Social Interaction and Power Relations in Companies: An Entropy-based Approach Using the Expert System Shell-SPIRIT

Abstract

The changed cooperation and organizational structure in companies are to be examined quantitatively with regard to their power structures and changes are to be revealed. The expert system Shell Symmetrical Probabilistic Intensional Reasoning in Inference Networks in Transition (Shell-SPIRIT, a.k.a., SPIRIT) is to be applied as the SPIRIT tool was already used for different applications in the context of information faithful knowledge processing, wherewith it’s achievement potential becomes visible, see, e.g., Rödder et al., Logic Journal of IGPL, Vol. 14 (2006). More recently the potential to model power structures and measure their power potentials in the SPIRIT tool has been recognized, see, e.g., Rödder et al., Decision Analytics Journal, (2021). This chapter will discuss from a practitioner’s perspective, how power structures in companies might have changed by Corona Virus Pandemic (a.k.a. Corona) in order to adequately model this in the SPIRIT tool. The chapter also presents and discusses the literature research, modeling results of power structures and rankings based on employee skills obtained from the SPIRIT tool, and their practical relevance.

1. Introduction

Do we see – based on an increased virtual corporate-related interaction since Corona [1] – a changed cooperation and organization in companies? Are effects on known power structures recognizable? In this chapter, these questions are to be examined quantitatively on the basis of a practical case study and changes are to be revealed. For the analysis, the expert system shell SPIRIT will be applied. SPIRIT has already been used for various applications in the context of information-based knowledge processing, which makes the performance potential of the program visible. The interested reader is referred to [2]. More recently, the potential of using SPIRIT for modeling power structures and measuring power potentials has been recognized, see for example [3, 4].

Beyond the already known analysis approaches according to [2], this chapter will deal with the direct consideration of (weighted) input factors in the expert system Shell-SPIRIT. The approach presented here is novel in the sense that in the past, power structures in SPIRIT were calculated quantitatively based on a given ranking. With this novel approach, rankings are no longer necessary. Instead, weighted input factors are used, which implicitly and directly result in a ranking in SPIRIT and thus directly reflect the power structure. In this sense, the process step of determining a ranking can be dispensed with and directly based on the more tangible input factors. This direct consideration of input factors enables new analysis approaches and shows the decision basis in a highly transparent way in one step, thus increasing the quality of quantitative decision-making.

In order to give this contribution practical relevance, we will first discuss, from the practitioner’s point of view, how power structures in a company might have changed pre- and post-Corona, in order to show and evaluate in SPIRIT then adequately. An initial literature review on the differences between physical and virtual collaboration shows that increased collaboration via digital communication channels is evident in everyday corporate practice.[1] Various studies have also recognized these changing conditions and examined the resulting effects on companies. One example is that digital leadership of employees is different than physical leadership. Accordingly, leadership styles must be adapted and adjusted to the changing conditions of digital collaboration.[5] But it is not only the requirements for employee leadership that are changing, general virtual collaboration also brings changes. For example, social dynamics can hardly be conveyed virtually, so that discussions are difficult to follow. Non-verbal communication is also significantly restricted, so that, for example, it is impossible to tell who is looking at whom, and it is also more difficult to understand who is more or less important in the discussion [6, 7].

Beyond these individual examples, [8] conducted an extensive meta-analysis of 255 studies looking at the differences between physical and virtual collaboration. Challenges of virtual collaboration were classified into five categories (geographical distance, temporal distance, perceived distance, the composition of decentralized teams and the diversity of participants), which show very clearly how multifaceted the change in collaboration due to virtual collaboration is. In addition, however, it is also significant that beyond the change in interaction, the key competencies of employees have changed [9].

In order to discuss the introductory questions and findings in a structured way, the article is organized as follows: After a concise introduction as well as an overview of the state of research in Section 1, Section 2 deals with different factors for power building. Section 3 deals with the differentiation of the power building process with respect to physically and virtually conducted meetings. Incorporating previously published articles on the topic of power in probabilistic networks, a first model is then shown in which the previously presented power structures are mapped. This even makes it possible to precisely quantify the power potentials of individual actors in such a network. Finally, the last section, Section 4, concludes with an outlook on possible further research activities in the research field under consideration.

1.1 Further explanation of the case study used

Virtual collaboration is widespread. The necessary technical requirements, such as a fast and stable Internet connection, digital data storage, location-independent work using laptops and various communication media, have already made it possible to work remotely for a longer time. The communication possibilities go far beyond written correspondence via e-mail (e.g., chats, telephone and video communication, which work bilaterally but also easily in larger groups) and enable efficient exchange. The Corona pandemic at the beginning of 2020 significantly changed the world, and also the way of professional collaboration (esp. in areas of activity that work with information, colloquially referred to as ‘classic office work’). The possibilities of remote work were not used across the board, but the home office share increased significantly from 2020 onwards.[1] For the above-mentioned case study, which is shown in the following, the following comments are based on Schröer/Reucher [10].

Corona has thrown a large number of companies in at the deep end and they have had to promote remote work and offer home office (at times even legally mandatory for such positions, e.g., in Germany [11]) in order to combat the pandemic. Apart from legal regulations, it is up to employers themselves to decide to what extent a home office is offered. However, it should be noted that in a time of shortage of skilled workers (also called ‘war of talent’), the home office option is desired or even required by many employees.[12].

As previously explained, the Corona pandemic has fundamentally changed the way collaboration is experienced in activities that can take place regardless of location. These changes are already having an effect on everyday corporate life, and it is to be expected that there will be no simple ‘going back’ – not without reason are we talking about a ‘New Way of Working’ post-Corona.[13] But what implications do these developments have for power structures in companies? Certainly, power in companies, similar to society or even politics, is defined by the formal, but also a lived, hierarchy. Regardless, it takes expertise and persuasion in the form of communication skills to get the necessary majority or (and more importantly) the crucial supporters in critical discussions. The assumptions for this contribution to the discussion are therefore formulated as follows:

• In traditional physical meetings, in which all discussion partners come together in one room, hierarchical power structures have a stronger impact than in virtual meetings. Accordingly, managers who tend to act via defined ‘chains of command’ and are less convincing with content in the discussion, have a greater interest in coordinating in physical meetings.

• Virtual meetings can have a positive effect on the outcome of the discussion if discussions are content-based and open-ended, with all contributions to the discussion being evaluated equally by the recipient, regardless of the sender.

Assuming that virtual collaboration will continue to be a significant component of companies post-Corona, the aim is to discuss, in relation to the hypotheses outlined above, whether the way in which social interaction functions in entrepreneurial organizations and the associated power structures, will undergo lasting change as a result of a ‘New Way of Working.’ For this first contribution to the discussion, we will initially proceed on an assumption-based basis and use simplified models. If these initial approaches show interesting theoretical as well as practical relevance, the discussion approach presented here will be analyzed in more depth.

In line with this initial situation, the focus of this chapter will be on modeling and comparing corporate power structures in physical and virtual meetings. The analysis of power structures is primarily based on [3]. Therefore, the changing professional collaboration since the beginning of Corona is examined as a case study (see above). Here, the question of what consequences virtual collaboration has on the power building process in organizations is discussed. If it turns out that the meeting form of the physical or virtual meeting has a sufficiently high impact on the power structures in the corresponding meeting, and if this can influence decision-making, there is a high interest in selecting the meeting form in favor of the meeting organizer. For this purpose, the first step is to show quantitatively that the two mentioned meeting forms with a comparable group of participants differ with respect to their power structures. The model used for this purpose should be designed in a comprehensible way so that it can be adapted and find a practical application. If input factors are clearly comprehensible and easy to adjust to the respective meeting participants, this model could be used to quantitatively evaluate the ‘optimal’ meeting form for the organizer in critical meetings.

1.2 Status quo of research

This section consists of two parts; both had to be examined with regard to their current state of research. A direct combination of the two research areas ‘social interaction and power relations in companies’ and ‘changes in remote work and its developments post Corona’ could not be found in the course of a literature search. Due to the significant change in the course of the ‘New Way of Working,’ the research question can thus be attributed a high topicality and significance in theory and practice.

First, the research status quo of quantitative analysis of social interaction and power relations in companies was drawn from [2, 3, 4] and [14]. For the second area of change in remote work and its developments post-Corona, the following sources were used in particular, [5, 6, 7, 8, 9, 12, 13, 15, 16, 17, 18, 19, 20, 21, 22].

2. Power building processes in companies

2.1 Selection of appropriate input factors

As already discussed in Section 1, the selection of suitable input factors is crucial for the practical significance of the new quantitative analysis presented here. This is because the newly developed analysis approach is based directly on the input factors. Only with the appropriate input factors, which give a complementary and non-overlapping composition (following the MECE rules [23]) of the facts to be analyzed on a selected aggregation level, derived results can be discussed meaningfully.

At this point, a distinction must be made between presenting the newly developed quantitative analysis on a theoretical basis using a theoretical case study and discussing the content of input factors for power structures. In the first case, it is considered more than sufficient to select meaningful, assumption-based input factors and to present their conceptual impact in the course of the quantitative analysis. For the second case, a more in-depth analysis of the input factors, including the corresponding literature research, is necessary, because only in this way can a purposeful discussion of the content be guaranteed. Since the application of weighted factors in SPIRIT has not been researched much yet, the focus in this elaboration shall be on the quantitative decision making and the input factors shall be used in a pragmatic way.

2.2 Power building: Changes through ‘New Way of Working’

As already described in the introduction (see above), since the beginning of the Corona pandemic, the number of working days not performed in the employer’s office has increased significantly. With the increase in the home office quota, classic structures in the daily work routine of employees who basically perform remote activities are broken down, thus increasing the share of virtual meetings [13]. For the following analysis in Section 2.3, which is essentially based on [10], two meeting variants are to be analyzed: The ‘physical meeting,’, in which all participants of a meeting are in one place, and the ‘virtual meeting,’ in which all participants are independently connected virtually. Hybrid meetings, in which participants are partly in one place and partly virtually connected, are not considered in this section.

For a more detailed explanation of the development of the ‘New Way of Working,’ the reader is recommended to read ‘The Impact of New Ways of Working on Organizations and Employees: A Systematic Review of Literature’ [24].

2.3 Factors of power building in different meeting structures

2.3.1 Basic assumptions of power building

The following elaborates on the pragmatic and assumption-based derivation of the input factors to be used in the power building process. As mentioned above, the focus in this section is on the quantitative analysis of power structures in the expert system Shell-SPIRIT, and not on the theoretical discussion of appropriate input factors for the power building process.

The power constellation of individual participants in a discussion is multi-layered and related to the two meeting variants mentioned above. Individual input factors have different effects. For this purpose, three clusters of factors for power building shall be distinguished, which will be explained in the subsequent sections. The factors described here have no claim to completely describe corporate power building but are intended to awaken an initial assumption-based basic understanding of it. Based on this, a first model will then be created and analyzed in the SPIRIT to see what different effects the factors have in the ‘physical meeting’ and ‘virtual meeting’ variants. Depending on the result, a more detailed investigation into the selection of input factors will then be critically examined and granularly derived in a later section.

2.3.2 Hierarchy structure

The first factor in the formation of power is the hierarchical structure. Based on the organizational chart of a company, disciplinary and functional responsibilities become clearly visible. Power structures and relationships can be derived accordingly. However, corporate practice shows that this is not sufficient to describe hierarchy structures comprehensively. Further sub-parameters are age, which often brings with it a natural seniority and length of service with the company. Finally, involvement in discussions and decisions is also taken into account.

2.3.3 Expertise

A second factor in power building is the expertise of discussion participants. Assuming that issues are discussed in a factual and open-ended manner, expertise is a significant building block for developing power and advancing one’s own issues. If individuals understand issues better and can place them in the overall corporate picture, it is possible to dominate other participants factually across hierarchies and to set one’s own agenda.

2.3.4 Communication skills

Another factor to be mentioned is the ability to communicate. With the ability to communicate one’s own knowledge about topics and facts in a comprehensible way and to present it to the discussion partners in a traceable way, one’s own power can be expanded. The communication ability can be based on aspects such as self-confidence or nervousness to speak in front of others or the ability to communicate a contribution in a foreign language (e.g., English as a non-native speaker). Depending on the topic of discussion and the level of detail, lack of expertise can also be covered up with strong communication skills (to a certain degree).

3. Power building process in physical and virtual meetings: a quantitative analysis

3.1 Approach to quantitative analysis

The case study below is essentially from [10]: The hierarchical position, the expertise and the communication skills of the discussion participants have different effects on the power structure of a group of participants, which also has an effect on the outcome of a technical discussion, as well as on the development of power structures of an organization. But how do these factors differ in the two meeting structures, physical and virtual discussion?

In order to quantitatively analyze the effects of different meeting structures in terms of power building, a freely chosen example is used.

In the case study, the technical area of the company, controlling and IT discuss the introduction of a new software for tracking research success. It can be assumed that, apart from team members, every position in the organization chart shown has personnel responsibility. For a clear organizational chart, only positions that are necessary for the case study have been listed here below the executive board. The above-mentioned participants are marked in the organization chart (Figure 1).

Figure 1.

Case study – Organization chart.

The different actors, which have their place in both official and lived organizational charts of the company, have individual expertise and individual communication skills. The following is an overview of these (Table 1).

	TN 1	TN 2	TN 3	TN 4	TN 5
Brief description	Experienced manager, long service and high reputation in the company	Employee new to the company with technical expertise, few communication skills	Young employees, with the company for a longer time, communication skills available	Young employee who is new to the company, expertise available	Recently joined the company, expertise and communication skills available
F1: Lived hierarchy:	1	0	1	0	0
F2: Expertise	0	1	0	1	1
F3: Communication skills	1	0	1	0	1

Table 1.

Case study – Meeting participant skills.

Simplified case example: 0 = capability is not presence; 1 = capability is presence

In the chosen analysis approach, the organizational structure applied is to be kept constant for the evaluation of the effect of different meeting structures on the power building of the meeting participants. The selected individual input factors of the participants are also to be held constant throughout, for an initial quantitative analysis (Figure 2).

Figure 2.

Overview factors of power building.

In the first simple model for quantitative analysis, only whether a participant has a factor for power building (= 1), or not (= 0), is evaluated. In order to increase the practical relevance, input factors can be considered on a more differentiated value scale in further elaborations, but this goes beyond the scope of this section. On the other hand, the assumed factors of the power building process are assigned a different value depending on the meeting variant. From these two components (ability and weighting) ‘a personal score’ per meeting participant is derived, which indicates the respective negotiation strength, and thus the power building.

Figure 3 shows that in the ‘Skills’ block, participant 1 (TN 1) is at the top of the hierarchy, but has no specialist knowledge, but is able to communicate. The entries of the other participants were read accordingly. In physical meetings, TN 2 and TN 4 are the most powerful, followed by TN 5, with TN 1 and TN 3 both being the least powerful. In comparison to the virtual meeting, however, TN 1 and TN 3 are now the most powerful, followed by the ‘equally powerful’ TN 2 and TN 4, with TN 5 bringing up the rear in the ranking here.

Figure 3.

Case study – meeting participant skills.

3.2 Mapping of power structures in the expert system Shell-SPIRIT

3.2.1 Preliminaries

First, a very brief overview of the concept of an entropy-based¹ knowledge acquisition process with a focus on power structures is given. Given a set of variables V = {V1…, Vn} with Boolean variables Vi = vi, and values vi = 0 or vi = 1, for all i = 1…n. Vi = 1 means, Vi has power and Vi = 0, Vi is powerless. Denotes R a set of (certain) conditionals {Vi = vi|Vj = vj [1.0], for some i, j = 1…n}, then the knowledge acquisition process in SPIRIT follows the principle of maximum entropy:

$${\mathrm{P}}^{\ast }=arg\left(Max\mathrm{H}\left(\mathrm{Q}\right),\mathrm{Q}\text{fulfills}\mathrm{R}.\left(\ast \right)\right)$$

Among all distributions for which R holds, P* is the one with maximum entropy, i.e., only the dependencies explicitly formulated in R will be emitted in P*, nothing else. In the case of R = {}, Q = P⁰ (uniform distribution). For more details, the reader is referred to [2].

Example: Let V1 and V2 be two binary variables with values 0/1, than P⁰ (V1 = v1, V2 = v2) = 1/4 for all four configurations v1 = 0/1,v2 = 0/1 and H(P⁰) = 2 bit. Now, let us assume, R = {V2 = 1|V1 = 1 [1.0]} “If V1 has power, than V2 also [with probability 1.0],” the result of (*) is P*(V2 = 0, V1 = 1) = 0 and for all other three configurations maximum entropy provides P*(V2 = v2, V1 = v1) =1/3 and H(P*) = 1,58 bit. In particular, we get P*(V1 = 1) = 1/3, P*(V2 = 1) = 2/3. The result may be surprising at first, because why should V1 be less powerful than V2? This will be discussed in more detail in the next section.

The following explanations, as well as the explanation of the conventional (Section 3.2.2 and Section 3.2.3) and novel (Section 3.2.4) analysis methods of the power building process in the expert system Shell-SPIRIT are mainly based on [10]: As presented before, there are different forms of how power relations can work. In the following overview, the basic relations between meeting participants are presented. A distinction is made between fixed factors (a), which remain constant regardless of the meeting variant chosen, and variable factors (b). The input factors of the variable factors of individual participants vary depending on whether the meeting takes place physically or virtually; see the explanations above (Figure 4).

Figure 4.

Case study – Power relationships.

Within the framework of a probabilistic model, different forms of power can also be depicted, namely the power constellations ‘all against all,’ ‘everyone against everyone’ or ‘give and take.’ For more information, see [3]. For power structures in companies, the form ‘give and take’ seems most realistic, so that this will be applied in the further course of the section.

For the analysis in the expert system Shell-SPIRIT in the following, the skills of the participants shown in Figure 3 will be used, with the weighting of the skills depending on the meeting variant and the resulting power ranking. The effect of the official organizational chart is indirectly included via the lived hierarchy factor. The change in key competencies due to increased virtual collaboration was discussed in detail at the beginning of this section [9].

3.2.2 Physical meetings (classic discussion pre-Corona)

In the expert system Shell-SPIRIT, (safe) rules are stored, which are based on the selected case study considering the rank orders (see Figure 3). Here, the rules with the conditional operator are as follows:

• Index 3: ‘If TN_1 has power, then he exercises it on TN_2,’ i.e., participant 1 (TN_1) exercises power on TN_2, since the latter has a higher rank.

• Index 4: TN_3 exercises power on TN_ 2, because he has a higher rank.

• The remaining rules are shown in Figure 4 below.

The results from the SPIRIT tool are shown in Figure 5. Each participant TN_i (i = 1…5) is represented as a Boolean variable with the semantics TN_i = 1, ‘participant i has power’ or TN_i = 0, ‘participant i has no power.’ Basically, we have to distinguish between the view ‘pm’ = ‘show probability’ (Figure 6) and ‘im’ = ‘show information’ (Figure 7). If p_i denotes the probability that participant i exercises power, the impact im = −ld p_i measures the power potential of actor i [3]. Here, ld denotes the logarithm dualis, which makes the unit of the impact im [bit]. For a deeper discussion, see [2] or [3].

Figure 5.

Power structures in physical meetings.

Figure 6.

Probability to exercise power in physical meetings.

Figure 7.

Impact on the exercise of power in physical meetings.

Figures 6 and 7 hold information. While Figure 6 shows the probabilities with which individual participants exercise power, Figure 7 provides the impacts of the individual participants on the exercise of power. Here we see that participants TN_1 and TN_3 are the most powerful with an impact of im = 2.000 bit each, followed by TN_5 with an im = 1.000 bit, ‘bringing up the rear’ are TN_2 and TN_4 with im = 0.4150 bit each. This corresponds exactly to the ranking list for the power structures in the physical meeting; compare the explanations in Section 3.1 and Figure 3 for a summary.

3.2.3 Virtual meetings (reinforced discussion form post-Corona/new way of working)

In addition to the previous section, in which power structures in physical meetings were modeled using the SPIRIT expert system shell, power structures in virtual meetings will now also be considered. Here, too, (safe) rules are formulated in the SPIRIT expert system shell, which are in turn based on the selected case study (see Figure 3). Here, the rules are to be understood as follows:

• Index 3: TN_1 exercises power over TN_2, since the latter has a higher rank.

• Index 4: TN_4 exercises power over TN_1, since the latter has a higher rank.

• The remaining rules are shown in Figure 8.

Figure 8.

Power structures in virtual meetings.

The two screenshots below show the quantified results from the SPIRIT expert system shell. The technical background has already been explained in Section 3.2.2.

Guidance on the basic interpretation of Figures 9 and 10 was given in Section 3.2.2. Specifically, Figure 10 shows that TN_5 is the most powerful with an impact of im = 3.000 bits, followed by TN_2 and TN_4 with an impact of im = 1.4150 bits each; TN_1 and TN_3 have the least power with an impact of im = 0.4150 bits each. Comparable to the physical meeting (see Section 3.2.2), the power order of Figure 3 could be reproduced correctly via the expert system Shell-SPIRIT.

Figure 9.

Probability to exercise power in virtual meetings.

Figure 10.

Impact on the exercise of power in virtual meetings.

As an interim conclusion, it can thus be stated that if the ranking of power structures is known, these can also be adequately represented in a probabilistic model. In the next section, we will show how power structures arise in the expert system Shell-SPIRIT when (weighted) input factors of the individual participants are known. Again, both cases, physical meetings and virtual meetings, are considered separately.

3.2.4 Power relationships based on (weighted) input factors

Changed power structures due to physical (3.2.2) and virtual (3.2.3) meetings could be shown in the two sections above. However, to determine the power structure, a ranking was necessary as an input factor. This input factor had to be determined in advance, so that the quantitative analysis of power structures always required two process steps, with the determination of the rank order as an input factor and the actual calculation of the power structures in SPIRIT. In the following, a new approach is presented, in which the entire computation is possible alone and in one process step in SPIRIT.

In order to derive the power structures in one work step, with reference to Figure 3 the three abilities F1 (lived hierarchy), F2 (expertise) and F3 (communication skills) are defined in each case as Boolean variables (1/0), whereby Fi = 1 means, ‘ability Fi available’ i = 1,2,3. In order to be able to consider now also the weighting G1:G2:G3 accordingly, F1_g1, F2_g2 and F3_g3 (gi = 1,…Gi, i = 1,…,3) Boolean variables are defined.

In the physical meeting, the weights were G1 = 3, G2 = 1 and G3 = 2 (see Figure 3), which finally resulted in the 3 + 1 + 2 = 6 variables F1_1, F1_2, F1_3, F2_1, F3_1, F3_2. Participant 1 (TN1) has the abilities F1 and F3, which are mapped rule-based via the index 15 (triple weighting of F1) and 19 (double weighting of F3). The other conditionals 16 to 21 and 27 read accordingly (Figure 11).

Figure 11.

Weighted skills in physical meeting.

After learning the rule set in the expert system Shell-SPIRIT, the following impacts² were calculated for the physical meeting per participant:

This results in the desired power ranking again. TN1 and TN3 are the most powerful with im = 3.6049 bit each, followed by TN5 with im = 2.3154 bit, and the ‘weakling’ is formed by TN2 and TN4 as equally powerful with im = 1.2512 bit each, which again corresponds exactly to the power rank order in the physical meeting, determined by aggregating the weighted input factors previously in Figure 3. Next to the explained quantitative analyses (Figure 12) in bits the probability values are shown in Figure 13.

Figure 12.

Impact to exercise power in physical meetings with weighted skills (in bits).

Figure 13.

Impact to exercise power in physical meetings with weighted skills (in probability values).

In the virtual meeting, the weights were G1 = 1, G2 = 3 and G3 = 1 (cf. Figure 3), resulting finally in the 1 + 3 + 1 = 5 variables F1_1, F2_1, F2_2, F2_3 and F3_1. Participant 1 (TN1) possesses the abilities F1 and F3, which are mapped rule-based via the index 14 (simple weighting of F1) and 15 (simple weighting of F3). The other conditionals 16 to 21 read accordingly (Figure 14).

Figure 14.

Weighted skills in the virtual meeting.

After teaching the rule set in the expert system Shell-SPIRIT, the following impacts³ are calculated for the virtual meeting per participant:

Again, applying weighted input factors yields the expected power order (see also Figure 3). TN5 has an impact of im = 2.2780 bit, TN2 and TN4 have im = 2.01510 bit. TN1 and TN3 have the least power with an impact of im = 1.6930 bit each. Next to the quantitative analyses (Figure 15) in bits the probability values are shown in Figure 16.

Figure 15.

Impact on the exercise of power in virtual meetings with weighted input factors (in bits).

Figure 16.

Impact on the exercise of power in virtual meetings with weighted input factors (in probability values).

4. Outlook

This chapter combines the research area of social interaction and power relations in companies and the change of remote work and its developments post-Corona based on a quantitative analysis. It was able to show with first simple assumptions that the power building in virtual meetings behaves differently than in physical meetings. In the same course, a new calculation method was presented, which does not determine the power potentials in SPIRIT on the basis of a given ranking, but directly on the basis of weighted input factors. This enables a faster and more transparent analysis process in the SPIRIT tool for a wide variety of application fields for quantitative decision-supporting models.

There are further exciting areas of application for the case study available, which can be elaborated on the basis of the findings presented in this chapter. First, factors of the power building process were assumed in a simplified way. In order to increase the practical meaningfulness, these factors have to be examined more closely and, if necessary, also defined with regard to regional or cultural differences. Furthermore, this chapter examined the power building process for a point in time. However, how the process of power building develops with different meeting variants over a longer period of time has not yet been considered. Accordingly, the analysis of power building over a longer period of time could be the focus of a further expansion stage. Also, questions could be raised about what impact it has if participants in the conversation have already met physically and then coordinated virtually, or only meet without any physical meeting at all. In addition, in the course of the quantitative analysis, impact values were determined, and for the moment, only evaluated in terms of which order of power results the different calculated impact values have. It has not yet been discussed how the concrete impact value or the impact delta to higher or lower power ranks is to be interpreted. It can already be seen that the impact values differ between the comparable scenarios (cf. Section 3.2) with and without weighted input factors.

Beyond this further consideration, it may also be possible to work out recommendations for action for managers in order to expertly use their individual power-building skills when weighing up the use of physical and virtual meetings. It will be more than interesting to see what other insights are gained in the process.