Abstract
This article explores the narrative structure of management cases. A selective sample of cases (n = 5) which focus on the turnaround of a high profile corporation is examined. The analysis considers the notions of: thematic framing and ordering (i.e., univocality, causal chains and a ‘problem-to-solution’ flow); verisimilitude (i.e., coherence and plausibility); the use of poetic tropes (i.e., the attribution of responsibility, agency and providential significance). The common features and general characteristics revealed in both ‘conventional cases’ and ‘critical cases’ are identified and discussed. Finally, the scope for embracing and applying alternative approaches (e.g., plurivocal, fragmented and less problem-centred) is presented.
Introduction
Management cases continue to be very popular and powerful educational devices (Ellet, 2007). In addition to being an aid to teaching, case studies also represent an important vehicle for engaging in meaningful research activity (Yin, 2009). The primary aim of this analysis is not to question the value and utility of management cases per se. Rather the intention is to derive insights into the overarching thematic patterns and common characteristics of management cases and to explore the potential for developing alternative forms of case study engagement within the field of management.
The starting point for this analysis is the premise that management case studies are, to a large extent, like stories insofar as they are designed to offer a coherent and plausible account of a series of temporally embedded events and/or happenings (Oswick, 2008; Polkinghorne, 1995). On this basis and drawing upon insights derived in the extant literature on storytelling in management, this contribution utilizes ‘narrative analysis’ (Boje, 2001; Czarniawska, 1997; Gabriel, 2000; Reissman, 1993) techniques and protocols to interrogate a sample of management cases.
There are four main parts to this article. First, in the following section details of the case study sample and methodological approach employed are provided. Second, the results and findings of the research undertaken are presented. Third, the effects of the predominance of highly narrativized management cases are discussed. Finally, some conclusions are drawn regarding the current hegemony around case study design and the scope for heterogeneous approaches is outlined.
The Sample and Methodology
The sample consists of five management cases. The sample has been selectively matched insofar as all the cases focus on Sears, Roebuck & Company and consider a specific period of time in which the company turned around its performance (i.e., from 1992 to 1998). The samples contain one critically-oriented case (Collins & Rainwater, 2005) which is written for an academic audience rather than explicitly having a practitioner or student emphasis. The composition of the sample is intended to offer a means of considering the same period and unfolding set of events through different lens (i.e., different case study perspectives). It also presents an opportunity to establish whether narrative structure differs according to the case purpose (i.e., educational or research-oriented), and the audience targeted (i.e., academics, practitioners or students).
In addition to the critical case of Sears provided by Collins and Rainwater (2005), there is a practitioner-oriented case and three student-oriented cases. The full sample of cases, presented in chronological order, is:
Ghemawat, P. (1993). Sears, Roebuck and Company: The merchandise group (A Harvard Business School Case). Sood, S. (1997). Sears, Roebuck and Co. (A Stanford University Published Teaching Case). Hallowell, R. & Cash, J.I. (1997). Sears, Roebuck and Company (A): Turnaround (A Harvard Business School Case). Rucci, A.J., Kirn, S.P. & Quinn, R.T. (1998). The employee–customer–profit chain at Sears (appearing in the Harvard Business Review). Collins, D. & Rainwater, K. (2005). Managing change at Sears: A sideways look at a tale of corporate transformation (appearing in the Journal of Organizational Change Management).
The process of interrogating the sample of cases was undertaken by combining two methodological approaches, namely discourse analysis (Fairclough, 2003; Potter & Wetherell, 1987) and thematic analysis (Boje, 2001). The texts were classified and analyzed in three basic phases of inquiry. First, the notion of the ‘dramatistic pentad’ (Burke, 1989) was employed to provide a general framework for comparing and contextualizing the chosen cases. Second, a protocol based upon the presence of ‘poetic tropes’ (Gabriel, 2000) was utilized to compare the specific characteristics and narrative structure of the different cases. Finally, data produced in the first two phases were further investigated and synthesized for evidence of ‘narrative-embedded verisimilitude’ (Brown, 2004). It is probably helpful at this juncture to elaborate a little further on these three phases of analysis.
Burke’s (1989) ‘dramatistic pentad’ provides a means of exploring an event in terms of five basic dramaturgical components (i.e., act, scene, agent, agency and purpose). For our purposes, we can think of ‘the act’ as being the main occurrence (i.e., what actually happened), ‘the scene’ as the physical locale and backdrop (i.e., the context), ‘the agent’ as a key stakeholder or an actor, ‘the agency’ is the process by which an agent acts (i.e., the ‘how’ of acting) and ‘the purpose’ is the intended or desired effect (i.e., the why of acting). The pentad has considerable general applicability within the social sciences and, although it has not explicitly been directed towards management cases, it has been deployed in the past to explore company-specific events (e.g., Walker & Monin, 2001).
So, what are ‘poetic tropes’? According to Gabriel (2000), they are interpretive devices which storytellers use to make sense of events. He explains: ‘Each one of these tropes represents a way of either making sense of specific parts in the narrative or making connections between different parts’ (Gabriel, 2004, p. 65). Gabriel has identified eight tropes: (i) the attribution of motive(s); (ii) the attribution of causal connections; (iii) the attribution of responsibility; (iv) the attribution of unity; (v) the attribution of fixed qualities; (vi) attribution of emotion; (vii) the attribution of agency (to objects); and (viii) the attribution of providential significance.
For Gabriel, poetic tropes are interpretive devices used by storytellers. For our purposes, we will be substituting ‘storytellers’ for ‘case study writers’ in order to examine whether comparable trope-based processes of sense-making arise in management cases (i.e., as forms of plausible narrative).
The final analytical phase will consider the concept of verisimilitude. The literal definition of verisimilitude according to The New Shorter Oxford English Dictionary is where something that has ‘the appearance of being true or real’. This is consistent with Denzin and Lincoln’s (1994) assertion that ‘Verisimilitude can be described as the mask a text assumes as it convinces the readers it has conformed to the laws of its genre’ (p. 580). In his application of the concept to texts, Brown (2004) posits:
My understanding of verisimilitude suggests that its attribution implies three related perceptions: that a text conforms to the rules of its genre, that it offers a vicarious experience, and that it provides good reasons for the events it describes. (p. 100)
Arguably, Brown’s three assertions about verisimilitude could be legitimately applied to case studies insofar as there are undoubtedly discernible conventions around the genre of case study writing and case studies are explicitly designed to offer a vicarious experience (i.e., learning from the process of systematically considering the actions and decisions of others). We might also expect that a management case study typically provides ‘good reasons for the events it describes’ (Brown, 2004).
In addition to the attributes of verisimilitude, it is the nature of the process of aspiring to achieve it, which is also of interest in this research. Gabriel (2004) has suggested that ‘…storytellers use “facts” in a plastic way, moulding them to the requirements of the plot which is itself a reflection of the needs of the audience’ (p. 66). This raises questions about how case writers use facts. To this end, we will draw upon work on the ways in which plots are manipulated in stories to examine the construction of verisimilitude in management cases. More specifically, we will utilize Gabriel’s (2004, p. 66) classification of how events are moulded: (i) Framing and Focusing—emphasizing events/characters and downplaying others; (ii) Filtering—removing certain events or characters; (iii) Fading—events and characters are brought in or out of focus; (iv) Fusing—two or more characters or events are merged to create a single one; and (v) Fitting—specific events or characters are re-interpreted or re-presented in line with the plot.
Results and Analysis
The use of the ‘dramatistic pentad’ (Burke, 1989) revealed some significant points of similarity and dissimilarity within the sample of cases (see Table 1). In all five cases, there was general consensus around ‘the act’ at a broad level (i.e., everyone agreed that Sears had made a considerable loss in 1992, returned to profits in 1993 and further improved profitability each year through to 1997). There was also broad agreement about ‘the scene’ (i.e., a context of strong competition in the retail sector) and ‘the agent’ (Arthur C. Martinez featured as the primary protagonist in four cases and was a secondary actor in the remaining one). However, the more detailed framing of core problem and the reasons for the turnaround were more varied. For Ghemawat (1993), the problem was one of re-balancing operations and success was attributed to reducing financial services and improving the merchandising portfolio. While others were in agreement in attributing the turnaround to Martinez, they tended to emphasize different forms of agency. For Sood (1997), Martinez effectively cut costs and streamlined retail operations. According to Hallowell and Cash (1997), Martinez led employees effectively and implemented a five-part strategic plan. For Rucci, Kirn and Quinn (1998), Martinez successfully reconciled and aligned employee and customer needs to enable the company to generate profits. Finally, Collins and Rainwater (2005) suggest that Martinez achieved success by subordinating and controlling labour.
With regard to the use of ‘poetic tropes’ (Gabriel, 2000) across the cases, there was remarkable consistency in the use of five of the eight attributes (i.e., motive, causal connections, responsibility, unity and fixed qualities) (see Table 2). In particular, there were explicit examples of ‘individuals seeking to influence events’ (i.e., exhibiting a motive) and ‘praiseworthy events were credited to single agents’ (i.e., the attribution of responsibility).
Causal connections, as a form of poetic trope, arise when ‘chronological sequence is turned into causal chain, earlier events causing subsequent ones’ (Gabriel, 2004, p. 65). This occurred in all the cases as one event led to another and then another, but it was most apparent in the transition from the articulation of problems to the presentation of solutions. The one case that differed in this respect was the critical case by Collins and Rainwater (2005); rather than presenting a single causal chain, it presented three contrasting causal chains with differing motives and forms of agency which in turn cast the main agent (i.e., Martinez) in three different roles (i.e., hero, villain and fool).
The other interesting facet of Table 2 was the complete absence of three of the poetic tropes (i.e., the attribution of emotion, agency to objects and providential significance) in the three conventional student-oriented cases (i.e., SRC1, SRC2 and SRC3). The attribution of emotion was found in the practitioner-oriented case (i.e., SRC4), but it was attributed to consumers, in terms of their enduring emotional attachment to Sears merchandise, rather than to a main agent in the case study. Equally, in the critical, academic-oriented case (SRC5) emotion was attributed to disenfranchised employees rather than the main agent in the case study. This is arguably one of the ways in which stories are significantly different to management cases. In stories it is important for the main protagonists to have feelings and to exhibit emotions in order to humanize them. But, in traditional management cases, the positioning of key actors is as ‘rational decision-makers’ who act in a relatively objective way.
‘Dramatistic Pentad’ (Burke, 1989) Classification of Sears, Roebuck & Co. Cases
Following on from the absence of emotion in cases, it is also noticeable that the attribution of agency to objects and the attribution of providential significance also only appear within the critical case (i.e., SRC5). This is perhaps tied to the need to retain the role of the main agent as the ‘key decision-maker’ and to secure objectivity and coherence in the plot. If agency is attributed to objects, this effectively detracts from the agency of the main protagonist (i.e., ‘things act on the agent’ rather than ‘the agent acting on things’). Put differently, it would further complicate the Sears case study, if Martinez’s agency was being undermined by non-human agents.
‘Poetic Tropes’ (Gabriel, 2000) Classification of Sears, Roebuck & Co. Cases
For similar reasons, the acknowledgement of providence significance is problematic in management cases because it introduces the notion of ‘fate’ and other unpredictable variables in the scenario. This impacts upon the perceived agency of the main agent (i.e., the ability to control things and events) and has the effect of diluting or disrupting the coherence and explanatory power of the case.
Discussion and Conclusions
The scrutiny of management cases using concepts from narrative analysis has produced some interesting insights. There are clearly strong points of similarity between stories and case studies in terms of narrative structure (especially around the role of key protagonists and the causal connection of events). The dramatistic pentad also demonstrates the way in which verisimilitude is achieved in case studies insofar as the relative concrete and uncontested components across the cases (e.g., that Sears went from loss-making to profit generating and that Martinez played a central role) are interwoven with more subjective and contrasting forms of rationalization around the problem framing (e.g., portfolio imbalance in SRC1, over-diversification in SRC2, lack of direction and leadership in SRC3 and insufficient attention to employee and customer needs in SRC4) and, through the attribution of causal connections, this in turn prescribes particular courses of remedial action. In this regard, the facts in case studies, just like stories, are moulded through verisimilitudinous processes involving the ‘framing’, ‘focusing’, ‘filtering’, ‘fading’, ‘fitting’ and ‘fusing’ of events and characters (Gabriel, 2000, 2004).
Beyond the points of similarity between stories and case studies, this analysis has also highlighted some points of significant difference. While stories typically attribute emotion, agency to objects and providential significance, these tropes were not embedded with case studies. It has been suggested earlier that the distinct absence of these poetic tropes is because, albeit in subtly different ways, they introduce elements of ambiguity, subjectivity and indeterminacy and, as such, they undermine the coherence of causal chains and claims to objectivity. In this regard, case studies are substantively different to stories. Stories do not need to be ‘true’ or ‘real’, but case studies generally do. Hence, management cases have a stronger need for verisimilitude than stories.
Ironically, it is the demand to demonstrate verisimilitude which can make management cases somewhat implausible. This arises because the pursuit of a singular, univocal account of a given situation is itself implausible. There are multiple realities and multiple alternative readings of reality (Boje, 1995). Indeed, the five cases analyzed in this article offer slightly different readings or versions (or realities) of the same situation.
This raises some interesting questions regarding the design of cases. The more straightforward questions revolve around: whether we overplay the level of agency associated with key protagonists; whether we privilege the articulation of causal chains; and, whether we downplay ambiguity, serendipity and alternatives in our desire for verisimilitude. There are also other fundamental questions. For example, rather than giving readers insights into ‘a problem’ should one of the tasks of a case study be to discover and/or generate problems? Should multiple, and even contradictory, framings of a problem be offered? Should multiple, alternative solutions be offered? Should case counter-pose alternative framings and explanations? Should cases encourage plurivocal divergence rather than promote univocal convergence?