Relative performance information and social comparisons: Exploring managers' cognitive, emotional and dysfunctional behavioral processes

Abstract

We explore the cognitive and emotional processes which manifest as a result of social comparisons involving relative performance information. We also explore how and why these processes may invoke dysfunctional behaviors. We mobilize social comparison theory and affective events theory to guide our study. We gather our data from a qualitative field study of a retail organization. Regarding cognitive processes, we find that, faced with a range of relative performance information, managers contemplate and select the most meaningful measurement comparisons for them. Managers also contemplate whether they can influence and attain their selected measures of relative performance. Such contemplation vis-à-vis leaderboard thresholds shape managers' emotions and dysfunctional behaviors. We conclude that perceived control over performance outcomes appears important in understanding how and in what ways social comparison effects unfold. We further conclude that social comparison processes push managers towards continuing performance improvements, even in the context of target achievement. Counterintuitively, perhaps, this effect may not always be organizationally desirable.

Introduction

In organizations, it is traditionally the case that a given manager can monitor his/her performance by reference to one or more designated performance targets (e.g., cost targets, sales targets). However, it is becoming increasingly commonplace that such performance information is not restricted to superior-subordinate eyes, but is made publicly available, potentially across the entire organization. For instance, store managers at Halfords plc, a UK-based retail company, are informed about each other's performance across multiple measures via intra-organizational leaderboards (Cullen et al., 2013). Similarly, Amazon, KPMG, and Omnicare regularly disclose such relative performance information (RPI) to their employees (Cardador et al., 2017; Berger et al., 2019). Organizations intend that, because of the innate human tendency to socially compare (Festinger, 1954), distributing RPI in this 'public' manner will improve effort and concomitant task performance, as managers strive for a higher leaderboard position. Nevertheless, dysfunctional side effects, including effort distortion and uncooperative actions, have been reported (Hartmann and Schreck, 2016; Wang, 2017; Hannan et al., 2019).

Experimental-based accounting research examining why dysfunctional side effects occur predominantly focuses on the different ways in which RPI can be provided (e.g., Hannan et al., 2013, 2019; Tafkov, 2013; Kramer et al., 2016). We take an alternative approach by qualitatively exploring the cognitive and emotional processes that manifest when performance measurement is based on publicly available RPI. This is an important issue, yet we know little about how managers' cognitive and emotional processes unfold in RPI work environments (see Mahlendorf et al., 2014; Luft, 2016). Notably, we seek to understand how and why these processes may invoke dysfunctional behaviors. We focus on exploring potential dysfunctional behaviors that may arise as initial access meetings with our case organization suggested these were both prevalent and costly.

RPI systems can provide two cognitive reference points (Hartmann and Schreck, 2016). In addition to comparing performance against designated targets, managers can also assess their performance relative to other managers' performance for a given target through explicit ranks. Guided by the principles of social comparison theory (SCT) (Festinger, 1954), we are interested in understanding if differences exist among measures of relative performance, in terms of the cognitive and emotional responses engendered. When presented with multiple measures of relative performance that are not necessarily correlated or may even conflict (Brown et al., 2014), managers may not compare all aspects of their performance to all colleagues in their assigned peer group (see Boedker and Chua, 2013; Luft, 2016). As Tafkov (2013) explains, comparison concerns are shaped by the measure's relevance and the colleague's commensurability. However, these conditions are generally imposed on participants within RPI experiments, so we cannot conclude from extant research how managers decide on what and with whom to compare.

Moreover, accounting research generally assumes that RPI-induced social comparisons are motivated by self-evaluative and self-enhancement cognitive goals only, yielding contrast emotions due to a competitive work environment (i.e., emotions stimulated by believing that colleague 'is not me').¹ We seek to explore the validity of this assumption. Social psychological and organizational behavior research indicates that it is possible managers' tendency to compare their performance with that of colleagues on measures of relative performance could be underpinned by self-improvement goals (see Wood, 1996; Buunk and Gibbons, 2007), and/or evoke assimilative emotions (i.e., emotions stimulated by believing that colleague 'could be me').² As Mahlendorf et al. (2014) acknowledge, accounting field research seldom explores the factors that may induce contrast or assimilation effects. Managers' perception of their ability to control performance outcomes is one pertinent factor (Smith, 2000), as RPI challenges the controllability principle (Choudhury, 1986: 195).³ An affective events theory (AET) lens subsequently enables us to reflect on how and why these cognitions and emotions impact observed dysfunctional behavior. That is, whether dysfunctional behavior occurs directly as affect-driven behavior or more indirectly through attitudes and evaluations about the job situation that manifest as judgment-driven behavior (Weiss and Beal, 2005; Redmond, 2017).

We collect our data mainly through semi-structured interviews with store managers at a retail organization that operates a store-level RPI system. This system publicly displays store managers' absolute performance against a target on several performance measures (e.g., sales, expenses, customer satisfaction). It also displays a store manager's relative performance rank versus other store managers within his/her assigned workgroup for each measure. The information is frequently distributed via leaderboards that highlight whether a store manager's performance is in the top, middle, or bottom quartile of his/her workgroup. The information, in part, impacts store managers' self-esteem, determines who is subject to ongoing performance reviews, and facilitates the allocation of discretionary public praise or condemnation (e.g., certificates, additional 'named and shamed' group emails).

We observed that social comparison processes are not necessarily straightforward, nor instinctive. Rather, managers contemplate and consider the most meaningful comparisons for them; in so doing, they select a subset of measures and colleagues for comparison purposes. At our research site, a measure became relevant for comparison purposes if the store manager's past performance secured a rank that located in a monthly leaderboard's extreme quartiles. As well as such rank positions having evaluative and reputational implications, the measure itself became a domain of self-worth in which self-esteem was staked. A colleague was deemed commensurable if s/he worked in a store of a similar profile to the manager's (e.g., a store of a similar age and with similar complexities, customer demographics, product mix, and profitability). Supplementary store reports and frequent internal lateral transfers of store managers between stores in a workgroup appeared to facilitate the observed selectivity of comparators.

We also observed how managers' emotions vis-à-vis selected RPI may depend not only on their leaderboard position but also, and importantly, on their perceived ability to control performance outcomes. For example, several of our top-quartile store managers perceived that they had high control over their performance vis-à-vis the measures in question. We observed these managers contrast their performance with selected workgroup colleagues, experience pride, and demonstrate boastfulness. Conversely, we noted how other top-quartile store managers perceived low outcome controllability. These managers appeared to assimilate, fearing that their performance would decline. Public praise and recognition of performance reinforced such emotions. The significance for organizations is the behaviors invoked in each case. We note how top-quartile store managers engaged in dysfunctional behavior to maintain their rank and self-esteem, as well as increasing the likelihood of progressing to 'bigger', 'complex' or even 'broken' stores. Bottom-quartile managers also engaged in dysfunctional behavior, but this time the reasons appeared to be focused on improving their rank, reducing shame through the pursuit of self-esteem, and minimizing the risk of demotion or dismissal. For managers who perceived low outcome controllability and experienced shame, frequent RPI availability and public condemnation intensified emotions, with such emotions being felt even if absolute performance was on-target. That only middle-quartile managers avoid potentially dysfunctional behavior is of theoretical and practical significance.

Our study offers three key insights. First, we tease out that the timing (frequency of distribution), presentation (utilizing meaningful thresholds), and use (discretionary performance feedback, assumed linkages to performance evaluations and incentives) of RPI shape the perceived relevance of the measure for social comparison purposes. This finding helps us explain why, relative to both high- and low-ranked managers, middle-quartile-ranked managers may have a lower propensity to compare and engage in dysfunctional behavior. At our research site, a middle-quartile rank did not appear to be a source of anxiety; such performance received minimal scrutiny and seemed less important for self-validation. Middle-quartile performers thus felt less ongoing pressure to improve their ranking. The result was a U-shaped interplay involving RPI, social comparisons and manifest dysfunctional behaviors. This finding contributes to debates about the relationship between RPI and effort (e.g., Hannan et al., 2008, 2013, 2019; Tafkov, 2013), specifically how the response function is shaped through signals such as rank ordering (Kramer et al., 2016), the proportion of winners (Knauer et al., 2017) and discretion over RPI provision (Hecht et al., 2019).

Second, we contribute evidence that suggests that the degree of control a manager perceives having over his/her performance outcomes influences whether assimilative or contrastive comparison judgments and associated emotions occur in a competitive work environment. This enables us to highlight that social comparison-based processes manifest themselves in ways which are broader in scope than has been recognized to date. Moreover, experimental RPI studies are relatively silent on the issue of perceived outcome controllability (cf. Hannan et al., 2019; Hecht et al., 2019), even though respect for the controllability principle has long been argued to determine how managers behave towards designated performance measures (e.g., Giraud et al., 2008). This insight has scope to be tested in future experimental studies.

Third, we shed light on tension in RPI systems that simultaneously provide information about absolute targets and relative ranks. Addressing Hartmann and Schreck's (2016: 17) call to "theorize further on the interaction between absolute and RPI," we suggest that managers care more about their relative rank than the absolute target, because it is the rank that determines the winners and losers (Tran and Zeckhauser, 2012). To this end, social comparison processes may continue to push managers towards yet further performance improvements, even when, formally, targets are achieved. Extant accounting research offers little insight into what may happen behaviorally in the context of target achievement. Rather, the focus has tended to be on understanding how best to motivate managers to meet pre-set targets (e.g., Argyris, 1952; Hofstede, 1968; Merchant and Manzoni, 1989; Pfister and Lukka, 2019). The organizational implication is that if social comparison processes continue to motivate beyond target achievement, expending time and effort seeking to set the 'right' target arguably becomes less critical. In this context, Hofstede’s (1968) classic refrain to set targets that are 'tough but achievable' has less import if social comparison processes via relative ranks ensure motivation persists beyond target achievement, even for 'easy' targets. That said, while these may be conceived as positive policy implications, the behavioral consequences reported in this study also highlight that associated motivations can result in dysfunctional outcomes.

The next section outlines our theoretical lenses. Section 3 describes our research methodology. Section 4 presents our study's findings, and Section 5 our discussion. Section 6 outlines our conclusions and recommendations for future research.