Business model innovation decisions: the role of group biases and risk willingness

2.1 Business model, business model innovation and decision-making

Business models (BMs) explain “how firms do business” (Zott et al., 2011, p. 1021). BMs define how companies create, deliver and generate value (Granig and Hilgarter, 2020) by exploiting business opportunities (Amit and Zott, 2001). BMs are a potential source of competitive advantage (Casadesus-Masanell and Ricart, 2010; Teece, 2010) and subject of strategic (Zott et al., 2011) and innovation management literature (Chesbrough, 2010). To fully understand the content of a BM, scholars suggest describing the individual elements in concrete terms. We used the nine BM elements of the BM Canvas defined by Osterwalder (2004): Value proposition, customer segments, key activities, key partners, key resources, customer relationships, channels, cost structure and revenue streams.

Due to dynamic changes in the environment, companies must constantly adapt their BM (Morris et al., 2005; Kraus et al., 2023; Snihur et al., 2023) and their BM elements. This adaptation, modification, change and redesign of an existing BM are referred to as BMI (Lindgardt et al., 2009). When internal and external opportunities or threats arise (Bucherer et al., 2012; Kraus et al., 2023), BMIs are an appropriate managerial approach to remain competitive (Geissdoerfer et al., 2018). BMIs increase the resilience of an organization (Casadesus-Masanell and Ricart, 2011) by identifying new ways to create and generate value (Amit and Zott, 2001; Böttcher et al., 2022; Teece, 2010). The implementation of BMI involves making strategic decisions. Although dynamic external changes require fast decision-making (Granig and Hilgarter, 2020), many companies fail to successfully implement BMI due to time-consuming decision-making (Doz and Kosonen, 2010). To solve this problem, recently, Snihur et al. (2023) developed a practical tool that helps executives determine whether and when they need to restructure their BM.

The BMI process has been described as an important construct and an intensely debated topic (Schneider and Spieth, 2013; Foss and Saebi, 2017). Several scholars have studied the complex organizational change process of BMI and described the process phases in different ways with similar characteristics (e.g. De Reuver et al., 2013; Frankenberger et al., 2013; Geissdoerfer et al., 2017; Wirtz and Daiser, 2018). For this study, it is advisable to choose a simple BMI stage model as the basis for exploring group interaction during decision-making, since the focus of this study is on group interaction during collective BMI decision-making and not on the BMI process itself. We therefore decided to apply the 4I framework of Frankenberger et al. (2013), who divide the BMI process into four steps: Initiation, Ideation, Integration and Implementation. At the beginning, an internal or external change triggers the BMI process (Kraus et al., 2023) and prompts the innovating company to understand and reevaluate the environment and stakeholders. Based on emergent internal or external opportunities and threats, the responsible organizational members must collect and generate ideas for the new or adapted BM in the second process phase (Frankenberger et al., 2013). In the third phase of the process, a new and viable BM is developed based on the most promising idea from the former stage. In this third phase, the company is also asked to integrate the interests of other external stakeholders such as suppliers and partners. Finally, the fully conceptualized new BM is integrated internally and externally, which usually involves high investments and risks that may create internal or external resistance (Frankenberger et al., 2013).

The design of new and the redesign of existing BMs is associated with uncertainty (Casadesus-Masanell and Ricart, 2010), ambiguity and risk (Schneckenberg et al., 2017). The literature distinguishes between decisions under risk and decisions under uncertainty (Luan et al., 2019). While all information is available for decisions under risk and the probability of all possible future states can be predicted, decisions under uncertainty imply imperfect knowledge and insufficient information. For decisions with great uncertainty, Gigerenzer and Brighton (2009) recommend the use of heuristics instead of complex management strategies, which are more appropriate in risky situations when past data can be used to develop multiple solutions. Some scholars suggest visionary approaches to planning processes that emphasize prediction (see Wiltbank et al., 2006), while others advise adaptive approaches that focus on experimentation and learning (Sarasvathy, 2001) and are better suited for decision-making in uncertain environments (Alvarez and Barney, 2005). Björkdahl et al. (2022) add to this debate by applying a problem-based view to search for a viable BM. They distinguish between backward- and forward-looking searches and concluded, based on three cases that when firms perceive potential failure to continue with their established way of doing business, they apply a forward-looking search. Conversely, executives shift to backward-looking search when they perceive high alternative costs. Hence, an organization's decision logic must adapt to dynamic changes in uncertain environments (Reymen et al., 2017). However, the dynamic relationships between BM development and decision logic remain poorly understood (Andries et al., 2013; Reymen et al., 2017).

Board members can make flawed BM decisions due to limited individual expertise and cognitive shortcuts (Tversky and Kahneman, 1974; Wu et al., 2018), so coping mechanisms are necessary to overcome the mental lack of insight and delayed action (Schneckenberg et al., 2017). As strategic decisions are associated with various risks (Reymen et al., 2017; Brillinger et al., 2020), BMI decisions are usually made in groups with multiple decision-makers (Maciejovsky et al., 2013) to increase the amount of information and mitigate individual cognitive biases. In the next section, we show how groups make decisions together and what advantages and disadvantages group decisions have compared to individual decisions.

2.2 Individual information processing, group decision-making and group biases

Leaders develop unique cognitive frameworks through which they process information on the basis of past experience and learning (Karhu and Ritala, 2020). For this, managers mentally frame information to make sense of information (Walsh, 1995). Interpretation of information is highly individual (Heubeck and Meckl, 2022). Individual mental models determine “how a given problem or decision is perceived” (Karhu and Ritala, 2020, p. 490). Kahneman (2012) and Walsh (1995) described two distinct ways to process information: the automatic processing mode and the controlled processing mode. In the automatic processing mode, people examine information at a superficial level, drawing on past experience in similar situations. Conversely, the current informational context shapes the controlled information processing mode. It is most applicable in new contexts for which decision-makers do not have prior knowledge patterns (Kahneman, 2012; Walsh, 1995). Heubeck and Meckl (2022) recently have shown that both managers’ human capital (e.g. managerial knowledge, competencies and capabilities) and social capital (e.g. sympathy, trust, reciprocity) decisively shape how managers cognitively assess BMI alternatives. Managers with higher levels of human capital will be more aware of the opportunities and risks of BMI than those with lower levels of human capital. As a result, managers with higher levels of human capital are more likely to resort to the controlled mode of information processing than their counterparts with lower levels. In line with Alguezaui and Filieri (2010) and Manev et al. (2005), Heubeck and Meckl (2022) confirm that also social capital facilitates information sharing and decision quality. By enhancing trust, cooperation and collaboration across the firm, managerial social capital enables greater breadth and depth of information. This richness of information enables to assess BMI alternatives more consciously and to challenge existing mental models. Consequently, executives with higher social capital are more likely to operate in the controlled information processing mode (Heubeck and Meckl, 2022).

While these findings contribute to understanding how information is processed and interpreted at the individual level, we need to understand how these individuals synthesize their information at the collective level (Heubeck and Meckl, 2022) in decision-making processes. Group decision-making is advantageous because the expertise of an individual is limited whereas the collective expertise of the group is beneficial (Wu et al., 2020). The quality of individual decision-making decreases when individuals misjudge a situation or their abilities, when they are overwhelmed, or when they do not consider enough information and alternatives (Jones and Roelofsma, 2000; Bang and Frith, 2017). Due to the described subjective managerial cognition, the knowledge and skills of all decision-makers are required to select the best option. Groups have a higher cognitive capacity because of more experience and access to more alternative solutions offered by the management team. Groups learn faster and adapt better to the organizational strategy than individuals (Maciejovsky et al., 2013). Some studies show that groups can analyze complex problems more accurately due to the diverse experiences and information (Minson et al., 2011; Klein and Epley, 2015) as group members discuss and interact.

However, there is also evidence in the decision-making literature that groupthink can lead to a premature search for agreement due to stress and anxiety, resulting in poor information search, synthesis and evaluation (Chapman, 2006). In addition, groups have limited expertise, experience and individual skills (Wu et al., 2018), so even when using the recommended discussion procedures (Brodbeck et al., 2007; Lu et al., 2012), incorrect decisions may be made (Bang and Frith, 2017). Many times, individuals have an initial preference before the group decision is made that represents their preferred decision alternative based on their own experience or information (Stasser and Titus, 2003). Individuals often stick to this initial preference and seek information to confirm it, although there are gender differences (Nicholson et al., 2022). Brodbeck et al. (2007) confirmed this confirmation bias at the group level, with most group members preferring the initial majority preference. Based on groupthink theory (Janis, 1982), groups strive for consensus while inadequately sharing and evaluating information, leading to false overconfidence (Coffeng et al., 2021). Due to the challenges of collective decision-making, common group biases that are detrimental to good decision-making have been identified in the literature. In this study, the effects of group interaction in strategic group decisions are examined as summarized in Table 1. In conclusion, there is a need to identify which of these pitfalls may occur in groups specifically in the strategic BMI context so that boards can more effectively exploit business opportunities.

2.3 Individual risk willingness and decision-making in the BMI context

Risk-taking is an essential prerequisite for entrepreneurial orientation (Asemokha et al., 2021). Sitkin and Pablo (1992) identified several individual characteristics that predict risk-taking in decision-making such as individual risk willingness or interaction with other group members. Since risk perception and risk behavior are subjective, individuals evaluate the same situation differently (March and Shapira, 1987). Various risk behaviors result from different risk perceptions, individual characteristics, decision contexts and domains (Ferrey and Mishra, 2014). Situational differences arise from different decision domains (Figner and Weber, 2011; Weber et al., 2002). Two opposing philosophies of risk behavior dominate the literature. In the prospect theory proposed by Kahneman and Tversky (1979), managers engage in risk-taking behavior when confronted with a potential loss during decision-making. In contrast, Staw et al. (1981) predicted risk-averse reactions based on their threat rigidity theory.

Individuals differ not only in their risk perception and risk behavior, but also in their willingness to take risks (Arslan et al., 2020). A person's risk appetite can range from risk-averse to risk-neutral to risk-willing. Regarding decision-making in groups, researchers disagree on whether groups are more risk averse than individuals (e.g. Baker et al., 2008) or vice versa (e.g. Zhang and Casari, 2012).

In the context of BMI, little is known about the role of risk aversion. Few articles have dealt with this aspect. However, risk willingness has a decisive influence on whether a group decides to modify an existing business opportunity or to explore new ones. Tykkyläinen and Ritala (2021) consider risk aversion as an important boundary for BMI in social enterprises. Chesbrough (2010) and Laukkanen and Patala (2014) emphasized that risk aversion is one of the main barriers to successful BMI. Therefore, emerging business opportunities may be forfeited due to an insufficient willingness to face the emerging risks. In contrast, Leppäaho and Ritala (2022) find that risk-averse family firms, triggered by crises, may temporarily exhibit risk-taking innovation behavior. However, we still do not know the details of how individual and group risk-taking behavior changes when managers interact. This study aims to shed light on this blind spot by examining individual risk willingness and collective risk-taking and their impact on strategic BMI decision-making through manager interaction.

2.4 Synthesis of the research areas

Since this study aims to link several different fields of research, it is necessary to synthesize multiple viewpoints then create a theoretical model. Our conceptual model is presented in Figure 1. To link the different research areas, we proceed from top to bottom:

The BMI process is triggered by an internal or external opportunity or threat (Kraus et al., 2023) that initiates the adaptation, redesign, modification or change of an existing BM (top left, 1st level of analysis; Lindgardt et al., 2009). As part of the BMI process, all nine elements of a BM (Osterwalder, 2004) must be determined (2nd level of analysis). At the beginning of the BMI process, the board decides in principle whether it intends to renew the existing BM. The board members interact and discuss what value the organization wants to create with the new BM (3rd level of analysis). In this process, the executive team can contribute (e.g. more information and experience) but also suffer from potentially inhibiting interaction of its members (e.g. group biases and poor information search). A particular focus of our study is individual risk willingness (bottom center, 4th level of analysis). Since BMI decisions have a strong strategic impact, they are typically not made by a single person but by the entire board. Board members differ in their level of individual risk willingness (risk-willing, risk-neutral, or risk-averse board members). As the board members interact, the management group is characterized by its own group risk willingness. The question is therefore whether the board is more or less risky due to internal social interaction and whether the board makes more or less risky BMI decisions than an individual board member? We expect the riskiness of the group to affect the redesign and final content of the nine BM elements. For example, we expect a risk-willing board to be more likely to seek out new customer groups than a risk-averse board. From a financial perspective, we expect a risk-averse board to be more conservative and cautious with its revenue streams than a risk-willing board would be. The distinction between the nine BM elements provides a more detailed schema of questions that need to be answered in collective BMI decision-making.

Bringing together the different areas of research, we can derive three research questions for this study:

• RQ1. What group effects are prevalent in BMI group decision-making?

• RQ2. What are the key characteristics of BMI group decisions?

• RQ3. What are the potential relationships between BMI group decision-making and managers' risk willingness?

3.1 Recruiting and sample

First, three interviewers received information about the research topic, objective and research questions, as well as training on the process of qualitative interviews. Second, the interviewers reviewed the websites and media reports of the German Commercial Register and the German Engineering Federation to identify potential interviewees from companies that had changed their BM. An explanatory document about the research project was attached to an initial e-mail sent to potential interviewees to solicit participation and gain the necessary background information from the interviewees. The first contact was followed up with a second contact to potential interviewees who agreed to participate. The crucial requirement for the interviewees was their leading position in the company (managing director, top manager, department, or team leader). Since the aim of the study is to analyze group biases in the mutual BMI decision process, at least two executives of a company's decision-making body had to be interviewed. This process yielded a sample of 26 interviewees from thirteen companies.

Regardless of industry and firm size, BMI may be relevant to all firms when internal causes and environmental changes require adaptation or reconfiguration of the existing BM. Therefore, the sample includes firms from different industries and firm sizes to bring together different perspectives on this topic. The sample includes six small (SM), five medium (ME), two large (LA) and one micro firm (MI). The majority (ten) of the thirteen companies belong to the manufacturing sector. The other companies are from the mobility, health and energy sectors. The sample includes two female and 24 male respondents, all of whom hold senior positions with an average age of 44.9 years. The sample is appropriate because all participants are directly involved in making strategic BMI decisions at their company, including decisions about developing and marketing of existing and new products and services, adjusting revenue streams, costs, resources and the company's value architecture. The acquisition of interview partners and data collection were discontinued once theoretical saturation had been reached. In particular, the coding of interview numbers 25 and 26 did not yield any new categories. Table 2 shows all sociodemographic information about the interviewees.

3.2 Data collection

This exploratory study includes data from (1) sociodemographic questionnaires and (2) semi-structured interviews with 26 participants in German, collected between December 2020 and April 2021. The sociodemographic questionnaire was sent to the participants before the interviews were conducted. The questionnaire requested data on the company (industry, number of employees, number of hierarchical levels), individual characteristics (e.g. age, gender, education, position in the company) and self-assessed individual risk willingness. Since both the context and the characteristics of the decision-maker influence individual risk willingness (Blais and Weber, 2006), situational differences were considered in the questionnaire (general, financial and professional risk willingness). Importantly, individual risk willingness must be part of the written questionnaire to avoid external influences by the interviewer.

Subsequently, 26 semi-structured interviews were conducted with an average interview duration of 30 min. The open-ended interview questions were developed after several discussions by a team of two researchers based on established literature. Questioning consisted of two parts. Part A explored board members' strategic BMI decision-making behavior (e.g. describing the typical strategic BMI decision-making process). Part B explored interviewees' perceptions of group interaction and emerging group bias (e.g. individual decision-making behavior vs collective decision-making behavior, influence of other decision-makers, desire for unity). The questions in Part A build on the existing BMI literature, in particular the BM canvas by Osterwalder (2004) and the BMI process description by Frankenberger et al. (2013). Part B draws from the literature on group decision-making and group bias. To examine the advantages of group decision-making, we drew on existing evidence from the studies of Jones and Roelofsma (2000), Klein and Epley (2015), Maciejovsky et al. (2013), Minson et al. (2011) and Wu et al. (2020), for the disadvantages of group decision-making, we relied on established work by Brodbeck et al. (2007), Chapman (2006), Coffeng et al. (2021), Lu et al. (2012), Stasser and Titus (2003) and Wu et al. (2018). The questions on group bias were based on Bang and Frith’s (2017) summary of key strategic group biases, particularly the literature that has identified causes and symptoms (e.g. Huang et al., 2019; Jones and Roelofsma, 2000; Montecinos-Pearce et al., 2020; see Table 1 for the full list of authors). Finally, the interview guide was revised and adapted in an iterative process to ensure consistency and comprehensibility.

3.3 Description of the managers' risk willingness

Participants were asked to self-assess how risk-willing they are in different contexts on a 7-point Likert scale (see Table 3), based on the established Socio-Economic Panel questionnaire (SOEP-IS Group, 2019). Response categories “1,” “2,” and “3” reflect a risk-averse person, “4” a risk-neutral person and “5,” “6,” and “7” a risk-willing person. Such a self-report scale is suitable for assessing individual risk tolerance (Dohmen et al., 2011). Most participants indicated their individual risk willingness differently depending on the three contexts (general, financial and professional risk willingness). Only seven respondents (Int_1.1, Int_2.1, Int_3.1, Int_3.2, Int_4.2, Int_5.1 and Int_6.1) estimate that their risk willingness does not change depending on the context. On average, the sample is slightly risk averse in general (4.61) and in the professional context (4.65); in the financial context, the managers surveyed are risk-neutral (3.96). Spearman Rho confirms a highly significant positive correlation between general and professional individual risk willingness (0.897) therefore we used the average of the two values for further analysis. The data show that in only nine of the 26 responses did the self-reported risk willingness (general context) match the perceived risk willingness of the colleague. For example, Int_2.1 and Int_2.2 agreed that Int_2.2 was risk-willing. This independent agreement was also evident between Int_4.1 and Int_4.2 and vice versa, Int_5.1 and Int_5.2 and vice versa, Int_7.1 and Int_7.2 and vice versa, between Int_9.1 and Int_9.2 and between Int_13.2 and Int_13.1. The majority (15 respondents) rated their risk willingness (general context) the same as the perceived risk willingness of the entire board.

3.4 Data analysis

This study was a thematic analysis following Gioia's guidelines (see Figure 2). The iterative approach allows for the identification of themes and connections between research domains at the organizational and individual level based on empirical data (Braun and Clarke, 2006; Gioia et al., 2013). After transcribing all 26 interviews verbatim, the data was analyzed in three steps. First, two researchers independently open-coded the data using MAXQDA 2022 software, resulting in two coding schemes. Then the two coding schemes were merged and analyzed for existing and missing coding overlaps. Second, all codes from the merged coding scheme were extracted, printed out, then analyzed to identify which codes matched thematically and structured. We then categorized related codes into more abstract themes and iteratively refined the themes by reevaluating the data. The iterative process resulting in the emergence of eight main categories. Validity was maintained by comparing the categories with the existing literature. Third, we assessed the relationships between these eight main categories resulting in the emergence of four overarching themes. We compared the overarching themes again with the data to verify that they were appropriate and to check whether further categories were needed. Figure 2 shows sample quotes from the open coding process, the eight main categories and the four final overarching themes.

4.1 Managers’ group interaction in BMI decisions

To examine which group effects prevail in BMI group decision-making, we compared the developed categories with literature on symptoms and causes of group biases. The results show that BMI group decisions of managing directors are mainly influenced by the symptoms of the following four group biases:

4.2 Development of a BMI decision matrix

The BMI decision matrix, which defines the key characteristics of BMI group decision-making emerged from the interviews. Eight categories were identified that merge into the four overarching themes (see Figure 3). Of the four overarching themes, two form an inductively derived axis: Axis 1 (y-axis) spans the range for group interaction between managers in BMI decision-making (overarching themes: managerial cohesion and conflict readiness). Axis 2 x-axis) represents the range decision bases of BMI decisions (overarching themes: information-based and emotion-based decisions). Based on these two axes, a BMI decision matrix emerged that encompasses the key characteristics of board members' BMI decisions. Figure 3 shows how BMI decisions can be grouped in the matrix. The numbers indicate how many of the 26 managers interviewed displayed signs of each overarching theme.

4.3 Relating individual risk willingness, group interaction and BMI decision base

We now answer the last research question by explaining how managers’ individual risk willingness and BMI decision-making are related. We mapped the individual risk willingness in our proposed BMI decision matrix as shown in Figure 4.

Contrary to expectations, the interviews did not indicate that social interaction changes the decision-makers’ risk willingness in the decision-making process. The interviews revealed that risk-willing decision-makers were not able to persuade other members of the decision team to make riskier decisions. Therefore, we cannot confirm that boards make riskier decisions than individuals (risky shift) and vice versa (cautious shift), as reported by Zhang and Casari (2012). Rather, our results show that additional information available to risk-averse decision-makers reduces the perceived uncertainty. Thus, if risk-averse decision-makers have gathered enough information, they may be willing to pursue a riskier option. Therefore, risky decisions appear to be more a function of the quantity and quality of available information rather than individual risk willingness.

The results also show that risk-averse, risk-neutral and risk-willing managers do indeed exhibit different decision-making behavior, even if the three risk willingness types hardly differ in terms of the basis of their decisions. Table 4 and Figure 4 illustrate that risk-willing managers have the highest desire for managerial cohesion and risk-averse managers the lowest. The results on willingness to engage in conflict show the greatest differences with risk-averse managers usually more willing to accepting conflict. Risk-averse managers reported that they question arguments intensively and then give their board the opportunity to refine solutions. Risk-averse managers perceive the need for managerial cohesion to be lower because their priority is a solid overall result for the company. They therefore seem to play an important role in the long-term internal acceptance of BMI’s decisions. Only if every team member understands the importance and background of a decision will everyone stay motivated and implement the BMI decision with vigor. The risk-neutral managers with the least willingness to engage in conflict are those in the team who mainly try to avoid confrontation. They seem to mediate between risk-averse and risk-willing managers by driving the process of weighing up the pros and cons.

5.1 Managerial cohesion and conflict readiness as important counterparts for effective BMI decision-making

The results show that good BMI decisions are made by balancing managers’ individual conflict readiness with managerial cohesion. Although four group biases indicate managerial cohesion in BMI decisions, conflict readiness is perceived as most relevant and, interestingly, is also predominantly perceived positively to improve decision quality. We think that conflict is beneficial for effective BMI decision-making. Effective BMI decisions require both individual conflict readiness and managerial cohesion, because the mere presence of conflict readiness without managerial cohesion would prevent a common consensus, which is essential for timely decisions.

Managerial cohesion occurs when managers prioritize group harmony over making the right decision (Park, 2000). Mullen et al. (1994) explained that managerial cohesion reduces conflict in groups. Conversely, Manata and Bozeman (2022) have recently shown that conflicts in groups strongly cause managerial cohesion. They measure only a weak opposing influence of managerial cohesion on conflict in groups. However, we think, if peer pressure limits an individual's confidence to pursue an alternative solution the resulting overly cohesive leadership could prevent good decisions. Peer pressure would encourage groupthink in the form of premature concurrence seeking due to stress and anxiety, as Janis (1982) also warns. We therefore also agree with Chapman (2006) that the premature concurrence seeking due to groupthink would promote inadequate information search and consequently inadequate evaluation of different solution alternatives.

Two reasons for groupthink that produces managerial cohesion are identified in the literature: either a pessimistic assessment of a group's ability to successfully solve a problem, which manifests itself in a defensive response in the form of “collective avoidance” (Chapman, 2006) or an overly optimistic assessment of the group's ability to solve a problem which results in an unrealistically high confidence for success (Whyte, 1998). Our results overlap with those of Whyte (1998) and indicate an overly optimistic assessment of business opportunities by board members. This is problematic because, first, overly optimistic assessments of the situation lead people to take excessive risks in strategic decisions (Whyte, 1998). Second, low-level skilled people tend to be more self-confident and perceive themselves as more competent than they really are, while competent people tend to underestimate their abilities, as described in the Dunning-Kruger effect (Kruger and Dunning, 1999; Ehrlinger et al., 2005). This paradox highlights a critical deficiency in human cognition which is particularly dangerous at the executive level: the gap between actual expertise and perceived competence. This phenomenon stems from two core aspects: perceiving “known unknowns” and “unknown unknowns” which relates to metacognition, a unique human trait of self-reflection. Many individuals lack this metacognitive capacity making them blind to the degree of their ignorance (Huang, 2013). This ignorance is especially pronounced in those with excessive self-perceptions of their abilities. Such individuals struggle to critically evaluate information and their own performance, leading to biased decision-making (Nold and Michel, 2022). Huang (2013) added that people in authority often make decisions without adequate knowledge, failing to notice crucial information, or overestimating their expertise. Understanding the extent of this discrepancy among incompetent individuals is key to understanding the roots and consequences of biased decisions at managerial level. Nold and Michel (2022) compare the behavior of managers and employees. They show that managers not only overestimate their own abilities, especially in the dimension of success, but they underestimate the abilities of their employees which makes them blind to relevant information from their employees.

However, we observed that boards need managers' approval to solve problems together as a top management team. This is consistent with Festinger (1950) who stated that group consensus is essential to the effectiveness of boards. Honest dialog and attempts to convince other board members should help prevent disagreements. However, we also found in the interviews that social norms may trigger the premature pursuit of agreement. In this sense, we agree with McCauley (1989) and caution that social norms may favor developing friendly relationship over a critical evaluation of ideas. Thus, although collective efficacy has a positive impact on team effectiveness as recently shown by Elms et al. (2022), social norms could lead to an undesirable BMI decision. For further elaboration of the complex relationships between managerial cohesion and organizational performance, see Mullen et al. (1994) and Banwo et al. (2015).

The interplay of managerial cohesion and conflict readiness in BMI decision-making shown in this study also overlaps with the definition of leadership unity by Doz and Kosonen (2010). Along with strategic sensitivity and resource fluidity, leadership unity (also labeled as collective commitment, de Diego Ruiz et al., 2023) is one of the three firm-level abilities that constitute strategic agility, which contributes to higher firm performance (Doz and Kosonen, 2010; Hock et al., 2016). Leadership unity enables bold decisions to be made quickly through mutual support from the top management team (Doz and Kosonen, 2010). Leadership unity among senior executives helps organizations cooperate and jointly commit to risky decisions necessary for BMI implementation (Clauss et al., 2021). Our interviewees made it clear that it is essential to efficiently distribute information among the management team and to stimulate constant dialog to achieve mutual commitment (Clauss et al., 2021).

5.2 Information-based dominates emotion-based BMI decision-making

The respondents stated that they make decisions based on information rather than emotion. We have our reservations as to whether this is always the reality. Nevertheless, the respondents’ perception is consistent with Coffeng et al.'s (2021) statement that the main task of boards is to carefully weigh the pros and cons of different strategic options. All relevant actors need to gather and analyze comprehensive information (Acciarini et al., 2021) both rationally and intuitively (Calabretta et al., 2017). We can apply Simon’s (1987) traditional rational and intuitive perspective on information sharing in decision-making to our study. While information gathering and analysis reflects the rational perspective of strategic BMI decision-making, emotion-based decision-making represents the intuitive perspective. In this context, we point out the connection between intuition and emotion in complex decisions as described by Matzler et al. (2007). In the literature, the terms “intuition”, “emotions” and “gut feeling” are used very closely together. Hence, a clear distinction between the terms is difficult. However, our results fit the findings of Andersen et al. (2022) and Korherr et al. (2022) who also observe a shift from intuition-driven to data-driven decision-making processes and that many companies now want to be guided less by their gut feeling and more by data. Andersen et al. (2022) literally describe that today decision-makers struggle with this shift ‘For some companies, it is a “crossroads” decision in terms of balancing when to follow intuition and when to follow data (Andersen et al., 2022, p. 30).

Our respondents rated overly emotion-based decision-making as an inappropriate, unfavorable decision strategy. Indeed, some scholars characterize emotions as undesirable biases (e.g. Shiv et al., 2005). However, other researchers disagree and assert that strong emotions lead to higher decision quality (Seo and Barrett, 2007). Vuori and Huy (2022) recently investigated how emotions can contribute to higher decision quality. Their observation could also be relevant to our results. They explain that top managers re-evaluate strategic options using data-driven analysis through group interaction which can gradually change their emotions. In this way, top managers regulate their emotions and achieve higher decision quality and radical change. This explanation suggests that, as we show with our BMI decision matrix, an interplay of emotion- and information-based decision behavior is indeed required to make sound BMI decisions.

As identified in the literature (Van Knippenberg and Schippers, 2007), our respondents believed that informationally diverse boards gain more perspectives through constructive dialog, thereby increasing decision quality. Participants’ perception is intuitive and common knowledge but is inconsistent with previous findings (e.g. Lu et al., 2012; Schulz-Hardt and Mojzisch, 2012; Sohrab et al., 2015) showing that groups do not share information and knowledge sufficiently (He et al., 2021; Silva de Garcia et al., 2022; Goncalves et al., 2023) and therefore make biased decisions (Brodbeck et al., 2007). We know that due to confirmation bias, individual managers tend to search for information that confirms their existing beliefs and actually stick to their initial preference (Stasser and Titus, 2003; Coffeng et al., 2021; Nicholson et al., 2022). In our study, managers perceived themselves as reflective, although they did not clearly indicate that they were aware of confirmation bias, either at the individual or group level. From this, we conclude that managers may have had a false sense of security and overconfidence in the quality of their decisions. This assumption is consistent with the findings of Coffeng et al. (2021), who demonstrated this overconfidence in a hidden profile experiment with nonprofit organizations. Since our sample comprises 24 male and two female participants, we are not inclined to conclude any gender differences. Though, Nicholson et al. (2022) observe that after receiving full information, female decision-makers were more able to correct their initial suboptimal decision. Compared to their male counterparts, women also show lower overall confidence in their selection decision (Nicholson et al., 2022).

Our surveyed managers felt that faster consensus is better achieved when fewer options were discussed indicating the initial majority preference even among highly experienced decision-makers. As Wittenbaum et al. (2004) explained, fewer options facilitate quick agreements, which is necessary in dynamic environments and the resulting time pressure. This is in line with Wu et al. (2022), who explain in their paper entitled “Time pressure changes how people explore and respond to uncertainty”. However, we must raise awareness that quick agreements may have a negative impact on decision quality if the information is not thoroughly collected and analyzed. Although various discussion procedures to overcome confirmation bias in groups are recommended in the literature, they remain mostly ineffective (Sohrab et al., 2015; Coffeng et al., 2021). Therefore, future experimental studies should develop effective discussion procedures that incorporate time pressure, as already cautioned by Coffeng et al. (2021).

Depending on their individual risk willingness, respondents had different perceptions on how emotion-based they were in making strategic decisions. However, our study did not distinguish between positive and negative affect. Other researchers show that information processing is poorer when negative emotions are present and that anxiety may induce a tendency towards excessive risk-taking, while others predict more risk-averse behavior when a positive mood is present and vice versa (Chapman, 2006). Future studies could therefore examine the interrelationship between individual risk willingness and emotions in strategic decision-making.

5.3 Individual risk willingness as a success factor for BMI decisions

It is well known that entrepreneurial behavior usually requires a risk-willing attitude in order to seize market opportunities (Janney and Dess, 2006). Other authors argue that entrepreneurs sense opportunities because they perceive lower risk (Krueger and Brazeal, 1994) or that they proactively minimize risk (Bateman and Crant, 1993; Acedo and Florin, 2007). Even though studies showed that entrepreneurs are generally more willing to take risks than managers in established companies (Begley and Boyd, 1987; Stewart and Roth, 2001), managers are also called upon to accept a certain degree of risk to enable intrapreneurship in the sense of BMI.

While some authors claim that high risk-taking attitudes hinder innovation efforts (e.g. Brooks et al., 2023), others state that a positive attitude to risk taking, coupled with effective risk management, can enhance innovation performance (Lendowski et al., 2022). Laukkanen and Patala (2014) suggested that an overly negative attitude towards risk hinders successful implementation of BMI. Our results clearly contradict this conclusion. Surprisingly, our interviewees negated the previous negative framing of risk aversion by evaluating it as a valuable “critical voice” for successful BMI decisions. This observation fits with the philosophical nature of distributive justice theory, more specifically Rawls’ (1971) notion that the more risk averse a person is, the more benevolent they are. We could imagine that board members may evaluate the cautious decision-making behavior of risk-averse colleagues as a protective shield against demonstrably wrong decisions. This argumentation fits with the findings of Brennan et al. (2008) whose experiment provided evidence that consideration of other-regarding concerns when monetary payoffs are common knowledge. Even if the strategic BMI decisions affect board members' pay at some point in the future, the potential monetary payoffs are not yet known at the time the BMI decisions are made. We therefore assume that board members are referring to the current, already known payouts and not to the potential payouts in the future.

We have found that risk-averse decision-makers who lack information perceive higher uncertainty than risk-willing decision-makers. As a result, risk-averse individuals are highly engaged in information gathering. We believe that this behavior is beneficial for successful BMI decisions because it increases the quality of the decision rather than the speed of the decision. This is consistent with the findings of Acedo and Florin (2007) who found evidence that risk perception in the context of strategic opportunities is influenced by an individual’s prior experience related to the opportunity, their cognitive style of information acquisition and analysis and their tolerance for ambiguous situations. Song et al. (2022) further discovered that differences in risk perception among team members can positively influence information adoption with learning willingness playing a mediating role. While Song et al. (2022) emphasized the positive impact of diverse risk perceptions on information adoption within teams, our findings extend this by demonstrating that individual risk aversion in decision-makers leads to a more thorough and quality-focused information gathering process, crucial for effective BMI. The desire for information by risk-averse managers observed in this study also confirms Kogan and Wallach’s (1964) view that risk-averse decision-makers consider a broader range of information (conjunctive decision behavior) than those with a higher risk willingness who consider only a narrow range of criteria (disjunctive decision behavior).

The risk-averse decision-makers do not appear to be the stubborn ones in the team who hinder change, as initially expected. Rather, they require good arguments as to why the BM should change and why they should accept the risks associated with the BMI. In doing so, they proactively try to increase the amount of information on which to base a decision. In the interest of business development, risk-averse managers are more inclined to welcome constructive dialog than their risk-neutral and risk-willing colleagues (1) if they are firmly convinced that their decision option is “really” the best one for the company's success, (2) if the other team members cannot convince them with strong pro arguments for their proposal and (3) if they perceive the counter arguments for the others' options to be very strong. Therefore, the perceived uncertainty by risk-averse individuals seems to have a positive effect on the outcome of the BMI decision because they wisely weigh pros and cons until all doubts are removed. This behavior is indeed beneficial as it can prevent board members from overestimating their business opportunities in terms of BMI.

We can also interpret the information-seeking behavior of risk-averse board members as an ex ante risk management strategy. More specifically, our results suggest that risk-averse managers intend to reduce risk before they make risky decisions rather than seeking solutions to deal with the consequences of risky decisions. In this respect, we refute the paradox of risk-averse vs risk-seeking risk management behavior explained by Van Winsen et al. (2016). In contrast to our results, Van Winsen et al. (2016) concluded for their sample in agriculture that risk-averse managers are less likely to adopt ex ante risk management strategies and more likely to rely on ex-post curative measures.

The role of risk-willing managers in strategic decision-making is complex and requires further investigation, as it may be influenced by various theories such as agency theory, behavioral theory and prospect theory (Hoskisson et al., 2017). Our findings shed light on this blind spot and contribute to the idea that risk-willing managers have the greatest need for managerial cohesion when making strategic decisions. We conclude from our results that risk-willing managers may need the security and strength of a team to support their preference for riskier decisions. They may assess decision-making in a group as a kind of safety net for riskier decisions and therefore feel stronger in a team than alone. While our interviews revealed that risk-willing managers promote managerial cohesion by emphasizing business opportunities and striving for unity, it is the risk-averse managers who show the highest conflict readiness. These findings contradict Lahno et al. (2015) who conclude that the degree of risk aversion per se has no correlation with conflict. However, they add that differences in risk attitudes between individuals can lead to interpersonal conflicts between them. We hypothesize that the risk-averse attitude is an important factor for the long-term survival of an organization, because risk-averse managers weigh the pros and cons extensively until uncertainty is sufficiently reduced. Though, our results also show that board members are not driven by the dynamics of a group discussion and get carried away quickly. However, risk-willing decision-makers might perceive this behavior as slowing down or preventing efficient BMI. Therefore, we cannot confirm risky or cautious shifts by the decision group as Zhang and Casari (2012) claim. Rather, individual risk willingness seems to be a stable companion in BMI decision-making.

6.1 Theoretical contributions

This paper enriches three areas of literature that are of great practical importance to business executives. First, this work contributes to the part of the BM literature that explains BM as a cognitive schema. While previous studies in this area of research have focused on how BMs are manifested by social interaction (Massa et al., 2017) we show how social interaction influences the design of a new BM. The empirical data suggests that social interaction determines the change of the collective schema of an existing BM to a new collective schema in the new BM. Our work shows that this cognitive change is primarily accompanied by two aspects of social interaction: managerial cohesion and board members’ willingness to engage in conflict.

Second, we contribute to BMI research that views BMI as an organizational process. Although several scholars have looked at BMI process models from different angles (e.g. De Reuver et al., 2013; Frankenberger et al., 2013), there has been too little known about the cognitive behavior and individual and team characteristics within these BMI process models (Spieth et al., 2023). Our research contributes to closing this gap by investigating how managing directors interact during the BMI decision-making process and how the board that collectively undergoes the BMI process can be characterized in general. To this end, we introduce the BMI decision matrix to the BMI literature which frames BMI decisions from two different perspectives: the social interaction perspective and the decision base perspective. In addition to managerial cohesion and conflict readiness we also shed light on the role of emotions and information as the board goes through the organizational change process of BMI. This study also adds value to the BMI literature from a methodological perspective because interview-based data has been lacking in this literature to date. Our 26 in-depth interviews with managing directors contribute to gaining deeper insights into the BM adaptation process (Saebi et al., 2017).

We contribute to the ongoing debate on the role of risk perception and risk willingness during BMI (see, e.g. Taran et al., 2013; Euchner and Ganguly, 2014; Brillinger et al., 2020). While previous BMI literature has emphasized that uncertainty and risk are a typical aspect of BMI, it has not clarified how managers deal with them during BMI decision-making (Brillinger et al., 2020). The present work sheds light on this blind spot. We show that managers' differing risk appetites influence their desire for board unity. Our work offers a new perspective on the relevance of risk aversion for BMI decisions by refuting the previous negative framing of risk-aversion for the implementation of BMI in an organization (Laukkanen and Patala, 2014). Instead, the results show that risk aversion represents an opportunity to improve the quality of group decisions and thus promote the successful implementation of BMI. Although we have known that BMI requires competent leadership, we lacked knowledge about the role of leaders in BMI and their interaction (Foss and Stieglitz, 2015; Bashir and Verma, 2019). In addition, the internal processes associated with organizational change, such as the BMI decision-making process, need further investigation (Torkkeli et al., 2015). We contribute to filling these gaps by providing insights into group interactions between managing directors during the strategic BMI process.

Third, this paper contributes to the literature on strategic group decision-making. To date, this line of research has lacked qualitative studies to understand complex strategic decision-making processes (Asemokha et al., 2021), especially on the effects of information sharing during debate (e.g. Coffeng et al., 2021). The managerial characteristics in decision-making processes have needed further investigations (Korherr et al., 2022). Our qualitative data sheds light on these blind spots by empirically investigating which group biases are prevalent in strategic BMI decisions in boards. Our results confirm previous studies that emphasize the high relevance of group confirmation bias even at the top management level (Stasser and Titus, 2003; Brodbeck et al., 2007; Coffeng et al., 2021). However, in the BMI context, we suggest that four group biases mainly influence BMI decisions: groupthink, social influence, hidden profile and group polarization. In addition, we have learned in this study that the degree of individual conflict readiness depends on managers' risk willingness. It is clear that particularly risk-averse managers accept conflict in BMI decision-making and thus play an important role in the long-term success of an organization. This finding offers a new perspective on risk aversion in innovation processes. Our study also offers a new characteristic for the classification of managers in strategic decisions and thus extends the current findings of Korherr et al. (2022). Future studies could further examine the opportunities that arise for organizations through risk-averse managers rather than confirming their negative role in an organization's innovation behavior.

6.2 Managerial implications

Regarding potential group biases in strategic BMI decisions, we conclude that managers should be aware of and pay attention to the following pitfalls: groupthink, social influence, hidden profile and group polarization. We confirm that executives prefer harmony to dissent (groupthink) and that they are influenced by the other executives with whom they interact. Consequently, managers tend to change existing attitudes to conform with others in the group (social influence). Our results also show evidence that executives predominantly focus only on information that is available to all board members (hidden profile) and their preferences reinforce each other (group polarization).

Although managers consider themselves highly reflective and think that they make decisions mainly based on information, previous studies have shown that due to their biased self-reflection they do not use all information (e.g. Kruger and Dunning, 1999; Nold and Michel, 2022; Sohrab et al., 2015). In fact, even highly experienced decision-makers do not always find the best solution, also due to too early agreements. An effective way to overcome the hidden profile effect is to create a supportive environment that allows for dissent. In the BMI context, this requires an entrepreneurial culture that encourages managers to explore and seize new business opportunities. In addition, managers should create structures that enable them to consciously absorb information from their employees in order to overcome their potential lack of self-reflection.

Some managers need more information to make strategic decisions than others. Managers are faced with the trade-off between fast decision-making to remain competitive or high decision quality. As decisions become increasingly data-driven, managers will be even more challenged to integrate technology to support decision-making. Certainly, business intelligence systems can help companies to select the most relevant data and pool information to increase the speed of decision-making needed to respond to threats or opportunities required to outperform competitors. Our findings align with recent works by Berg et al. (2023) and Kanbach et al. (2023) that understanding artificial intelligence (AI) technologies is indispensable to accelerate learning to support value creating BMI. We also agree with Haefner and Gassmann (2023) and expect that AI technologies will fundamentally change how organizations collect and process information in innovation processes.

Managers are the ones who are supposed to determine the right direction for their company. It is essential to pay attention to the diverse voices within the decision-making team. It is common knowledge that critical perspectives are indispensable for company success. We build upon Manata and Bozeman (2022) who demonstrated the positive impact of group conflict on group cohesion and that opposing perspective are not mere opinions but pivotal contributions that can significantly enhance the quality of BMI decision-making. Hence, by integrating various viewpoints managers can facilitate a more comprehensive and robust decision-making process.

In this sense, it becomes fundamental for managers to shift their perception of risk-averse colleagues. Often, those who are cautious in their approach are viewed as barriers to progress. However, we suggest that it is more productive to view risk-averse individuals as enablers of successful BMI implementation. They are not only “part of the game” but their caution and attention to detail is valuable in identifying potential pitfalls. Consequently, risk-averse individuals help create a viable BM that not only fosters innovation but also mitigates risks which increases the probability of success.

6.3 Limitations and future research

One of the limitations of the study is the limited sample of 26 respondents from four industries, all in Germany. Samples from other geographic areas or cultures may yield different results. Future work could extend the data by applying a survey design to quantitatively investigate the role of individual risk willingness on BMI. An interesting question here would be what type of individual risk willingness promotes different BMI strategies (e.g. exploitation versus exploration strategy). Furthermore, this study used a qualitative research design and thus reflects managers' perceptions of group interaction during strategic decision-making. Experimental studies could further explore the topic by developing and applying specific BMI decision-making tasks for board members. Future studies could also develop an approach that measures the actual ratio between the information- and emotion-based decision base. We also see the need for future studies to identify ways for overcoming or mitigating the Dunning–Kruger effect in strategic decisions.