Brand deletion: How the decision-making approach affects deletion success

This is the accepted version of the manuscript: Temprano, V., A.I. Rodríguez-Escudero y J. Rodríguez-Pinto (2018). “Brand deletion: How the decision-making approach affects deletion success”. BRQ Business Research Quarterly, vol. 21 (2), pp. 69-83, https://doi.org/10.1016/j.brq.2018.03.003 BRAND DELETION: HOW THE DECISION-MAKING APPROACH AFFECTS DELETION SUCCESS VÍCTOR TEMPRANO-GARCÍA ANA ISABEL RODRÍGUEZ-ESCUDERO JAVIER RODRÍGUEZ-PINTO Department of Business and Marketing University of Valladolid Avda. Valle Esgueva, 6. 47011 Valladolid (Spain) Corresponding author: VÍCTOR TEMPRANO-GARCÍA. Email: [email protected]va.es Financing: European Social Fund (ESF) and Junta de Castilla y León (Spain), ORDEN EDU/828/2014, and European Regional Development Fund (ERDF) and Junta de Castilla y León (Spain), project reference VA112P17. The authors declare no conflict of interest in this article. Acknowledgements: Authors want to express their immense gratitude to Prof. John Rudd (Head of the Marketing Group, Warwick Business School) for his insightful comments on earlier versions of this paper. We also thank the support of Corporate Excellence - Centre for Reputation Leadership in the data gathering process. 1 BRAND DELETION: HOW THE DECISION-MAKING APPROACH AFFECTS DELETION SUCCESS 1. Introduction After decades in which building a diversified portfolio of brands coupled with deep product lines was a customary practice in many firms (Keller et al., 2011; Morgan and Rego, 2009; Rosenbaum-Elliott et al., 2015), with the turnaround of the century we have witnessed a reversal in this trend. Thus, brand deletion (hereinafter BD), i.e., discontinuing a brand from a firm’s brand portfolio (Shah, 2013), in their different types, emerges as a strategic decision in the realm of brand portfolio management. Brands can be deleted through a brand name change. Examples of this move are the migration by Unilever from Starlux to Knorr, or the rebranding by the Santander banking group from Abbey in the UK or Banesto in Spain to the global brand Santander. In other instances, the deletion occurs through a total brand killing, so the company retires from the market both the brand and the product lines commercialized under that brand, as did General Motors (GM) when eliminating some emblematic brands, such as Hummer, Pontiac or Saturn. In the cases of Saab, Opel or Vauxhall, the deletion was made through brand disposals, as these brands were sold by GM and the new owners assumed their commercial use. Disposal was also used by the Danone Corporation to delete brands such as San Miguel, in the beer market (Barrett, 2000), or Lu and Príncipe, in the biscuits market (Jones, 2007). These examples are not exceptional and serve to illustrate how more and more companies from diverse manufacturing and service sectors are embarking on BD strategies. Three main world-wide environmental factors have impelled BD and led many firms to focus on competing more strongly through their most strategically important brands. Firstly, days of rapid growth drew to a close for many industries and 2 companies, and efficiency became a priority (Hooley and Saunders, 1993), which forced many diversified corporations to get rid of businesses that were unrelated to their own core competences (Varadarajan et al., 2006). Secondly, market globalization facilitated significant cost reductions derived from outsourcing and relocating operations in emerging countries, but it also triggered substantially increased market competition. The reaction of many corporations has been to focus their efforts on fewer but stronger brands with a global presence, divesting from local or regional brands (Depecik et al., 2014; Özsomer et al., 2012). Thirdly, the rising market share of private label brands has further prompted manufacturers to compete with a smaller set of strong brands rather than with a larger set of weak ones, thus deleting underperforming and secondary national brands (Sloot and Verhoef, 2008). However, despite the strategic nature of the BD decision and the fact that many companies are suffering the environmental circumstances mentioned above, managers are not knowledgeable about how they should tackle the challenge to delete a brand, and scholarly research on BD is so scarce and fragmented that it is hard to talk of a cohesive body of knowledge on the subject. Very few researchers have conceptually or empirically addressed issues of BD. Table 1 summarizes the main contributions found in this emerging but relevant research area. As regards conceptual works, Kumar (2003) adopts a normative approach and outlines a set of recommendations guiding managers during the BD process. The other conceptual works are geared toward classifying the explanatory factors underlying the BD adoption propensity either in general (Shah, 2015; Varadarajan et al., 2006) or in multinationals (Ketkar and Podoshen, 2015). Shah (2017a) identifies the outcomes which could serve to define a BD as successful and suggests a set of decision and implementation factors which could have an impact on these outcomes. Empirical papers are also scant and 3 tend to focus on the outcomes of BD, either considering consumer evaluations as a performance measure (Mao et al., 2009; Mishra, 2017) or analyzing the impact on the firm’s value by looking at stock market reactions after the announcement of a brand disposal (Depecik et al., 2014; Wiles et al., 2012). These studies reveal factors which affect customer and investors reactions to a BD (e.g., type of BD, brand weakness, scope, relatedness to other brands in the portfolio), yet treat BD as an exogenous event. In other words, these studies do not explicitly consider the BD decision maker’s point of view and fail to examine why or how a BD decision was taken and how it was implemented. One notable exception is the qualitative research by Shah (2017b) and Shah et al. (2017) in which, based on grounded theory, a causal model of the BD strategy is proposed and several questions are explored such as the context, reasons and fit of this strategy, the types of BD, as well as its implementation and consequences. Even if we look at a related field of study, namely research into brand delistings –i.e., the retailer’s decision to remove an entire brand from its assortment, leading to the unavailability of the deleted brand within the retailer’s stores–, we see how this has been neglected in the literature, with empirical research having opted to focus on customer reaction to the delisting (Sloot and Verhoef, 2008; Wiebach and Hildebrandt, 2012). (Insert Table 1 here) Given the scarce and fragmented literature addressing BD and the importance of such a strategic decision vis-à-vis gaining or sustaining competitive advantage, further academic research exploring the factors which drive the success of BDs is clearly a must. The “black box” of how BDs are decided and executed must be opened. The present research is primarily concerned with the decision-making process. Based on the strategic decision-making literature (Child et al., 2010; Elbanna, 2006; Kester, 2011), 4 we propose a model in which we consider how three different approaches to decisionmaking –namely, rational, intuitive and political– are related to BD success. We define BD success as the extent to which the company is satisfied with the outcomes and has achieved the objectives established at the time the decision was made. The use of these decision-making approaches have already been object of research in the strategic management literature, but the specificities of the BD strategy makes relevant to scrutinize their effect on performance in this particular context. Deleting a brand is usually a controversial and emotionally charged process (Shah, 2017a). Divergent perceptions, opinions and feelings are likely to emerge during the process. Whilst the BD may be considered indispensable for some stakeholders, for others it represents a failure or an unnecessary breach in the company’s history. The extent to which rationality, intuition and politics drive (or are absent from) the decision-making can seriously affect the results from the deletion. Thus, our first objective is to empirically investigate how the way in which the BD is decided impacts on the outcomes of this decision. In addition, as outlined above, deleting a brand from a firm’s portfolio may be carried out in a number of different ways (i.e., killing the brand, selling it, changing the brand name), each of which entails differing benefits and risks and which may require particular approaches. Based on a contingent perspective of the decision processperformance linkage (Fredrickson, 1983; Miller and Friesen, 1983), the second objective in this research is to investigate the interaction between the diverse approaches to decision-making and the type of BD to be executed. Therefore, we examine whether the influence of these approaches on BD success varies depending on whether the brand is totally killed off or sold to another company or deleted through a brand name change. This research contributes to the strategic marketing and brand portfolio management literatures by addressing a managerially relevant and topical problem, namely, pruning a 5 firm’s brand portfolio, but which has thus far received little scholarly attention. We expand on the scant empirical research conducted to date into the results of BD and contribute to a better understanding of which factors under managerial control may render a BD more successful. More specifically, we delve more deeply into the BD decision-making process and analyze main and interactive effects of rational, intuitive and political approaches, which, alone or in combination, affect BD success. Furthermore, unlike prior quantitative research, which gathers experimental data from consumers concerning their reactions to fictitious BDs (Mao et al., 2009; Sloot and Verhoef, 2008; Wiebach and Hildebrandt, 2012) or which relies on secondary data and press reports (Depecik et al., 2014; Wiles et al., 2012), we have gathered information from managers about 155 real cases of recent BDs by companies in a range of different industries. Albeit in retrospect, this method did enable us to ascertain managers’ perceptions about the particular internal and external circumstances of each case of BD in our sample when the decision was made and implemented together with their evaluation of how appropriate each decision proved to be. 2. Theoretical framework The BD decision can be considered as a critical strategic choice because, in addition to being infrequent, non-routine and complex, it involves an important relocation of a firm’s resources, alters its brand architecture, seriously affects diverse stakeholders, and can have a major impact on long-term market and financial performance (Shah, 2015). In any case, this kind of decision must not be conceived as a stand-alone one-shot decision problem with a single optimal solution, but as a strategic decision-making process which different firms may approach in different ways. According to the strategic decision-making literature (Child et al., 2010; Elbanna, 2006; Kester, 2011), three dominant currents prevail: synoptic formalism, the incremental 6 perspective and logical incrementalism. Synoptic formalism emphasizes deliberate, effortful and analytical procedures such as formal planning activities, generating alternatives and evaluating quantitative data as well as all relevant information as a basis for decision-making in order to reach an optimal decision (Atuahene-Gima and Li, 2004; Wiltbank et al., 2006). Rationality is the representative trait of the synoptic model (Elbanna and Child, 2007a). On the other hand, the incremental perspective advocates strategic decision-making based on adaptation through a gradual and complex process of learning, which is intuitive in nature (Atuahene-Gima and Li, 2004; Fredrickson, 1984; Rajagopalan and Spreitzer, 1997). Finally, Quinn (1980) proposes the logical incrementalism as an approach that combines elements of rational planning and intuitive behavior, and incorporates a third component: political maneuvering based on power and social interactions. As a group, people may differ in preferences and interests regarding the decisions to be taken within the company (Quinn, 1980). In this common scenario, politics may have a major influence on the strategic planning process and its outcomes (Eisenhardt and Zbaracki, 1992). In sum, logical incrementalism states that decisions are undertaken based on rationality, intuition and politics. Thus, following the literature (e.g., Elbanna and Child, 2007b; Kester et al., 2011) three approaches may be at work in the strategic decision-making process: rational, intuitive and political (see Table 2). Rational decision-making is an approach characterized by an attempt to exhaustively gather the information relevant to the decision, and the reliance upon analysis of this information when making the choice (Dean and Sharfman, 1996). Intuitive decisionmaking refers to more incremental adaptations based on an intimate understanding of the situation (Eisenhardt and Zbaracki, 1992). It is a process based on experience, but is not necessarily biased and irrational, even if it may be difficult to articulate the reasons 7 underlying the decisions taken when following this approach (Khatri and Ng, 2000). Therefore, although it is sometimes contaminated with presumptions and naïve preferences, intuition does represent a complex psychological phenomenon which helps in problem-solving by drawing from the store of knowledge in our subconscious and is rooted in past experience. In a political approach, the key assumption is that organizations are coalitions of people with competing interests. People with conflicting preferences may employ political tactics so as to shape decisions in line with their preferences. Consequently, we define political behavior as intentional attempts to enhance or protect the self-interest of individuals or groups (Hickson et al., 1986). (Insert Table 2 here) The literature has acknowledged that just one approach may not be sufficient to describe the complexity of the strategic decision-making process (Brews and Hunt, 1999; Eisenhardt and Zbaracki, 1992). As Hart and Banbury (1994) argue, strategic decisionmaking requires integrating both formal planning and incremental adjustments. Nutt (2002) contends that strategic decisions are made through the combined use of rational processes, judgment (intuition) and bargaining (politics). Our proposal is that rational, intuitive and political approaches to decision-making can coexist in a decision process, thus making it relevant to examine the specific impact each approach has on BD success, and even the interactive effects between them. Hence, as a first objective, hypotheses H1 to H4 of our research elaborate on how the different approaches to BD decision-making influence BD success (see Figure 1). As indicated in the introduction section, deleting a brand may come in a number of different forms: total brand killing (i.e., both the brand and the product lines commercialized under that brand are retired from the market), brand disposal (i.e., the brand disappears from the firm’s portfolio but still remains on the market because it is 8 sold to another company that assumes its ownership and commercial use), and brand name change (i.e. the brand is eliminated in order to sell the same –or very similar– products or services under another brand name or trademark of the same or from a different company). Both total killing and disposal entail determining an irreversible strategic decision since it usually involves the company getting out of the market and closing or at least downsizing one or more of its business units. Alternatively, when the BD is undertaken through a brand name change, the company reduces the number of brands in its portfolio currently used for commercial purposes, although it retains ownership of the deleted brand name and remains in the market. This form of BD is also strategic since it alters the company’s brand architecture and its branding strategy, but compared to total killings or disposals, brand name changes are less risky as they can be reversed or implemented gradually. Given the unequal consequences and risk of the different BD types, we argue that this variable conditions how rational, intuitive and political approaches to the decision might impact on the success of the deletion. Therefore, a second objective in our study is to explore the moderating effect of the type of BD on the relationship between the different approaches to BD decision-making and BD success. This objective is specified in H5 to H7 (see Figure 1). 3. Hypotheses development 3.1. The decision-making process and BD success Rational decision-making is the process by which firms use objective information and empirical evidence to build decisions. There is considerable consensus concerning the positive effect of this kind of decision-making approach and performance. For example, in the subject of NPD portfolio decisions, Kester (2011) and Dean and Sharfman (1996) find a positive relationship between rational decision-making and portfolio decisionmaking effectiveness. Similarly, in the strategic planning research area, Miller and 15 they did not wish to disclose any information concerning this type of decision or because the managers were too busy to comply with our request. As a means of exploring the manager’s point of view regarding the relevance of the variables identified in our literature review on BD decision-making, we conducted eight in-depth interviews with executives, five of whom worked in firms operating in service industries, and the other three in the manufacturing industry. As regards size, three of the interviewees were top managers in medium-sized companies and five in large companies. These interviews also served to refine and pre-test the questionnaire designed to gather data for the empirical analysis. The final version of the questionnaire, in which the unit of analysis is a case of BD recently carried out by the respondent firm, was sent to the 232 companies that agreed to participate, along with two letters of support by Interbrand and the Leading Brands of Spain Forum and a letter thanking them for participating in our research, and explaining the benefits of joining our research in terms of full access to the research findings. After a follow-up by telephone and personal visits to their offices, we obtained 155 complete questionnaires, provided by 111 respondent firms, yielding an effective response rate of 48%. Table 3 shows the sample characteristics. Respondents were asked about their direct participation in the BD decision and implementation as well as their knowledge of the reasons and facts surrounding the deletion. Mean scores for these questions were, respectively, 5.75 and 6.38 out of 7, indicating that the key informants in our sample are a valid source of information. (Insert Table 3 here) Sample representativeness was assessed as follows. We conducted a proportion test among the companies in the sample and in the population using the industry as the strata variable. The results in Table 4 show that the wholesale and retail trade sector is 16 significantly underrepresented in the sample, while the information and communication sector is significantly overrepresented. The reason might be the inclusion of a large number of wholesalers in the “Wholesale and retail trade” group (NACE code 45, 46 and 47). For these companies, the strategy of using the brand as an asset on which to base the value proposition is of little importance when compared to other industries. It is uncommon for wholesalers to own several brands (they sometimes own only one) and, when they do own them, they are used primarily for identification purposes rather than for differentiation. The decision to eliminate a brand rarely occurs, hence the low representation of this industry in the sample. In contrast, in the information and communication sector, knowing that Atresmedia, one of the leading private media groups in Spain, was taking part in our research had a snowball effect, encouraging other companies within this sector to also participate. (Insert Table 4 here) To assess the quality of the gathered data, we compared the correlation between the data on sales and employees extracted from the Amadeus database, and the data on sales and employees reported by respondents. The correlation for sales is .89, and the correlation for employees is .88, providing an indication of the reliability of the answers given by informants. In addition, following the recommendation of Armstrong and Overton (1977), we examined the potential influence of non-response bias by comparing early (33%) and late respondents (33%) via a t-test. No significant differences at p<.05 were found between the two groups regarding the constructs examined in this study. Since a single informant provided the data for each BD case, we also examined whether common method bias (CMB) could be an issue in our survey. We attempted to a priori minimize method bias by using some of the best practices described in the literature (Podsakoff et al., 2003; Rindfleisch et al., 2008). In particular, we protected respondent 17 anonymity and, as indicated above, ensured respondents were executives in a position to provide accurate information and opinions. In addition, item wording was carefully revised to prevent biased connotations, the dependent and independent variables were in separate parts of the questionnaire, and different scale formats were used (see Table 5). As post-hoc analysis, we used Harman’s one-factor test and results indicate that little common method variance (CMV) is observed in our data. According to Fuller et al. (2016), it is very unlikely that such a small CMV could substantially bias the estimated relationships. The observation of positive and negative construct intercorrelations (see Table 6) further supports the conclusion that the possible impact of CMB is minimal. 4.2. Construct measurement Measurement instruments are presented in Table 5. As stated before, the literature on BD is extremely scarce. Therefore, scales previously used in strategic decision-making literature were adapted to operationalize the three constructs measuring the firm’s approach to BD decision-making. In particular, we adapted the scales used in the research by Papadakis et al. (1998), Khatri and Ng (2000) and Dean and Sharfman (1996), which have also been used by Kester (2011) in the context of research into NPD portfolio decisions. We elaborated a new scale to measure BD success, which reflects the level of satisfaction of the company with the outcomes of this decision. The choice of BD success as the dependent variable provides for a close link between the strategic decision-making process and its outcomes as well as it avoids the causal ambiguity associated with more general indicators of organizational performance (Dean and Sharfman, 1996; Elbanna and Child, 2007a; Shepherd and Rudd, 2012). Please, note that a perceptual construct of BD success is used 1 . 1 In contrast with other alternative measures, such as economic performance, our scale of BD success is applicable to any BD decision, whatever the industry, the context or the reasons to make this decision. In 18 We incorporated the firm´s prior economic situation as a control variable since this can affect the greater or less urgency to accomplish the BD as well as the reactions and perceptions concerning its impact on company performance. We operationalized this control variable with a three-item scale adapted from Moorman and Rust (1999) and Verhoef and Leeflang (2009). Considering Varadarajan et al.’s (2006) proposition, we also controlled for the effects of the firm having previous experience in similar strategies, as it could be expected that the accumulation of relevant knowledge will positively influence performance (Finkelstein and Hambrick, 1996; Golden and Zajac, 2001). A single-item scale, adapted from Dayan and Elbanna (2011), was used to operationalize the firm’s experience in BDs. Finally, we controlled for the effects of formalizing the execution of the BD. Establishing standardized rules, protocols, deadlines and control mechanisms during the deletion process should help to ensure it is undertaken in an effective and timely manner (Argouslidis, 2008; Argouslidis and Baltas, 2007; Avlonitis and Argouslidis, 2012; Gounaris et al., 2006). Formalization was measured with five items adapted from the works by Argouslidis (2008), Argouslidis and Baltas (2007). (Insert Table 5 here) 5. Analysis and results The validity of the measurement instruments was assessed using the Partial Least Squares technique with the SmartPLS 3 software (Ringle et al., 2015). Reliability was examined by verifying that Cronbach’s α and composite reliability (CR) values were all above .70 and that average variance extracted (AVE) exceeded the recommended minimum of .50 (Bagozzi et al., 1991). As reported in Table 6, discriminant validity this sense, a BD can be considered successful/unsuccessful even though negative/positive economic results had been observed after the deletion. For example, if the BD was motivated by a lack of strategic fit with the corporate strategy, the success of the decision does not necessarily reflect in superior profits or sales figures, at least in the short term, but improved corporate reputation should be observed. 19 was assessed by applying the well-known Fornell and Larcker (1981)’s criterion, which provided satisfactory results, as well as the criterion recently proposed by Henseler et al. (2015). According to this new criterion, based on the HTMT (heterotrait-monotrait) ratio of correlations, discriminant validity is established when the HTMT ratios do not exceed the .85 recommended threshold and their 90% bootstrapped confidence intervals do not include the value 1, conditions that are clearly met in our survey. (Insert Table 6 here) Moderated hierarchical regression analysis with the SPSS software (v.23) was used to test the hypotheses depicted in Figure 1. As the nature of the main effects differs for models with and without interactions and may involve false and misleading conclusions (Henseler and Fassott, 2010), we sequentially introduced different blocks of variables to check their respective explanatory power. Firstly, in Model 1 we estimated the main effects of the focal and the control variables in our model. Secondly, in Model 2 we added the interaction effect between rational and political decision-making approaches (Model 2). Finally, in Model 3 we incorporated as predictors of BD success the twoway interactions between the three different approaches to BD decision-making and the type of BD (measured with a dummy variable in which a zero was assigned to BDs undertaken through total killing or disposal, and one was assigned to the cases of BDs through a brand name change). All model variables were mean-centered to avoid difficulties concerning interpretation of coefficients resulting from simultaneously including linear and interaction terms of the same variables in the same model (Echambadi and Hess, 2007). The standardized parameter estimates of the hypothesized and control relationships are presented in Table 7. (Insert Table 7 here) H1 predicts a positive relationship between rational decision-making and BD success. 20 H1 is confirmed since said effect is positive and significant (β=.19, p<.05). H2, which postulates that the relation between intuitive decision-making and BD success is also positive, is likewise supported (β=.20, p<.01). The relationship between political decision-making and BD success is negative and significant (β=-.20, p<.01), as hypothesized in H3. The results in Table 7 also show a negative and significant interaction effect between rational and political decision-making on BD success (β=-.13, p<.05) thus providing support for H4. The nature of this interaction has been examined using Aiken et al. (1991) procedure, which tests for the significance of regression coefficient estimates for the independent variables at one standard deviation below and above the mean of the moderating variable. At a low level of rational decision-making, the relationship between political decision-making and BD success is non-significant (β=-.09, n.s.), whereas at a high level of rational decision-making, a large and highly significant negative effect of political decision-making is found (β=-.32, p<.00). The moderating effects of the type of BD predicted in H5 and H6 are rejected since the interactions of this dummy variable with the rational and the intuitive decision-making variables are both non-significant. In contrast, H7 is supported by the data since a positive interaction effect is observed between type of BD and political decision-making (β=.13, p<.05). This means that the negative effect of political decision-making on BD success is less negative when the firm is changing the brand name than when it is killing or selling the brand. In order to further explore this moderating effect, we compared the effect of the political decision-making approach on BD success across the two different types of BD. Henseler (2012)´s multi-group analysis based on PLS (PLS-MGA) was used to perform this comparison since it relies on a nonparametric test that does not require any distributional assumption. Table 8 shows the results for this test, revealing a 21 significant difference across groups (p<.05) in the effect of political decision-making on BD success. In particular, results show that the negative effect of political decisionmaking on BD success is stronger (β=-.39 p<.01) when the brand was killed or sold to another company 2 than when the brand was deleted through a name change, where the impact becomes insignificant (β=-.05, n.s.). (Insert Table 8 here) 6. Discussion Rational, intuitive and political approaches, all of which potentially present in BD decision-making, are not equally recommendable because, whereas both rationality and intuition exert a positive effect on BD success, politics provokes a negative impact. There is no controversy in the literature surrounding the result of the positive impact of rationality in success. What is more noticeable, however, is the positive effect of the intuitive approach on BD success. Our findings are in line with the literature highlighting the benefits of intuition in the cases of strategic (or non-routine) decisions characterized by incomplete knowledge. For example, intuition can be brought in after rational processes have done the groundwork and can provide data and analyses concerning the BD as the basis for intuitive processes (Sauter, 1999). Intuition could be a form of intelligence which decision-makers can use when they cannot access rational processes or can be used simultaneously (Fredrickson, 1985; Parikh et al., 1994). In addition, when compared to rationality, intuition has the advantage of requiring less resources and being fast, whereas rationality struggles to deal with discrepant 2 A priori the cases of total brand killing and of brand disposal were categorized as a single group of more risky and irreversible BDs. As a robustness check, we run separate analyses for the 53 cases of total brand killing and the 18 cases of brand disposal and the findings are consistent. That is, the standardized parameter estimate of the relationships between political decision-making and BD success is negative and significant for total brand killings as well as for brand disposals. 22 information since this proves troublesome when determining the weight of such information (Bingham and Eisenhardt, 2011). In line with previous research, we find support for a negative relationship between political behavior and the outcomes of BD (e.g., Dean and Sharfman, 1996; Gandz and Murray, 1980). Nevertheless, the strength of the negative relationship between political decision-making and BD success is contingent on two variables. First, it depends on the level of rationality. When a firm relies on a rational decision-making approach and bases the BD decision on objective and comprehensive information, having to negotiate and make concessions to particular groups might be perceived as a deviation from an optimal choice and may cause a feeling of frustration among those having access to the data that proves the BD is necessary and urgent. However, when the BD decision is adopted without robust evidence of the convenience of this initiative, political behavior is easier to justify as a result of the lack of objective information. Second, in cases of brand name changes, political behavior does not play such a dysfunctional role as it does in cases of total brand killing or disposal. Political tactics such as negotiation or bargaining may prove appropriate vis-à-vis “selling the issue” and facilitating the path to implementing the changes required by the BD. Thus, this approach should not always be discarded since politics can serve as a mechanism for organizational acceptance of difficult decisions and for promoting the necessary strategic alignment (Eisenhardt and Zbaracki, 1992; Elbanna, 2006; Nutt, 1998). As Mintzberg (1998) points out, politics should be evaluated according to its effect on an organization’s ability to pursue the appropriate mission efficiently in the long term since private political interests do not necessarily come into conflict with the common interests of the firm. 23 7. Conclusion, managerial implications, limitations and future research This paper offers an important contribution to the scarce BD literature. From an academic point of view, this study builds upon the more general literature on strategic decision-making and provides evidence in favor of Quinn’s (1980) logical incrementalism. Compared to the synoptic formalism or the incremental perspective, which respectively emphasize the role of rationality and intuition, the logical incrementalism offer a more comprehensive and realistic view of decision-making as it acknowledges that rational, intuitive as well as political approaches are present and to some extent combined in the making of strategic decisions such as BD. In this sense, our investigation demonstrates that the way in which a firm approaches the decision to delete a brand from its portfolio affects the deletion outcomes and that all the three approaches are related to the perceived success of the decision. In particular, rational and intuitive decision-making approaches positively contribute to a successful BD. Therefore, this work adds empirical support to the mainstream of management research that defends a strategic decision-making process based on rational-intuitive assessment of information (e.g., Taggart and Valenzi, 1990). However, using a political approach exerts a negative influence on BD success. This negative effect is particularly harmful when decision-makers have objective information and empirical evidence to build a decision-making rationale but let the BD become contaminated by the possibly spurious interests of particular groups. Moreover, the negative effect is contingent on the type of elimination. The impact of the political approach is not statistically different from zero in the cases of BDs involving a brand name change. This probably occurs because, despite the generally detrimental influence of politics which is found, political behavior may well prove necessary and beneficial in terms of getting the BD decision accepted 24 by the different organizational stakeholders and facilitating its implementation as well as adaptation to the new situation. From a managerial point of view, the most important implication of our research is that managers have the power to influence the success of the BD. In this vein, we recommend that firms establish information systems which help decision-makers facing a BD decision to access the relevant data in order to decide accordingly and to provide evidence of the appropriateness of the strategy to be implemented. It is also necessary to make sure that management team members have the relevant expertise to rapidly and accurately assess the available evidence and to facilitate effective intuition, thus avoiding naïve intuition based on biased assumptions. As Khatri and Ng (2000) indicate, intuition can be developed through repeated exposure to the complexity of real problems, such intuition proving advantageous in the context of challenging and complex decisions such as a BD. In general, the political approach to BD decisionmaking should be minimized, particularly when managers have suitable information to base their decision on rational evidence. One exception to this general recommendation of minimizing political behavior is the situation in which the use of political tactics helps to overcome certain stakeholders’ initial rejection of the BD. This may be the case of BDs which simply involve a brand name change and where the business continues to operate in the market. In these instances, a political approach might not automatically lead to poorer performance. However, politics could jeopardize the success of the BD in cases involving total brand killing or disposal. The results of this study must be interpreted bearing in mind certain limitations. First, we have used subjective measures based on the perceptions of a single respondent. Thus, even though we have verified that the respondents in our sample are knowledgeable informants and that there are no indications of a substantial CMB, our 31 Mishra (2017) Quantitative To explore the consequences of brand deletion. C,D MBA students / Controlled experimental design. Evaluations of organizational performance depend on the strenght of the deleted brand, whether the brand is merged, sold or eliminated, and whether the firm communicates the logic of the deletion. Shah (2017b) Qualitative To explore the causes of BD in firms with a ‘house of brands’ portfolio. A Managers and archival data / Grounded theory. Brands are deleted because of financial factors, as well as because non-financial factors related to the consumers’ needs and preferences, the brand portfolio strategy and the firm’s overall strategic direction and goals. Shah et al. (2017) Qualitative To understand the phenomenon of BD. B Managers and archival data / Grounded theory. Strong brands are a source of competitive advantage, but weak brands diminish the firm´s competitive advantage. Deleting weak brands releases resources that can be reallocated to the strong brands to boost performance. * NOTE: Variables set: A–Causes of BD; B–BD decision-making; C–BD implementation; D–BD outcomes. TABLE 2 Approaches to the strategic decision-making process Construct Definition Rational The extent to which the decision process involves collecting information relevant to the decision and the reliance upon analysis of this information when making the choice. Intuitive Mental process based on a “gut feeling” and personal experiences to build a subjective decision-making rationale. Political Decision-making results when an unequal distribution of power allows more powerful groups or individuals to make decisions that reflect their personal interests. Source: Based on Dean and Sharfman (1993), Khatri and Ng (2000), Hickson et al. (1986) and Kester (2011). FIGURE 1 Research model Rational decision-making BD success Intuitive decision-making Political decision-making Type of BD Total brand killing or disposal vs. brand name change H1: + H2: + H3: – H5:- H6:+ H7:+ H4: + BD DECISION-MAKING APPROACH First objective Second objective 32 TABLE 3 Sample characteristics Brand characteristics Deleted brand N % Type of BD N % Created 108 69.70% Total brand killing or disposal 71 45.80% Acquired 47 30.30% Brand name change 84 54.20% TOTAL 155 100.00% TOTAL 155 100.00% Geographical scope N % Local/regional 23 14.80% National 95 61.30% International 37 23.90% TOTAL 155 100% Firm characteristics Industry N % Family business N % Manufacturing 39 35.10% Yes 75 67.60% Service 72 64.90% No 36 32.40% TOTAL 111 100.00% TOTAL 111 100.00% Number of employees (2014) N % Turnover (2014) N % <50 5 3.60% <= 10 6 2.70% <250 32 28.83% <= 50 26 23.42% >251 71 63.96% >50 67 60.36% N.A. 3 2.70% N.A. 12 10.81% TOTAL 111 100.00% TOTAL 111 100.00% Market targeted % Consumer 55.70% Industrial 44.30% TOTAL 100.00% 33 TABLE 4 Population and sample distribution by industry: Proportion test Population Sample NACE Code N % of total N % of total 10,11,12,13,14,15. Manufacture of food, tobacco and wearing apparel. 82 14.39% 19 17.12% 20,21,22,23,24,25. Manufacture of chemical, pharmaceutical, plastic and metal products. 68 11.93% 12 10.81% 26,27,28,29,30,31,32,33. Manufacture of electronic and optical products and machinery and furniture. 23 4.04% 5 4.50% 35,36,38,41 Electricity supply, water collection and waste management. 6 1.05% 2 1.80% 45,46,47. Wholesale and retail trade 190 33.33%* 24 21.62%* 49,52,53,55,56. Transportation, storage and housing services. 18 3.16% 3 2.70% 58,59,60,61,62,63. Information and communication. 19 3.33%* 12 10.81%* 64,65,66,69,70. Financial, insurance and professional activities. 129 22.63% 24 21.62% 71,73,74,77,79,81,82,85,86. Scientific, technical support education and health activities. 35 6.14% 10 9.01% TOTAL 570 100% 111 100% * Significant differences: p <.05. TABLE 5 Construct measurement Construct (Scale adapted of …) Items Mean (S.D.) Rational decision-making* (Papadakis et al., 1998; Kester, 2011) The management team made a comprehensive assessment of the economic, financial and market situation of the deleted brand. The decision was made in a systematic way (sequential, organized, logic and analytical). The decision was based on evidence and objective information. Multiple information sources were incorporated. 5.16 (1.80) 5.42 (1.65) 5.60 (1.41) 5.01 (1.69) Intuitive decision-making* (Khatri and Ng, 2000; Kester, 2011) We based our decision on what we felt to be right. We made the decision based on our own experience rather than on evidence. When making the decision, we took into account firm member experience. 4.15 (1.93) 3.63 (1.87) 4.68 (1.84) Political decision-making* (Dean and Sharfman, 1996; Kester, 2011) We had to negotiate and make concessions to get the decision approved. Our decision was conditioned by the stance of certain groups or individuals. We had to accept the position of particular groups or individuals to gain approval for the deletion. 3.26 (1.78) 3.19 (1.91) 2.91 (1.81) BD success** Deletion of this brand has been good for the future of the company. The company achieved the goals for which the decision was made. The deletion decision is considered a complete success. 8.31 (1.87) 8.42 (1.67) 8.18 (1.96) Firm´s prior economic situation* (Moorman and Rust (1999); Verhoef and Leeflang (2009) Our market performance was satisfactory. The company was performing well financially. The company was experiencing substantial growth. 4.97 (1.60) 4.98 (1.65) 4.58 (1.84) Experience in BDs*** (Dayan and Elbanna, 2011) Degree of experience in BD decisions. 5.70 (2.55) Formalization* (Argouslidis, 2008; Argouslidis and Baltas, 2007). A standardized or normalized procedure was used to execute the BD. An action plan was elaborated to guide the deletion process. Milestones or deadlines that had to be met were set up. The responsibilities of the members involved in the BD were pinned down. The evolution of the deletion process was regularly monitored. 5.00 (1.81) 5.40 (1.78) 5.40 (1.70) 5.34 (1.79) 5.36 (1.72) Note: * 7-point Likert scales (1: totally disagree, 7: completely agree); ** 10-point Likert scale (1: totally disagree, 10: completely agree). *** 7-point Likert scales (1: very low, 7: very high) 34 TABLE 6 Correlation matrix and discriminant validity Cronbachα CR AVE 1 2 3 4 5 6 7 1. Rational decision-making .88 .92 .74 .86 .40 .13 .26 .08 .05 .54 2. Intuitive decision-making .81 .88 .72 -.33 .85 .11 .10 .27 .13 .36 3. Political decision-making .90 .94 .83 -.07 -.01 .91 .21 .10 .22 .08 4. BD success .92 .95 .86 .24 .08 -.20 .93 .08 .06 .25 5. Firm´s prior economic situation .94 .95 .87 -.02 .25 -.09 .09 .93 .04 .13 6. Experience in BDs - - - .03 -.13 .21 .02 -.04 - .20 7. Formalization .94 .96 .81 .50 -.31 .07 .24 -.14 .20 .90 Note: The diagonal elements (in bold) are the values of the square root of AVE. The values below the diagonal are the zero-order correlation coefficients. The elements above the diagonal (in grey) are the values of Henseler et al.’s (2015) HTMT ratio of correlations. TABLE 7 Standardized parameter estimates Model 1 Model 2 Model 3 Hypothesized relationships Rational decision-making → BD success Intuitive decision-making → BD success Political decision-making → BD success Rational decision-making* Political decision-making→ BD success Type of BD→ BD success Rational decision-making *Type of BD→ BD success Intuitive decision-making *Type of BD → BD success Political decision-making *Type of BD → BD success .19* (H1) .20** (H2) -.20** (H3) .21* .22** -.20** -.13* (H4) .22** .25** -.20* * -.14* .08 .00 (H5) -.02 (H6) .13* (H7) Control relationships Firm’s prior economic situation → BD success Experience in BDs → BD success Formalization → BD success .04 .05 .20* .04 .05 .19* 03 .06 .18* R2 of BD success .15 .17 .19 * p<.05, ** p<.01 (one-tailed test). 35 FIGURE 2 Interaction effect between rational and political BD decision-making TABLE 8 Standardized parameter estimates of multigroup analysis for total brand killing or disposal vs. brand name change Group 1 Total brand killing or disposal (71 cases) Group 2 Brand name change (84 cases) Political decision-making → BD success -.39** -.05 ** = p<.01, * = p<.05 (one-tailed test). Significance levels based on bias-corrected bootstrap confidence intervals. The model estimated is Model 2. No significant difference between groups is found for the rational and political decision-making interaction.