for the further development of adequate measures of social well-being at work. Limitations concern the specific (cultural) setting of the cases and the use of existing data. Practical implications – Preliminary suggestions for fostering social well-being include change management, participatory design, being alert of the identified risks of activity-based offices and supporting privacy regulation, identity marking and a sense of community, as well as a diversity of informal face-to-face interactions balanced with quiet spaces. Originality/value – This article contributes to the conceptualization of social well-being in contemporary offices by discussing established social well-being theory and analysing real-world data, using a method novel to management research.

. For human beings, interactions and relationships with people around them are crucial to their health and well-being (Baumeister and Leary, 1995;Diener and Seligman, 2004;Ryff and Keyes, 1995). This also applies to the work environment (Kahn, 2007). As Rath and Harter (2010) state: "We are social beings, and our need to be connected to others does not disappear when we enter the office." The organization benefits as well, because face-to-face contact at work boosts mood and productivity (Pinker, 2014), and high-quality connections are a source of organizational improvements (Dutton and Heaphy, 2003). Additionally, friendship (Lopes Morrison, 2005) and perceived inclusion (Chen and Tang, 2018) at work increase engagement. For many, the current COVID-19 pandemic requiring long periods of social isolation and working from home has increased awareness of their social needs in the workplace: they miss meeting co-workers in the office (Gensler Research Institute, 2020).
Fostering employee well-being requires proper measurements covering this important social dimension, to determine the need for and impact of interventions and to monitor well-being over time. Still, conceptualization and measurement of social well-being at work are in its infancy (Fisher, 2014) compared to physical and mental well-being. Organizational literature covers many social aspects of employee well-being, such as professional isolation (Golden et al., 2008), loneliness in the workplace (Wright et al., 2006), incivility (Schilpzand et al., 2016) and conflict (Ayoko et al., 2003), but there have been few attempts to conceptualize the social dimension as a whole. This holistic view is important because a positive experience of one aspect of social well-being could be undermined by a negative experience on other aspects. Also, relatively few connections have been made with the office environment (Appel-Meulenbroek et al., 2018). In organizational research, the role of the physical environment has long been outside the usual scope (Ashkanasy et al., 2014;Khazanchi et al., 2018), while it is clear that the social and physical work environments are intertwined (e.g. Ayoko and Ashkanasy, 2020;Gifford, 2014, p. 342;Sander et al., 2019). Yet, in research on healthy offices, social aspects are underexposed (Bennett, 2018;Colenberg et al., 2020;Forooraghi et al., 2020). As noted recently, there is a need for clearer definition of well-being constructs in building research (Hanc et al., 2019).
Developing a valid multiple-indicator measurement requires clear conceptual grounding (DeVellis, 1991;Rosas and Camphausen, 2007). The purpose of this article is, therefore, to aid scale development by identifying components of social well-being in office environments. First, current social well-being theory and related characteristics of activity-based offices, given the current trend in office environments, are analysed. Second, social well-being components are identified as they have emerged from case-study data. The results are discussed in the light of scale development and workplace management.

Social well-being theory Established theory
The most prominent conceptualization of social well-being is by Keyes (1998), published half a century after the 1948 World Health Organization's acknowledgement of a social dimension to well-being, next to physical and mental well-being (WHO, 2006). Individual social well-beingas opposite to societal or sociological well-beingtraditionally has been measured through satisfaction with social support and adjustment to the social environment, which is closely related to mental health (Larson, 1993). The positive mental health approach of functioning well in social life has been the premise of Keyes' theory. He deducted five dimensions of social well-being from philosophy, social psychological theory and cultural analysis, reflecting the individual's (1) integration in a community, feeling part of a group, (2) acceptance of diverse characters and qualities of other people, feeling comfortable with others, (3) perceived contribution to the community, feeling a valued group member, (4) actualization or belief in the JMP 36,4 community's evolution, feeling hopeful about its progress and (5) perceived coherence of their social world, feeling they can understand what is happening around them. In psychological research, social well-being is positioned as the outer-directed aspect of well-being that complements hedonic inner pleasure and eudaimonic inner growth (Fisher, 2014;Gallagher et al., 2009). Keyes levels social well-being with long-term and competence-oriented eudaimonic well-being (Ryff and Keyes, 1995;Ryff and Singer, 2008), together labelled flourishing (Keyes, 2002), and separates it from transient moods of hedonic well-being or happiness (Diener and Ryan, 2008). This separation characterizes two main streams of psychological well-being research, although it is still under debate if these also reflect two types of well-being (Biswas-Diener et al., 2009;Gallagher et al., 2009;Magyar and Keyes, 2019).

Application to work environments
Recent measurements of the social dimension of employee well-being (Czerw, 2019;Kazemi, 2017;Rautenbach, 2015;Redelinghuys et al., 2019) rely on Keyes' (1998) conceptualization. However, there is reason to believe that the phenomenon might be organized in a different way, at least in the context of specific work environments, which might explain unsatisfying consistency coefficients of aforementioned measurements (Kazemi, 2017;Page and Vella-Brodrick, 2009). For instance, Cockshaw et al. (2012) found that workplace belongingness and general belongingness are distinct constructs, although both are related to depression. Gallagher et al. (2009) found a better fit for their well-being model when positive relationships, part of eudaimonic well-being (Ryff and Singer, 2008), were moved to the social dimension, serving as an addition to Keyes' dimensions. Considering this, as well as the notion that highquality connections (Dutton and Heaphy, 2003;Stephens et al., 2011) include both positive short-term interactions and longer-term relationships, Fisher (2014) proposes the following conceptualization of social well-being at work: "feeling embedded in meaningful communities and having satisfying short-term interactions and long-term relationships with others". This definition taps into hedonic well-being and acknowledges the role of emotions and affective events at work as well as the eudaimonic notion of meaning and purpose. Integration of shortterm and long-term aspects also characterizes Morrison and Macky's (2017) socially oriented well-being measure aiming at workplace evaluation. As Kazemi (2017) discusses, maybe Keyes' dimensions reflect conditions contributing to social well-being and not the sense of social well-being per se.

Social interactions, relationships and belongingness
Fisher's (2014) definition might be a useful start for further conceptualization of social well-being at work, but first its dimensions require elaboration. Feeling embedded in a meaningful community refers to the need to belong (Baumeister and Leary, 1995); its satisfaction called belongingness (Malone et al., 2012). It may comprise concepts such as a sense of community, affective and normative organizational commitment and group cohesion, and on the negative side loneliness, social exclusion and ostracism. Sense of community results from feelings of inclusion, importance and mutual benefit, as well as shared emotions with others at work (Blatt and Camden, 2007), which refers to social interactions. Social interaction can range from noticing other people's presence to deliberate exchange of information. In essence, social interaction refers to seeing, hearing, smelling and touching other people (Argyle, 1968). It involves verbal and nonverbal behaviour between two or more individuals and can be contextual or enabling (De Jaegher et al., 2010). In open workspaces, conscious co-presence is an important contextual aspect, whereas verbal and non-verbal interactions enable transaction of information, including knowledge, feelings and emotions. Social interactions can be experienced positively or negatively. Positive interactions at work hold the subjective experience of vitality, positive regard and Conceptualizing social well-being mutuality, and they help building and maintaining relationships (Stephens et al., 2011), which provide emotional and instrumental social support (Dutton and Ragins, 2007). Friendships, referring to informal and expressive relationships that include personal and confidential communication (Khazanchi et al., 2018), are particularly beneficial for well-being. In summary, Fisher's (2014) conceptualization includes cognitive and emotional long-term and short-term reactions to the presence and behaviour of other people at work and to representations of a work community, being a formal team, department or organization or an informal group of co-workers.

Contemporary office context
Currently, many people in the Western world work in offices which increasingly feature open workspaces and desk-sharing. The ongoing trend is an office concept referred to as activity-based working (Engelen et al., 2019) or activity-based flexible office (Wohlers and Hertel, 2017), offering a diversity of open and enclosed spaces designed to support different work activities, usually accompanied with a policy of desk-sharing (Brunia et al., 2016). Field assumptions behind this office concept are that open spaces and flexible use of workstations increase communication, which should be beneficial to collaboration and performance. This seems a simple causal relation but is actually a complex mutual interaction, since the physical work environment does not determine employee behaviour (Gifford, 2014, p. 341;Vischer, 2008). Moreover, simply more communication is not always better for (social) well-being.
Although the diverse activity-based working environment is relatively new, open-plan and flexible (non-territorial) offices have been around for a few decades now, and research on the impact of open workspaces and desk-sharing on social interaction and well-being has been cumulating. Partly in accordance with aforementioned assumption, open workspaces featuring desk-sharing have been found to better support affinity than private offices (Bodin Danielsson and Bodin, 2009). Good opportunities for communication and interaction have also been found advantages of activity-based offices (Engelen et al., 2019).
However, open workspaces have been associated with increased noise and lack of privacy (Bodin Danielsson and Bodin, 2009) too, which outweighed the benefits for communication (Kim and de Dear, 2013), worsened interpersonal relations (De Croon et al., 2005) and increased conflict (Bodin Danielsson et al., 2015). Furthermore, employees in flexible, or nonterritorial, offices experienced more uncooperative behaviours, feelings of distrust and negative relationships compared to employees with their own workspace (Morrison and Macky, 2017). Hirst (2011) noted that the flexible use of workstations, or hot-desking, may create social tensions, mutual indifference and a sense of isolation from colleagues. Haapakangas et al. (2019) found that, after moving to an activity-based office, employees' belongingness and satisfaction with communication had decreased, and employees who came from private offices experienced negative effects on social support. When moving from open plan with assigned desks to a non-territorial activity-based office, reasons for dissatisfaction with the working environment were the high workstation-occupancy ratio, teams being split up, difficulties in finding colleagues and perceived injustice due to nesting (Rolf€ o et al., 2018).
In summary, the activity-based working environment can impact social well-being in many ways, both positively and negatively. However, it remains unclear what social well-being as a whole comprises in these specific work environments. Considering the possibly context-sensitive nature of social well-being and empirical research on specific demands and resources of activity-based offices regarding social aspects of well-being, this study aims to answer the following research question: "What are components of social wellbeing in activity-based offices?" JMP 36,4 Method Using existing case-study data This study used interview data from recent post-occupancy evaluations of two activity-based offices in the Dutch public sector. The choice for public offices is based on the observation of Engelen et al. (2019) that this sector is underexposed in activity-based office research. These two cases represent the latest standards for Dutch governmental offices, such as an employee-desk ratio of 0.9 to 0.7, a minimum of 6 m 2 floor space per desk and an average occupancy of 75% on peak days (Rijksvastgoedbedrijf, 2015).
Both cases are renovation projects where employees were brought together from several big and small offices located elsewhere into one large office (over 1000 workstations) emphasizing open-plan layout and featuring desk-sharing. In both cases for some employees, desk-sharing was new, while others already were used to it. Although officially non-territorial, most floors were assigned to (sub)departments, sharing building facilities. The evaluation was conducted by the Center for People and Buildings around 6-12 months after the organizations had moved into the new offices and included occupancy measures, a building assessment and an online survey followed by deepening group interviews (focus groups). This article used data from these live documented 60-90 min interviews with 2-8 employees each, who volunteered to participate and were grouped by department. The participants were not questioned directly about their social well-being, but aspects of social well-being came up while they were discussing their experiences with the new work environment. They were asked to name positive and negative aspects of their new work environment, to explain their feelings and to discuss possible solutions for perceived problems. Focus groups like this allow participants to bring up issues they deem significant and to challenge each other's views, offering the researcher insights in the way people collectively make sense of a phenomenon (Bryman, 2016, pp. 501-503).
Identifying key concepts through concept mapping To identify communalities in the interviewees' perceptions of social well-being, concept mapping was performed, a conceptualization methodology used in a wide variety of disciplines (Trochim, 1989(Trochim, , 2017. It consists of a multi-step process of quantitative and qualitative analysis resulting in a conceptual map of related concepts. Concept mapping is a useful technique for specifying target constructs in an inductive approach to scale development (Rosas and Camphausen, 2007) and for thematic analysis (Jonsen et al., 2009). Also, it offers a solid method for establishing content validity (Rosas and Ridings, 2017) and improves reliability of text analysis by coupling human judgement to statistical analysis and by engaging the target group (Jackson and Trochim, 2002).
Following the approach of Jackson and Trochim (2002), in our study the first step (see Figure 1) involved unit creation by cutting different interviewees' statements (a maximum of two sentences) on social well-being from the interview transcriptions and pasting them into an Excel file. Criteria for statement extraction were (1) content relating to aforementioned descriptions of social interactions, relationships and belongingness, following Fisher's (2014) definition of social well-being, and (2) substantial literal difference to previous extractions, since this study aims to identify themes rather than gather data on frequency or importance of topics. The extraction procedure was terminated after the analysis of the 19th group interview, at the number of 182 units, considering the capacity of human sorters to handle a maximum of 150-200 units each (Jackson and Trochim, 2002). Also, at that time no substantially different statements were coming forward, so criterion b seemed to be saturated. Saturation is recognized as a guiding principle in determining sample size of (group) interviews (Bryman, 2016, p. 418;Mason, 2010).
The second step included manually grouping the employees' statements, which were printed on paper strips to provide the sorters with a better overview than on screen Conceptualizing social well-being Figure 1.
Overview of the concept mapping procedure taking seven steps JMP 36,4 (Trochim, 2017). Trochim (1989) has suggested that a minimum of ten sorters is needed for a reasonable output reliability. In this study, 11 office workershaving experienced activitybased offices and being native Dutch speakersvolunteered to sort the 182 statements by content similarity and to label their self-determined groups. For each of them, the sorting task took around 90 min, which underlines the maximum capacity of 200 units.
Steps 3-5 (see Figure 1) included computer analysis of the statements' mathematical similarity. First, the 11 binary similarity matrices that indicate if statements belong to the same group or not, according to the sorter, were aggregated into one distance matrix. On this matrix, t-stochastic neighbour embedding was performed using Ward's algorithm and Python scikit-learn (Pedregosa et al., 2011), chosen to realize maximal distinction and internal consistency of the groups in a clear visualization (Van Der Maaten and Hinton, 2008). Their distances translated into coordinates, the statements were plotted in a two-dimensional space, including the sorters' labels. The hierarchical cluster analysis started with the 182 statements as single units, merging the two closest ones in every next iteration. From the 20-3-cluster stage, every iteration was studied to determine when further merging was considered illogical based on content. In the last two concept mapping steps, human judgement was used again to freeze the merging procedure at a logical moment and to label the clusters based on their statements' content.

Additional analysis for concept interpretation
Parallel to the concept mapping, the individual statements' content was reviewed to support the interpretation of the concepts to emerge. Two native Dutch speakers independently determined if the statement was phrased in a negative, neutral or positive way (inter-rater consistency κ 5 0.814) and if it referred to short-term interactions, long-term relationships or feelings of belonging (κ 5 0.811). Both reviewers entered their judgements into the Excel file. After termination of the cluster analysis, the statements were sorted by their cluster. Figure 2 shows the 14 key concepts that emerged from the interview data, visualizing the statements' similarity, as judged by the sorters, by their spacing on the map (the closer the similar). The number of 14 concepts results from the decision to freeze the clustering process ( Figure 1, step 6) before the statements on privacy and identity of the environment, two clearly different aspects, would have been merged. This number of 14 happens to be close to the sorters' average of 13 groups. The 14 key concepts are situated in three regions of meaning (Jackson and Trochim, 2002), or overarching categories, based on the 3-cluster stage of the analysis: social needs, experiences with (anti-)social behaviour of others and perceived social affordances of the physical environment. The number of statements per cluster rather reflects the duration or nuances of the conversation on this topic than the importance of the theme, since only distinctive statements were extracted, no matter their frequency. Table 1 shows that according to the additional interpretative analysis, the 14 concepts together reflect long-term eudaimonic well-being, represented by belongingness and relationships, as well as short-term hedonic well-being represented by social interactions, while most of the statements emphasize short-term social interactions.

Identified: 14 concepts in three areas
In the statements on social interaction, the interviewees described observations of different behaviours and ways of communicating and their emotional reactions to the presence and behaviour of random others. Most of the statements on relationships relate to encounters, proximity, community and norm setting, referring to maintenance of social ties through meeting and co-locating with colleagues and to clear rules of conduct as a way to stay Conceptualizing social well-being on good terms. Statements on belonging are prominent in cohesion, polarization and identity and mainly refer to familiarity with both people and environment, to group rights and obligations and to social exclusion. There is a negative tendency in the majority of statements on social well-being. Further, the identified concepts within the three main areas are described based on the included statements.

Social needs
One of three main topics of the extracted statements refers to the employees' thwarted or fulfilled social needs. As to negative experiences, they express how they feel disconnected from their close colleagues while having an emotional and functional need to be seated together. In their statements on connectedness and physical proximity, they refer to an increased social distance caused by the new office environment. They observe that to escape from the office noise and crowding, or because it feels useless to go to there if they cannot be with their co-workers, working from home has increased, resulting in less frequent co-worker contact. While in the office, the use of headphones makes co-workers look inapproachable, and contacts are formalized because deliberately meeting each other now requires appointments. Regarding community and cohesion, the interviewees observe less socializing, they miss "the traditional jokes", celebration of birthdays and staying in touch with personal events in their colleagues' lives. They express feeling like a visitor in their own office and missing the visibility of their team's work. Statements on these thwarted social needs reflect deterioration of close relationships and lack of belongingness as an observed downside of the new working environment in these cases.
Positive statements in the social needs area focussed on desirable encounters, such as the increase of spontaneous positive interactions while being on the move and at the coffee Each of the 182 employee statements on social well-being, grouped into 14 key concepts and three overarching categories that emerged from the concept mapping analysis JMP 36,4 machine and the ease of meeting colleagues due to spatial openness and mobility. Interviewees also enjoy choosing their workstation based on, for instance, visibility (the manager who wants to be approachable to his team) or inspiration (in the proximity of interesting colleagues), increasing the chance of positive social interaction. They perceive the new working environment to invoke more informal communication as well as more diverse connections.

Co-workers' (anti-)social behaviour
The statements on co-workers' behaviour in the shared workspace are largely about negative social interactions, such as claiming workspaces, disregard and downright hostile reactions (polarization), excluding outsiders. They describe experienced, witnessed and instigated incivility, as distinguished by Schilpzand et al. (2016), which seems to be connected to scarcity of workstations and may be unintended judging from interviewee statements such as: "It is not clear to me if it's sabotage or ignorance". As territory marking is restricted by the office concept lacking possibilities to personalize the workstation or to lock drawers, the employees apparently resort to other strategies for marking and controlling their workspace. The statements on territoriality refer to identity marking, such as spreading out belongings and using one's voice as an audible marker and to both anticipatory and reactionary defences as defined by Brown et al. (2005). One of the interviewees describes a creative or perhaps Conceptualizing social well-being desperate strategy: "I came across a note saying 'out of order', but it turned out to be a reservation." Another refers to social pressure: "Co-workers told me I was a fool to give up my workstation, so now I leave my stuff too." They experience difficulties in correcting such behaviours of others, because they do not feel they have the right to do so, do not feel comfortable doing so, have had negative experiences while doing it or do not want to put the effort into it. Therefore, they desire more clear rules and stronger action taken by the managers (norm setting). The concept of indifference includes decreased responsibility, for instance, cancelling meeting rooms and reporting issues, emotional social support dealing with the new situation or the lack of support, as well as a positive side effect of less social cohesion being less social pressure to answer charges of absence. The only other positive statement in this overarching category refers to making an example by giving emotional social support.

Perceived social affordances
The third group of social well-being concepts that emerged from the concept mapping refers to social affordances (Gibson, 2014;Spreitzer et al., 2020), properties of the environment facilitating or obstructing positive social interactions or feelings of belonging. The interviewees' statements on the perception of these affordances are all negative. They refer to a lack of spatial enclosure and to the high spatial and social density, increasing the negative consequences of social interaction (noise, feelings of crowding) while not being able to control the amount of interactions. Many statements concern unwanted talking in an open workspace and the feeling of having nowhere to go for a (private) phone call: "My colleague prefers making calls in the toilet for the disabled." Regarding belongingness there are two different remarks on the unwelcoming d ecor and a complaint on getting reprimands for hanging things on the wall.

Theoretical implications
In addition to deductive models of social well-being (Fisher, 2014;Gallagher et al., 2009;Keyes, 1998), this study serves as an inductive source of outer-directed well-being aspects that are significant to employees working in activity-based offices. The social well-being components identified reflect the broad spectrum as well as the significant nuances of social well-being as it is perceived in contemporary activity-based offices.
The results indicate that social well-being might not only include long-term, relatively stable eudaimonic well-being but also short-term hedonic well-being. Our data show that interviewees bring up how specific social interactions at work directly impact their well-being as well as how interactions, or lack of social contact, add up to thwarted belongingness and disintegration of social cohesion. This aligns with literature on relationship and community building (Blatt and Camden, 2007;Stephens et al., 2011). In their statements the interviewees emphasize short-term interactions and emotional responses. This points at the significance of daily hassles in the office or affective events (Ashkanasy et al., 2014) for social well-being and favours Fisher's (2014) conceptualization of social well-being including both short-term and long-term well-being. After a period of opposition between these two philosophical traditions, namely hedonics questioning the conceptual and methodological sophistication of the younger eudaimonia and eudaimonics considering hedonic pleasure less important to a good life, several scholars now think both should be integrated (Henderson and Knight, 2012;Lambert et al., 2015). As Waterman (2008) argued, eudaimonia does not exist without hedonia, which the concept mapping appears to confirm.
Regarding social well-being components, the results show that of Keyes' (1998) five dimensions, only two areto some extentrepresented in the identified concepts. Although JMP 36,4 the statements differ significantly from Keyes' operationalization into survey questions, it could be argued that the concepts of cohesion, community, connectedness and incivility refer to (1) social integration and that the perceived polarization, territoriality, indifference and norm setting relate to (2) social acceptance. None of the identified concepts seems to relate to (3) social contribution, (4) social actualization or (5) social coherence. Possibly this is the case because these dimensions require reflection on their own role and understanding from a higher level of abstraction, which apparently did not come up for discussion in the group interviews, at least not explicitly related to presence and behaviour of co-workers, supervisors or the organization and may be less significant to employee well-being.
Additionally, a new dimension has risen from the data, reflecting a connection with the physical context. This dimension comprised undesired social interaction, lack of support of relationship building and thwarted belongingness, as well as increased possibilities for positive interactions invoked by the environment. The concepts of the environments' visual identity, privacy support and noise and crowding caused by co-presence and positive encounters reflect the known pros and cons of activity-based offices as summarized earlier. This indicates that social well-being might not be a general phenomenon, reflecting functioning in social life anywhere, but a local phenomenon bound to (physical) context, reflecting the entanglement of social and physical environment.

Limitations
The cases' office workers were not directly asked about their social well-being, thus their statements refer to topics put forward by one or more group members. These topics might have been their most prominent or most recent experiences, or the ones that were easiest to talk about, particularly in a group interview. These could have been the negative experiences, because those are remembered more easily and in one of these cases the evaluation showed that the users were relatively dissatisfied. It is possible that especially dissatisfied employees volunteered to participate in the group interviews. Furthermore, the longer the discussion, the more different statements could have been made about the specific topic, possibly at the expense of other topics. This might explain the emphasis on short-term interactions, since these are more concrete and probably more frequent than thoughts about long-term relationships or community awareness. On the other hand, the topics that have come up do reflect participants' significant experiences.
Other limitations concern the cases' specific cultural setting (Dutch government offices), relatively dissatisfied users and size (over 1,000 workstations), as well as the one-researcheronly extraction of units and labelling of clusters. Using multiple interpreters and repeating the study with data from other settings, for example, smaller offices where people know each other better, offices where there is less scarcity on workstations, that have more enclosed spaces or where the users are more satisfied with their environment, could yield additional social well-being components to be considered in further conceptualization.

Directions for future research
The context-bound properties and hedonia-eudaimonia integration of social well-being found in this study require further exploration and validation, for instance, by developing and testing an occupational social well-being scale that captures the specific conditions in an activity-based office and addresses both hedonic and eudaimonic well-being. The study can be taken as a first step of inductive item generation, to be completed with deductive item generation from existing scales and theories and several iterations of testing and validation, as suggested by Boateng et al. (2018). An item pool could be built up by mapping items from existing scales onto the 14 key concepts. Additionally, the 14 concepts could be used as a conceptual coding scheme for analysing qualitative data on social well-being aspects in contemporary offices, which could lead to new items or concepts serving scale development.

Practical implications
The interviewees' statements on the impact of their new working environment provide knowledge for team managers as well as facilities and human resource management. To prevent decrease of social well-being when moving into an activity-based office, managers should carefully prepare their team for the new situation and discuss possible difficulties as described in this article. In the new environment they should be alert to the signs of weakening ties, environmental stress, conflict and withdrawal. They should keep in touch with their team, monitor social well-being, acknowledge the sense of loss of the old environment and act upon troubles, for instance, addressing an evolving incivility spiral (Andersson and Pearson, 1999) as soon as possible.
To facilitate building positive relationships and nurture feelings of belonging in activitybased offices, workplace managers and designers should carefully balance the visual and physical openness and standardization of the environment with the employee's need for privacy, identity marking and a sense of community. The interior surfacing should provide adequate sound absorbing and the work environment should offer a variety of spaces for personal conversations, uplifting chit-chat (Methot et al., 2020), spontaneous encounters and social events, as well as possibilities for and marking of team and organizational identity. Participatory design that involves employees and managers could be a fruitful way to inform the design as well as enhance belongingness and feelings of ownership.
Regarding the return to the office after COVID-19 lockdown, managers should balance protection against the virus with support of relationship-building activities and give special attention to new employees' integration and staying in touch with home-workers. Now increased working from home is expected to stay, the office's social function as a meeting and community place will become even more important. This requires appropriate facilities for a larger proportion and diversity of face-to-face interactions, as well as hybrid meetings combining conventional and video conferencing, but this still has to be balanced with adequate spaces for rest and working individually. The office has to offer a welcoming place for work and bonding with co-workers and organization. A positive side effect of COVID-19 social distancing rules might be a reduction of crowding and related noise problems.
Although from a hygiene and belongingness perspective, fixed workstations might be better than hot-desking, with substantial working from home this probably is not financially efficient. However, introducing identity rich home zones for teams may serve as a semi-territorial compromise, while a well-considered location of teams and mixing of departments and centralized facilities will still support casual encounters.

Conclusion
This article contributes to conceptualization of individual social well-being at work. It provides a first step in development of a social well-being scale tailored to the context of contemporary offices. Concept mapping of case-study data revealed 14 key concepts reflecting employees' social needs, their confrontations with co-workers and the physical environment's social affordances. This indicates thatcontrary to established theorysocial well-being includes components of both short-term hedonic and long-term eudaimonic well-being and is nested in a spatial context, and thus, workplace characteristics are to be considered when measuring and managing social well-being at work.