FREEDOM OF CHOICE IN CUSTOMIZED OFFERS: SELF-CONCEPT CLARITY AS A MODERATOR by

FREEDOM OF CHOICE IN CUSTOMIZED OFFERS: SELF-CONCEPT CLARITY AS A MODERATOR

by

Lennard Winters

Supervisors: Prof. dr. B.A. Nijstad

&

Prof. dr. P.C. Verhoef

Rijksuniversiteit Groningen Faculty of Economics and Business

MSc thesis

Marketing Management and Human Resource Management

ABSTRACT

Personalization is a popular trend in both marketing and HRM but customized offers are not always effective and are not effective for everyone. In this research, choice and self-concept clarity are investigated as predictors of the evaluation of customized offers and loyalty towards the company. A customized offer that provides targets the freedom to choose their exact offer from a choice set is expected to result in a more positive evaluation and higher loyalty than an offer without this choice. It is expected that adding attractive alternatives to the provided choice set leads to an increase of attractiveness but also to a more difficult choice while the addition of unattractive alternatives decreases the attractiveness but makes the choice easier. Specifically, it is expected that the addition of attractive alternatives is beneficial for people with higher self-concept clarity while the addition of unattractive alternatives is beneficial for people with lower self-concept clarity.

An experiment, in both a marketing and an HRM setting, was conducted among 416 bachelor students at the University of Groningen. The results show that freedom of choice may have a positive effect on evaluation. Moreover, a choice set may lead to higher loyalty through more positive evaluation. This only applies to people with high SCC when the choice set included attractive alternatives, and only to people with low SCC when the choice set included unattractive alternatives. The results suggest a recommendation for companies to sending customized offers with a choice set, and that companies could effectively differentiate based on personality characteristics.

INTRODUCTION

For both the consumer and job markets, it is often more important to retain loyal customers and committed employees than to acquire new ones. Reichheld and Strasser (1990) argue that acquiring new customers is typically more costly than keeping current customers, and that long-tenure

customers are more profitable. Moreover, company efforts to retain employees may result in lower withdrawal cognitions (Yamamoto, 2011) and higher job satisfaction (Barber, Dunham, &

Formisano, 1992), which ultimately benefit firm performance (Kumar & Arora, 2012).

One way for companies to retain customers and employees is to focus on the relationship. In a relationship both parties involved should be willing to sacrifice for the sake of an enduring

relationship. To build better customer relationships, companies can invest in Customer Relationship Management (CRM) (Payne, 2006). For employee relationships the corresponding term is

Employee Relationship Management (ERM) (Strohmeier, 2013). Recent advances in information technology have allowed companies to maintain detailed and accurate information about customers’ preferences (Clemons & Weber, 1994) and to, thereby, customize communication specifically directed at individual consumers’ preferences and needs at relative low cost (Payne & Frow 2005). Literature on this topic emphasizes the broader scope of CRM beyond technology (e.g. Iriana, Buttle & Ang, 2013) prompting the following definition of CRM: “the core business strategy that integrates internal processes and functions, and external networks to create and deliver value to targeted customers at a profit. It is grounded on high-quality customer-related data and enabled by IT” (Iriana, Buttle & Ang, 2013, p.468)

The technological developments that contribute to the improvement of customer relationships are not limited to marketing, but have also been used within the employee domain (Strohmeier, 2013). Within Human Resource Management (HRM), ERM refers to the “strategy, programs and

technology [that] effectively manage how firms relate to prospective, current and former

employees” (Rogers, 2008, p48). As in the company-customer relationship, there are clear calls for “personalization” in ERM (Gillenson & Sanders, 2005). The call for individualization in ERM should lead to customization of all activities directed toward the individual employee (Strohmeier, 2013). Strohmeier (2013 p. 97) agrees that HR process of compensation should lead to the

Thus, CRM and ERM both focus on the individual target (customer or employee) and consist of customized communication. These practices try to create mutual value for both the company and the target (Rogers, 2008; Keim & Fritsch, 2008; Iriana, Buttle & Ang, 2013), in order to foster a good relationship (van Tongeren, 2010). Targets receive offers customized to their individual preferences. In CRM, this may consist of, for example, a customized marketing offer in which a regularly-bought product is offered at a discount. In ERM, customized offers may consist of a flexible benefit system and thus customization in the compensation process (Wilky & Christie, 2000). Rather than providing each employee with the same benefit package, employees are given the option to choose, for example, additional payment, medical care, extra vacation days, stock options, dentist care, lodging and/or a company car.

The effectiveness of a customized offer likely depends on the characteristics of the offer itself and the individual for whom it is customized, there are for both characteristics important shortcomings in theory. In terms of offer characteristics, there is a reasonable amount of research aimed at the appearance of the offer (Sela, 2010), the product that is offered, and how well this offer fits the individual’s needs (Simonson, 2005). When choice is provided to targets, it will arguably better fit their needs (Simonson, 2005). However, a customized offer that provides the target the freedom to choose the exact content of their offer from a choice set seems absent in literature. Besides,

researchers warn for negative reactions to the specificity of customized offers and, therefore, the perceived self-interest and intrusiveness of the offer provider (Sela, Simonson and Kivetz, 2003; Van Doorn & Hoekstra, 2013). Providing the target of a customized offer with a choice set may be a reasonable strategy to overcome some of a target’s resistance towards the self-interested and intrusive behavior of a company and thus may be a gap in literature. A company may offer a choice set with several options, including the most preferred option, to the target (whether customer or employee), and ask the target to choose one from among them. Alternatively, without freedom of choice, the target is provided with a single offer which only can be accepted or refused. Given that people prefer to make their own choices over having them dictated externally (Botti & Iyengar, 2006), it would follow that offers with a choice set will be more attractive and lead to more positive consequences.

However, the presumption that people are never worse off, and are usually better off, as a result of making their own choices may not necessarily be true. Human beings have always shown

options (Iyengar & Lepper, 2000) or because cognitive conflicts arise, freedom of choice may lead to dissatisfaction with decision outcomes (Chernev 2003; Huffman and Kahn 1998; Shafir,

Simonson, & Tversky 1993). Target characteristics may play a role for the presence of both positive and negative effects as will be discussed in the next paragraph.

In literature it is unclear for who customized offers are effective because not much attention has been given to target characteristics. This is an important shortcoming, because CRM and ERM are individual-centric strategies aimed, by definition, at individual customers and employees. It seems, therefore, logical that target characteristics should be important determinants of success in CRM and ERM. This paper will focus on the individual’s self-concept clarity (SCC) as target

characteristic. This refers to the extent to which an individual’s self-beliefs are clearly and

confidently defined, internally consistent, and stable (Campbell, Trapnell, Heine, Katz, Lavallee & Lehman, 1996). Unconfident and inconsistent self-beliefs make it reasonably more difficult to choose and therefore SCC is relevant for this paper. Furthermore, SCC correlates with preferences stability (Kihlstrom & Cantor, 1984; Markus & Wurf, 1987) and is therefore relevant for

customized offers in general. Indeed, company attempts to understand targets’ preferences and assumes that these preferences are stable and can be derived from previous behavior. The effectiveness of customized offers may thus depend on an individual’s SCC.

We argue that a choice set is beneficial, in comparison to a single offer, for the evaluation of a customized offer, and for the loyalty towards the company that provides the offer. Furthermore, it is expected that a choice set that includes the most preferred option with several attractive alternatives will lead to an attractive offer. However, the attractive alternatives choice set implies a difficult choice, especially for people with low self-concept clarity. Under this condition, the customized offer results in a more positive evaluation and higher loyalty -than without the freedom of choice- when self-concept clarity is high. Additionally, it is expected that a choice set that includes the most preferred option with several unattractive alternatives may lead to a less attractive offer. This

unattractive alternatives choice set implies an easy choice because only one option is relevant, beneficial for people with low self-concept clarity. People with higher self-concept clarity will quicker disregard the unattractiveness of the alternatives, thus under this condition the customized offer only results in a more positive evaluation and higher loyalty -than without the freedom of choice- when self-concept clarity is low.

of choice within customized offers on offer evaluations and loyalty towards the company, and to what degree does this depend on self-concept clarity?

The contribution of this research is twofold. On the one hand, the question of how, and for whom, choice is beneficial or harmful will contribute to the theoretical understanding of human choice behavior. Moreover, it will provide better insights into theory on customized offers. On the other hand, it offers practical applications for companies to improve their customized marketing and HR practices and, thereby, gain competitive advantages.

The variables above have been summarized in Figure 1, and will be elaborated upon in the following sections.

Figure 1. Research model.

Evaluation

Choice Loyalty

THEORY

Central to CRM and ERM practices is the construct ‘relationship’. In ancient Greece, Aristotle argued that the following characteristics define friendship: “[friends] wish and grant each other the best (and know that from another), for the sake of usefulness, or pleasure, or the good” (Aristoteles, 350 B.C. 1156a1-5). Three important lessons can be gleaned from this ancient definition. Firstly, friendship can be built on utility and is, therefore, applicable to the relationship between a company and its customers and employees. Secondly, a customized offer is reasonably beneficial to this relationship because the company ‘grants’ the target (be it a customer or employee) something that is positive. Thirdly, Aristotle’s definition implies reciprocity. Indeed, both in theory and practice, evidence indicates that CRM and ERM practices lead to better relationships. Additionally, more loyal customers (Reinald & Kumar, 2000) and employees (Yamamoto, 2011) lead to higher profitability for a company (Boulding, Staelin, Ehret & Johnston 2005; Seiders, Voss, Grewal & Godfrey, 2005; Kumar & Arora, 2012).

This definition of friendship also introduces important points of attention for a company. A good relationship cannot exist without reciprocity, without which one party would be abusing the other (van Tongeren, 2010). Moreover, according to Aristotle, both parties must ‘know from another’ that they wish the other party the best. These drawbacks also apply to CRM and ERM practices. Indeed, when targets perceive a company’s behavior as self-interested (Sela, Simonson and Kivetz, 2003), unfair (Yamamoto, 2013) or intrusive (van Doorn & Hoekstra, 2013) it leads to negative responses. This will be elaborated later.

CRM

As briefly described in the introduction, CRM entails building customer relationships with the help of data and information. This data and information about individual customers is available due to recent technological developments. Companies are increasingly recognizing that customers have different values (Reinartz, Krafft & Hoyer, 2004) and are attempting to maximize these values (Verhoef & Lemon, 2013). Companies analyze data and use this information to develop customized offers (Reinartz, Krafft & Hoyer, 2004; Simonson, 2005), retain customers, and stimulate increased spending (Verhoef & Lemon, 2013).

based on customer satisfaction and wallet share, and 4) the sales approach should focus on the fulfillment of customer needs and the sale of more products to the same customers (Shah, Rust, Parasuraman, Staelin & Day, 2006).

Over the past few years, CRM has been widely adopted by companies, has received a great deal of attention from marketers and consultants (Kivets & Simonson, 2003), and is one of the most visible themes in marketing literature (Allaway, Whipple & Ellinger, 2011). Its popularity can be explained by the fact that successful customer relationship management results in higher customer loyalty (Peppers and Rogers, 1997), leading to a larger customer base (Reinartz & Kumar, 2000) and, subsequently, higher profitability (Seiders et al., 2005). This is highly important to maintain profitability and a sustained competitive advantage in the future (Anderson et. al., 1994). As an example for successful CRM, Boulding et. al (2005) offer an overview of ten articles in different industries and contexts, utilizing various measurements and research methods. A significant relationship between CRM practices and enhanced firm performance was found in eight of ten articles. Research included cross-sectional and longitudinal methods, and performance

measurements included customer retention and satisfaction, customer knowledge, and company profits. As Boulding et. al. (2005, p.157) conclude: “We believe that this is a powerful result”.

ERM

In these times of fierce competition, organizations are at a greater risk of losing capable human resources to other organizations (Fox, 2014). Therefore, ‘retention’ is the key word for HRM (Yamamoto, 2009) and CEOs (Kumar & Arora, 2012) to keep in mind. ERM practices should help an organization to retain its well-performing employees, thereby allowing the company to gain a competitive advantage (Strohmeier, 2013). Indeed, higher retention rates results in higher revenues, through higher productivity (Strohmeier, 2013) and less knowledge loss (Yamamoto, 2009), and lower costs than acquiring new employees (Kumar & Arora, 2012).

In their study of 215 smaller businesses, Allen, Ericksen and Colins (2013) provide empirical evidence for the relation between ERM practices and firm performance. They found substantial revenue growth two years after the implementation of commitment-based HR-practices, which was partly mediated by the retention of employees.

organizational treatments with positive attitudes and consequential high performance levels (Shore, Bommer, Rao & Seo, 2009). Within ERM, customized compensation should be important. Indeed, appropriate compensation that take employee individual needs into account (Barber, Dunham, & Formisano, 1992) results in higher retention rates (Rogers, 2008; Keim & Fritsch, 2008;

Yamamoto, 2013).

One option to consider when aligning compensation with employee needs is the introduction of a flexible benefit system (Strohmeier, 2013). In a flexible benefit system, employees can choose their preferred benefits, such as additional payment, medical care, more days off, stock options, dentist care, life insurance and/or a company car. This should increase employee satisfaction (Sturman, Hannon & Milkovich, 1996) and decrease employee turnover (Yamamoto, 2011). This concept is supported by prior studies, which have shown that employee benefit practices have a positive effect on retention and, therefore, on firm performance (Mitchell 1982; Lincoln & Kalleberg 1996; Fairris 2004). Literature underlines the importance of customized benefits for the individual employee when attempting to successfully implement a flexible benefit program (Yamamoto, 2011; Barber, Dunham, Formisano, 1992; Strohmeier, 2013).

It is undeniable that CRM and ERM programs generate more loyal customers and employees, positively affecting firm performance. These relation management practices entail sending customized offers to customers and employees, respectively. We will shortly address the most important determinants of success of these customized offers. The determinants rely strongly on marketing literature since customized offers are used for a longer period, and more frequently, in marketing than in HRM.

Evaluation of customized offers

The evaluation of an offer depends on the actual attributes of that offer. Alongside the offer’s perceived fit with a customer’s preferences, this leads to an evaluation of the offer (Simonson, 2005). A higher fit with preferences results in a more positive evaluation (Sturman, Hannon & Milkovich, 1996; Simonson, 2005). The perceived fit with customer preferences may be higher than the actual fit when targets feel explicitly addressed. This is particularly the case when targets

receive an offer with attributes that they believe differentiate them from others (Simonson, 2005) because these attributes are relevant for the self and identity (Belk, 1988; Sirgy, 1982). Lastly, the offer’s presentation context may be important. For example, a bad option can be presented

overvalued but the bad option is undervalued when presented after the good option (Chernev & Gal, 2010). It is also known that using a less favorable product as a benchmark results in a more positive evaluation. For example, Simonson and Tversky (1992) demonstrate that consumers are more likely to spend $6 on an elegant Cross pen when they also have the option of spending $6 on a less

attractive pen. Thus, positive actual attributes, a high perceived fit with individual needs, and a smart presentation context all lead to the positive evaluation of an offer.

Currently, it seems favorable to always customize an offer, but that view is incorrect. As previously mentioned, customizing an offer requires data and information about individuals in order to

ascertain individual preferences. Data does not always correctly predict customer preferences, however, and Simonson (2005) warns marketers about the risk of irritating customers and damaging the relationship by making incorrect assumptions about customer preferences. Aristotle’s definition of friendship suggests that, in a good relationship, a company offers targets what they want. In such a relationship, knowledge of customer preferences must be gathered. A lack of customer knowledge indicates little attention for the customer-company relationship, and may lead relationship

management practices to backfire because of the company’s perceived self-interest (Sela, Simonson & Kivetz, 2003). When consumers are convinced of a marketer’s self-interested behavior, they prefer a less unique customized offer because, in their view, it gives them advantages that were ‘unforeseen’ by the marketer (Friestad & Wright 1994; Wright 2002). Sela, Simonson & Kivetz (2003) have demonstrated this adverse customer reaction to customized offers. In their field experiment, people were willing to accept a neutrally framed attribute offer more often (27%) than when the offer was aimed at customer’s specific attributes (4%).

Even when a company’s assumptions are correct, customized offers may backfire to the company if targets perceive intrusive behavior of the company because of company’s knowledge of their behavior (van Doorn & Hoekstra, 2013). Therefore, a customer’s trust is important when

Acceptance

An evaluation represents an intention while the acceptance of an offer indicates actual behavior. Therefore, acceptance is a concrete outcome of evaluation. An offer can be positively evaluated but still be rejected and vice versa. Accordingly, acceptance is also a simplified approximate of

evaluation and is measured in addition to evaluation.

Choice

As shortly discussed in the introduction section, a customized offer may provide the target the freedom to choose their exact offer from a choice set. Under the single offer condition the most preferred option is offered without a choice set, thus without freedom of choice, the target can only decide whether or not to accept. Under the attractive alternatives condition, the choice set consists of the most preferred option and the subsequent mostly preferred options. Under the unattractive alternatives condition the most preferred option is in the choice set accompanied with the least preferred options. Freedom of choice will have an influence on both a customized offer’s actual attributes and perceived fit with target preferences, thereby also influencing the evaluation. The actual attributes are directly determined by the targets’ choices (Kramer, 2003). Because the target has personally influenced the offer, the perceived fit with preferences will be higher (Levitt, 1960).

Consumer choice is said to be the cornerstone upon which the private enterprise system rests (Elliot, 1965). In today's market democracy, people face an ever-increasing range of options across multiple domains to choose from. This range includes careers, places to live, holiday destinations, and an immeasurable number of consumer products (Scheibehenne, Greifeneder & Todd, 2010). The preference for freedom of choice can be explained by classic psychology research, which has demonstrated that people prefer making their own choices to having them externally dictated (Botti & Iyengar, 2006). Three psychological processes related to the concept of choice should be

from behavior (Kivetz & Simonson, 2003). In any case, choice eventually results in increased psychological well-being (Cialdini, Trost, & Newsom, 1995).

In summary, freedom of choice lead to a more positive evaluation compared to no choice. Moreover, these effects seem especially applicable to the evaluation of a customized offer. Therefore, freedom of choice is in this paper defined as the provision of a choice set instead of a single offer. Thus, the following hypothesis will be the starting point of this paper:

Hypothesis 1: Freedom of choice has a positive relation with evaluation.

While common sense suggests that consumers that more closely consider their options make more satisfying decisions (Carmon, Wertenbroch & Zeelenberg, 2003), deliberating on a difficult choice can have undesirable consequences because it may cause confusion, poor decisions, and, eventually, dissatisfaction (Iyengar & Lepper, 2000). Moreover, deliberation increases the appeal of the

disregarded option and causes the target to experience after-choice discomfort (Carmon, Wertenbroch & Zeelenberg, 2003). In this case, choosing can feel like losing because choosing option A excludes option B. As a consequence, targets avoid choosing and, thereby, the possibility of experiencing dissatisfaction with their choice (Chernev, 2003; Shafir, Simonson, & Tversky, 1993). Therefore, offers with a difficult choice are not always positively evaluated. As mentioned before, under the attractive alternatives condition, all options are relevant. Next to an increased attractiveness such a choice set leads to a more difficult choice. On the other hand, under condition of the unattractive alternatives choice set, the addition of unattractive option decreases the

attractiveness of the offer but makes the choice easier. Additionally, offering targets a choice in general can negatively influence the extent to which targets feel addressed by the customized offer. That may consequently lead to a lower perceived fit and, thus, to a less positive evaluation

(Simonson, 2005).

In addition to positive effects, choice may thus have negative effects that lead to negative

evaluations. It is important to consider under what conditions negative effects have greater impact and under what conditions positive effects have greater impact. This may reasonably depend on target characteristics. SCC is in particular important because its connection with stability of, and insight in, preferences.

Self-concept clarity

The self-concept encompasses the ways in which people represent themselves – the fluid collection of schemas, views, beliefs, preferences, and aspirations that comprise a person’s sense of identity (Markus & Wurf, 1987) and controls the processing of self-relevant information (Kihlstrom & Cantor, 1984; Markus, 1977). The conceptualization allows a distinction between the contents and structure. The contents include knowledge and evaluative components on the self-concept such as self-esteem (Campbell et. al, 1996). The structure explains how the content is organized and can exist as, as examined in this paper, self-concept clarity (Campbell et. al, 1996). It is thus a

characteristic of people's beliefs about themselves (i.e. their self-concepts). It is a relatively stable personality trait and correlates with stability of preferences (Kihlstrom & Cantor, 1984; Markus & Wurf, 1987).

SCC is thus related to the structure of the self-concept and can be defined as; “the extent to which an individual’s self-beliefs are clearly and confidently defined, internally consistent, and temporally stable” (Campbell et. al, 1996, p.141). Simonson (2005) sums up two problems that hamper the provision of a customized offer that will fit the target’s needs: (1) the degree to which consumers have stable, well-developed preferences and (2) the consumers’ self-insight into those preferences, including their perception of the stability and clarity of their preferences. The definition of SCC entails exactly the problems mentioned by Simonson and is, therefore, highly applicable to the context of customized offers.

As previously discussed, a customized offer is generally positively evaluated but it may sometimes generate negative reactions. Incorrect assumptions and perceived intrusive behavior could be factors that have a negative effect on evaluation. Moreover, choice can generate both positive and negative reactions from a customized offer’s target. The following paragraphs contain predictions about the interaction between choice condition and SCC on the evaluation of the customized offer. Only the two conditions with a choice set, and thus the freedom to choose, will be discussed because there is no theoretical support for any interaction effect between SCC and evaluation when a single offer is provided.

Attractive alternatives

caused by choice stress or by the feeling of loss that arises when alternative(s) disappear upon choosing one option, are especially applicable for targets that do not know their preferences (Chernev, 2003; Shafir, Simonson, & Tversky, 1993). Moreover, the positive effects of choice are less clearly present because an explanatory process for the positive effects of choice is that people prefer to behave consistently and that choice provides the opportunity for this. Individuals with lower SCC, however, feel more uncertain about how to behave in a manner that is consistent with their self-beliefs (Setterlund & Neidenthal, 1993). Consequently the process of self-determination is arguably less pleasant since these targets are uncertain about what determines them (Berkowitz & Lundy, 1957).

Thus, under this condition negative effects are not a given for customized offers, and people with lower SCC will experience more negative effects and fewer positive effects of choice, therefore, evaluating an offer less positively.

On the other hand, negative reactions to choice are, arguably, less applicable to targets with higher SCC. Targets with high self-esteem know what they want, do not experience the choice as difficult, and assess it confidently, which is consistent with previous behavior (Setterlund & Neidenthal, 1993). Therefore, they will not feel as much after-choice discomfort in letting other options pass by because they know that these were less attractive to them. Chernev and Hamilton (2009) state that fewer options are better evaluated when they are relevant. This demonstrates that a difficult choice does not always lead to negative reactions. In fact, people feel personally addressed when all choices are relevant, resulting in a more positive evaluation.

Note that, regardless of personality, the positive effects of customized communication are present because the company is willing to sacrifice for the sake of an enduring relationship. Moreover, remember from Aristotle that in a relationship each party grants the other ‘the best’, the attractive alternatives represent a close fit to target needs and thus what is ‘best’ for the target. However, because targets still can make their own choice, the company may not be perceived as self-interested or intrusive.

Thus, it could be argued that the positive effects of a customized offer are undiminished while the negative effects are avoided. Higher SCC results in fewer negative effects and more positive effects of choice during the evaluation of the attractive alternatives. Therefore the hypothesis reads:

Unattractive alternatives

It is also worth considering whether a customized offer providing unattractive alternatives still provides positive evaluations. The choice complexity decreased because the choice includes unattractive alternatives that are explicitly not preferred. Thus, the negative effects of choice that arise due to choice complexity will be negated, even for people with low SCC (Rosenbloom & Hallman, 1981). However, positive effects of freedom of choice are undiminished applicable. As previously mentioned, self-esteem and self-concept clarity are correlated (Campbell, 1990; Nezlek & Plesko, 2001; Stinson, Wood, & Doxey, 2008). The concept clarity of people with low self-esteem are characterized by relatively high levels of uncertainty, instability, and inconsistency (i.e. low levels of clarity), and thus exhibit greater susceptibility to external influence than individuals with high self-esteem (Berkowitz & Lundy, 1957; Janis, 1954). Consistent with these findings, it has been shown that individuals with low self-concept clarity were more likely to accept a false generic, but seemingly specific, personality description (Cuperman, Robinson & Ickes, 2014). Because their self-concept is inconsistent, uncertain, and instable, it could be argued that the external influence of the offer itself, changed targets’ preferences towards the presented options, which leads to a higher perceived fit than actually the case. Because the individuals do not explicitly view the less-preferred options as incorrect assumptions of their preferences, negative effects of a customized offer are evaded.

Targets with higher self-concept clarity have a better understanding of their preferences (Kihlstrom & Cantor, 1984; Markus & Wurf, 1987) and evaluate an offer as positive when it fits their needs (Simonson, 2005). When the majority of options is less attractive, targets with higher SCC are more likely to notice the incorrect assumptions about their preferences. They may question and mistrust a company that offers random products or services. It is reasonable to argue that this type of offer encourages the perception of self-interested and intrusive company behavior. According to

Simonson (2005), when targets receive such an offer unasked, the risk of damaging the relationship is assumed to be high and is labeled as abuse of the relationship by van Tongeren (2010).

Thus, despite the positive effects of choice, possible negative effects of customized offers become more applicable when SCC is higher. Therefore the hypothesis reads:

Loyalty

Before the construct ‘loyalty’ and the application for the marketing and HRM setting will be

discussed, it is important to understand that loyalty arises, via satisfaction, out of the evaluation of a customized offer. Literature on customer satisfaction research recognizes two basic

conceptualizations of satisfaction: transaction-specific and cumulative satisfaction (Johnson, Herrmann & Gustafsson, 2002). The transaction-specific approach defines satisfaction as the targets’ evaluation of their experience with, and reactions to, a product (Olsen & Johnson, 2003). Cumulative satisfaction describes the customers’ overall consumption experience with a product or service over time. Satisfaction can, thus, be seen as an attitude based on previous experiences (Bolton, 1998).

The definition of loyalty can be rather unclear when discussing human relationships (Hart & Thompson, 2007). Loyalty is a broad psychological concept that can have various applications across different settings. This is reflected in literature by different measurements of both customer and employee loyalty. A general definition of loyalty is provided by Hart and Thompson (2007, p.297) who state, “Loyalty is an attitude that resides in the mind of the individual and is an individual-level construction of perceived reciprocal obligations”. In addition to attitude, these ‘reciprocal obligations’ imply a behavioral aspect. It is unlikely that a target can reciprocate only with an attitude; an attitude must be communicated through behavior. Indeed, a compositional approach of attitudinal and behavioral aspects is recommended when investigating the concept of loyalty (Dick & Basu, 1994).

Most often, in research on customer loyalty, three aspects are measured: intention to stay in the relationship, willingness to pay more, and customer advocacy intentions (Fornell , Johnson,

Anderson & Bryant, 1996). Willingness to stay in the relationship may result in the self-sacrificing customer behavior (Cialdini, 2009). Repurchasing at and same company despite higher prices is an example of sacrificing behavior. According to Morgan and Rego (2006), this intention to

repurchase is the most important indicator of customer loyalty. The final behavioral aspect of customer loyalty is the willingness to recommend a company to a friend or colleague.

To define employee loyalty, Meyer and Allen’s (1991) three-component structure is most often used (Jak and Evers, 2010) because it is known to occupy much of the same conceptual space as loyalty (Hart & Thompson, 2007) and can even be viewed as synonyms (Kumar & Shekhar, 2012). In Meyer and Allen’s (1991) commitment construct there is a distinction between three components of commitment: affective, continuance, and normative commitment:

“Affective commitment refers to the employee's emotional attachment to, identification with, and involvement in the organization. Employees with a strong affective commitment continue employment with the organization because they want to do so. Continuance commitment refers to an awareness of the costs associated with leaving the organization. Employees whose primary link to the organization is based on continuance commitment remain because they need to do so. Finally, normative commitment reflects a feeling of obligation to continue employment. Employees with a high level of normative commitment feel that they ought to remain with the organization.” (Meyer and Allen, 1991, p. 67)

It is important to note that the first two components have strong attitudinal aspects. The willingness to stay in the relationship may result in self-sacrificing employee behavior (Zhang et al., 2014). The behavioral aspect of employee loyalty is, therefore, related to the intention to remain employed at a company (Costen & Salazar, 2011).

Loyalty is an attitude based on enduring satisfaction, which results in proactive behavior that benefits the future relationship. Thus, satisfaction is a long-term and extended form of evaluation, and loyalty is a long-term and extended form of satisfaction. This results in the following

hypothesis:

Hypothesis 3: There is a positive relationship between evaluation and loyalty.

When an offer is positively evaluated, it results in a positive attitude towards the company,

especially after several positive experiences. The constructs of evaluation, satisfaction and loyalty are extensions of one another. Therefore, the positive relation of freedom of choice on evaluation should also have its impact on loyalty, which imply the following hypothesis:

Hypothesis 4a: The relationship between choice and loyalty is mediated by evaluation.

effects will indirectly, via evaluation, be applicable on loyalty. This leads to the following hypotheses:

METHOD

Participants and research design

The data was gathered with an experiment conducted among bachelor students at the University of Groningen. There were 416 respondents between the ages of 17 and 33 (M=20.7; SD=2.58), of which 222 were male and 194 were female.

First, respondents’ preferences were measured to enable the researchers to customize the offering. Next, self-concept clarity was measured in order to investigate possible moderating or interacting effects. Subsequently, respondents received a customized offer and were asked to evaluate this offer. Lastly, respondents’ loyalty towards the provider of the offer was measured. Respondents received two customized offers – one marketing offer and one HR offer.

The customized offers could be presented within three conditions: attractive alternatives,

unattractive alternatives, and single offer. Under the attractive alternatives condition, respondents was a choice provided between their four highest preferred options. Under the unattractive

alternatives condition, one of the options was their most preferred option but the three other options were least preferred. Finally, for the single offer condition, respondents only received their

preferred option. For every condition respondents could choose to reject the offer. The manipulation of the conditions was checked by investigating the perceived difficulty of the choice with

statements such as, “It was clear which option was the most attractive” (for marketing: 3 items, α=.79; for HRM: 7 items, α=.80).

Respondents received a pre-existing questionnaire on self-concept clarity. The questionnaire, taken from Campbell (1996), is widely used and includes statements such as, “On one day I might have one opinion of myself and on another day I might have a different opinion” (12 items, α=.79).

Procedure

Respondents’ ranking for nine options was needed. In the marketing setting, these nine options were represented by nine magazines. In the HRM setting, the nine options were nine secondary

According to Dutch websites, www.tijdschrift.nl and www.123tijdschrift.nl, the nine magazines below are unisex and highly popular. They were, therefore, chosen for use in this experiment:

 VT wonen  Runnersworld  National Geographic  Privé  Televizier  Elsevier  Delicious  Donald Duck  Quest

For the additional benefits, nine options relevant to the Dutch business environment and

recognizable for the participating students were chosen. Moreover, the following employee benefits are often used in flexible benefit programs (Yamamoto, 2011; Milkovich & Sturman, 1995)

 Extra days off

 Facilities for leisure activities (sport and recreation)  (contribution to and help with) Medical insurance  (contribution to and help with) Dentist insurance  (More expensive) Lease car

 (contribution to and help with) Life insurance  (contribution to and help with) Accident insurance  Stock options

 (contribution to and help with) Lodging

Respondents then received an offer as if it had been sent by e-mail. The offer was explicitly presented as a customized offer aimed at a long-term customer/employee relationship, and formulated to trigger feelings of reciprocity by including the sentence, “We hope the pleasure of having you as our customer/employee will stand for a long time, therefore, especially for you: actual offer”. For the marketing setting, the actual offer entailed a 25% discount on a six-month magazine subscription. In the HRM setting, salary could be transferred into secondary benefits. The monetary value of the benefits was 150% of the transferred salary.

Thus, the four most-preferred and three least-preferred options were used for the experiment. It should be noted that nine options were considered during the preference measurement; this left two unused ‘middle’ options. These ‘middle’ options were assumed to be neutral for the respondents and imply that the better than neutral preferred options are truly preferred and the less than neutral preferred options are truly not preferred.

The order of the marketing and HRM settings was counterbalanced to overcome biased order effects. The condition under which the respondents received their offer was randomly assigned for both settings.

Measures

DV: Evaluation. After receiving the first customized offer (either the marketing or HR offer) respondents were asked to evaluate it. The evaluation included the most important statements about overall evaluation, evaluation relative to respondents’ expectations, and evaluation relative to ideal products (Fornell et. al, 1996). These were also used in the American Customer Satisfaction Index and consisted of statements such as “I am satisfied with this offering” and “This offering fits with my preferences” (for marketing: 6 items, α=.90; for HRM: 6 items, α=.94 ). The same measurement was used for both settings.

DV: Acceptance. Acceptance was a dichotomous variable measuring only whether or not a

respondent accepts the customized offer. Acceptance measured ‘actual’ behavior – either rejection or acceptance of the offer.

DV: Loyalty. In the marketing setting, customer loyalty was also measured. The customer loyalty construct contained three parts: repurchase intention, willingness to pay more, and likelihood of recommendation. According to Reichfeld (2003), willingness to recommend is the only item you have to measure to determine customer loyalty. Morgan and Rego (2006) question these findings, however, demonstrating that repurchase intention is most important (2 items). However, many scholars, such as Ittner and Larcker (1998) advocate likelihood of recommendation as a

respondents’ willingness to remain a customer (3 items) despite price increases of 5%, 10%, and 15% (Ittner & Larcker, 1998). To ensure that the altitudinal aspect of customer loyalty was

measured, Allaway, Whipple and Ellinger’s (2011) construct of emotional attachment (4 items) to a company was added. This was measured using statements such as, “I care about the long term success of this company”. Altogether, 10 items were used to construct one customer loyalty scale; α=.85.

In the HRM setting, employee loyalty was measured. Meyer and Allen’s (1991) three-component commitment measurement is the most common measurement of employee loyalty (Jak & Evers, 2010). The measurement included questions on affective aspects (6 items), such as: “I feel ‘part of the family’ in this organization” and normative aspects (5 items), such as: “I was raised with the thought that it is worthwhile to stay loyal to an organization”. Although the construct includes a third component – the continuance component – this was excluded from the measurement. In an experiment, respondents must empathize with a scenario and the items of the continuance

component were assumed to be too farfetched for this setting. Meanwhile, the continuance aspect was measured using Costen and Salazar’s (2011) construct on willingness to stay with the

TABLE 1

Experienced difficulty in marketing setting

RESULTS

The results of the experiment in the marketing setting and, subsequently, the human resources management setting are reported in this section.

Marketing setting

To check whether the choice manipulation was successful, the experienced difficulty of the choice was investigated. A linear regression analysis was performed with experienced difficulty as the dependent variable. The predictors were two dummy variables for the single offer and unattractive alternatives conditions, the standardized Z-score for self-concept clarity, and two interaction

variables for the two conditions with SCC. When the dummy of the condition took the value of one, only the effects of that condition were measured because the other two conditions then took the value of zero. This enabled us to compare each of both conditions to the attractive alternatives condition. Additionally, gender and the order in which the respondents completed the two settings were included as control variables. As can be seen in Table 1, the unattractive alternatives condition differed significantly from the attractive alternatives condition. In other words, the respondents experienced less difficulty with the unattractive alternatives choice set than with the attractive alternatives choice set. This indicates that the manipulation was successful. Moreover, SCC was shown to have a negative relationship with experienced difficulty. That means that, for the attractive alternatives choice set condition, respondents with a lower SCC experienced more difficulty in making a choice. Furthermore, this negative relationship between SCC and experienced difficulty was found to exist across all settings, regardless of the condition.

Evaluation

Given that we investigated customized offers, on average these offers should be seen as reasonably attractive. In order to remain realistic, they could not be too attractive but in order to prompt reactions leading to relevant results for offer evaluation and, in particular, target loyalty, the offer

needed to be somewhat attractive. This appeared to be the case. On a 10-pointscale, average offer evaluation was on 7.04, and 63% of respondents gave a rating of 7 or higher (Morgan &Rego, 2006). Further, 41% of respondents accepted the offer.

To test the hypotheses concerning evaluation, regression analyses on acceptance and overall evaluation were performed. Since an offer could be simply rejected or accepted, the offer acceptance variable is a dichotomous dependent variable. Therefore, a binary logistic regression was applied. The broader construct of evaluation was analyzed using linear regression, with the overall evaluation of the offer as the dependent variable. For both analyses the same predictors were included: two dummy variables for the attractive alternatives and unattractive alternatives

conditions were computed to analyze differences compared to the single offer condition, with respect to the dependent variable. When the dummy of the condition took the value of one, only the effects of that condition were measured because the other two conditions then took the value of zero. This enabled us to compare each of both conditions to the single offer condition, two

independent regression analyses were performed to investigate the interaction terms for both choice set conditions separately. Moreover, standardized SCC scores and the interaction between dummy variables and SCC were included. Gender and the order in which the settings were completed by respondents were employed as control variables. Outcomes are shown in Table 2.

TABLE 2

Hypothesis 1 states, “Choice has a positive relationship with evaluation”. For the evaluation construct, overall evaluation and acceptance can be examined. In the experiment, the unattractive and attractive alternatives conditions did not include more respondents accepting the offer than under the single offer condition. The dummy variable for the unattractive alternatives condition was twice significant; the dummy variable for the attractive alternatives condition was not. This means that the evaluation under the unattractive alternatives condition was more positive than the single offer. However, the evaluation did not differ under the attractive alternatives condition. Tough the results were mixed, the hypothesis was rejected.

Hypothesis 2a predicted a positive relationship between the attractive alternatives choice set and evaluation, but only for people with high SCC. Similarly, hypothesis 2b predicted a positive

relationship between the unattractive alternatives choice set and evaluation, but only for people with low SCC. In Table 2, the regression results are summarized. There were two independent regression analyses performed to investigate the interaction terms for both choice set conditions separately. The table demonstrates that none of the interaction terms between a dummy variable and SCC had a significant impact on evaluation. This means that individual characteristics did not influence the evaluation for any of the conditions. Therefore both hypotheses were rejected. The order of the experiments, however, did have an effect on acceptance but at a lower reliability interval with α=.10. The experiment illustrated that when the marketing setting was presented second more marketing offers were accepted.

Loyalty

To investigate whether offer characteristics, individual characteristics, and offer evaluation predict loyalty, linear regression analysis was performed using customer loyalty as the dependent variable. The abovementioned predictors were included and this time, overall evaluation, previously used as a dependent variable, was added as predictor. The results are summarized in Table 2. The table illustrates that in step 1 only the same predictors as for evaluation were investigated, while in step 2 the evaluation of the offer was added as predictor of loyalty.

Hypotheses 4a, 4b, and 4c concerned a mediation effect, whether or not moderated, on loyalty. Since a significant relationship between choice and evaluation was requisite but not found in the marketing setting, those hypotheses had to be rejected.

The linear regression analyses, using evaluation and loyalty as dependent variables, were performed with dummy variables for the unattractive alternatives and attractive alternatives conditions. As mentioned, those dummies allowed for a comparison to the single offer condition. Despite not being predicted by the theory, there may be differences between the unattractive and attractive alternatives conditions. To investigate those possible differences, a one-way ANOVA test and a Tukey post hoc test were conducted. These provided no significant results, demonstrating that there were no

HRM setting

To determine whether the choice manipulation was successful, the experienced difficulty of choice was investigated. A linear regression analysis was performed with experienced difficulty as the dependent variable. The predictors were two dummy variables for the single offer and unattractive alternatives conditions, the standardized Z-score for self-concept clarity, and two interaction

variables for the two conditions with SCC. When the dummy of the condition took the value of one, only the effects of that condition were measured because the other two conditions then took the value of zero. This enabled us to compare each of both conditions to the attractive alternatives condition. Additionally, gender and the order in which the respondents completed the two settings were included as control variables. As can be seen in Table 3, the unattractive alternatives condition differed significantly from the attractive alternatives condition. In other words, the manipulation was successful because the respondents experienced less difficulty within the unattractive alternatives condition than within the attractive alternatives condition.1 Furthermore, male respondents experienced more difficulty than their female counterparts with choice in the HRM setting.

TABLE 3

Experienced difficulty in HRM setting

Evaluation

As with the marketing setting, the attractiveness of the offer was investigated first to ensure that the offer was not too attractive or too unattractive. This appeared to be the case. On a 10-pointscale, average offer evaluation was on 7.14, and 56% of respondents gave a rating of 7 or higher (Morgan &Rego, 2006). Further, 61 % of respondents accepted the offer.

To test the hypotheses on evaluation, regression analyses were performed to measure acceptance and overall evaluation. Because an offer could be simply rejected or accepted, the offer acceptance

1

1) Zero-order correlation between SCC and experienced difficulty was significant and negative.

variable is a dichotomous dependent variable. Therefore, a binary logistic regression was

performed. The broader construct of evaluation was analyzed using linear regression analysis, with overall evaluation of the offer as the dependent variable. For both analyses the same predictors were included: two dummy variables for the attractive and unattractive alternatives conditions were computed to analyze the effect of different conditions on the dependent variable. When the dummy of the condition took the value of one, only the effects of that condition were measured because the other two conditions then took the value of zero. This enabled us to compare each of both

conditions to the single offer condition. Also included were standardized Z-score on SCC and the interaction of conditions’ dummy variables with SCC. Again, gender and the order of the settings were included as control variables. Outcomes are displayed in Table 4.

Hypothesis 1 states “Choice has a positive relationship with evaluation”. Overall evaluation and acceptance should be examined when measuring evaluation. The Wald statistics for acceptance and the B-values for evaluation in both choice set conditions were significant for both linear regressions, tough for one of the linear regressions the unattractive alternatives choice set only was significant at α=.10. This means that more respondents accepted the offer under both the unattractive and

attractive alternatives conditions than for the single offer condition. Furthermore, in the choice set conditions the evaluation was more positively than in the single offer condition. Therefore,

hypothesis 1 was accepted.

Hypotheses 2a and 2b predicted a relationship between SCC and evaluation, under a specific condition. Hypothesis 2a reads: “Freedom of choice has a positive relationship with evaluation, but when a choice set includes attractive options it only applies to people with high SCC” and

TABLE 4

Regression on acceptance, evaluation and loyalty

Figure 2. Effect of freedom of choice of an attractive alternatives choice set on evaluation, moderated by SCC.

Moreover, gender and the order of the experiments had an effect on evaluation and acceptance. When the HRM setting was presented first, or respondents were female, the evaluation was higher and more offers were accepted in the HRM setting.

Loyalty

In order to investigate whether offer characteristics, individual characteristics, and offer evaluation were predictors of loyalty, a linear regression analysis was performed with employee loyalty as the dependent variable. The abovementioned predictors were included and overall evaluation,

previously used as the dependent variable, was added as predictor. Results are summarized in Table 4, which illustrates that in step 1 only the same predictors as for evaluation were investigated, while in step 2 the evaluation of the offer was added as predictor of loyalty.

Hypothesis 3 predicted a positive relationship between evaluation and loyalty. This hypothesis was supported by the regression analysis results in Table 4. The differences in R square of the

regressions means that evaluation was responsible for 9% of the differences in loyalty.

Additionally, respondents who accepted their offer were significant more loyal than those who refused. Thus, hypothesis 3 was accepted.

The linear regression analyses, using evaluation and loyalty as dependent variables, were performed with dummy variables for the unattractive alternatives and attractive alternatives conditions. As mentioned, those dummies allowed for a comparison to the single offer condition. Despite not being predicted by the theory, there may be differences between the choice set conditions. To investigate

Figure 3. Effect of freedom of choice of an unattractive alternatives choice set on evaluation, moderated by SCC.

those possible differences, a one-way ANOVA test and a Tukey post hoc test were conducted. These provided no significant results, indicating that there were no significant differences between the two conditions.

So far, the results for the HRM setting showed that SCC moderates the relationship between choice and evaluation, and evaluation predicted loyalty. Therefore, it was worthwhile to investigate the mediation of evaluation on the relationship between choice and loyalty, whether or not moderated, as stated in hypotheses 4. To investigate the mediation effect, multiple regression analyses was performed with loyalty as outcome variable and evaluation as mediator. The SPSS macro as discussed in Preacher and Hayes (2004) was used to obtain a bootstrapped confidence interval. With bootstrapping, no assumptions about the shape of the sampling distribution of the statistic are necessary because the sample is seen as a pseudo population. Two dummy variables were computed for the choice conditions. When the dummy of the condition took the value of one, only the effects of that condition were measured because the other two conditions then took the value of zero. This enabled us to compare each of both conditions to the single offer condition, two independent analyses were performed to investigate both choice set conditions separately. Additionally, gender and order were included as control variables. To investigate the conditioned indirect effects, a multiple regression was performed twice. The same variables were included as for the

unconditioned indirect effect as abovementioned, and SCC was added as moderator. Results are summarized in Table 5.

TABLE 5

95% confidence intervals for mediation effects and moderated mediation effects

Table 5 demonstrates the boundaries of the effects on a 95% confidence interval, this means that on significance level α=.05 the B-value is between those boundaries. Hypothesis 4b states, “The relationship between choice and loyalty is mediated by evaluation, but when a choice set includes attractive options it only applies to people with high SCC.” Under the condition of an attractive alternatives choice set the direct relationship between choice and loyalty did not significantly differ from a single offer because the value of zero is in the confidence interval, thus no support for a direct effect was found. The confidence interval for an indirect effect, however, did significantly differ from zero because the interval did not include the value of zero, thus support was found for a mediation effect. The confidence interval for an effect of the moderated mediation model was between .00 and.08. This means that the effect did significantly differ from zero, thus support for the moderated mediation was found. In other words; a more positive evaluation led to higher loyalty in the attractive alternatives choice set than without the freedom of choice, but only for people with high SCC. Therefore, hypothesis 4b was accepted.

Hypothesis 4c states, “The relationship between choice and loyalty is mediated by evaluation, but when a choice set includes unattractive options it only applies to people with low SCC.” As displayed in Table 5, in the unattractive alternatives choice set condition the direct relationship between choice and loyalty did not significantly differ from zero because the confidence interval included the value of zero. Therefore, no support was found for a direct effect. The confidence interval for a conditioned indirect effect did significantly differ from zero, but only for low SCC. The confidence interval for an effect of the moderated mediation model was between -.10 and -.02. This means that the effect did significant differ from zero and thus support was found for the moderated mediation. In other words; a more positive evaluation led to higher loyalty in the

unattractive alternatives choice set than without the freedom of choice, but only for people with low SCC. Therefore, hypothesis 4c was accepted.

Effects: Attractive choice set

alternatives Unattractive choice set

alternatives

Lower limit Upper limit Lower limit Upper limit

Direct on loyalty B = -.21 B = .06 B = -.16 B= .09

Indirect via evaluation; B= .02 B = .11* B= -.04 B = .04 - Under condition of high

SCC

B= .06 B= .20* B= -.08 B = .04

- Under condition of low SCC

B= -.01 B= .12 B = .04 B = .17*

DISCUSSION

For both human resource management and marketing environments, personalization has been a popular trend. Due to technological developments, communication can be personalized at a low cost as, for example, a customized offer. In this experiment, targets received a customized offer in one of two settings – marketing or HRM. This paper offers several hypotheses concerning how, and for whom, a customized offer is most effective to achieve a more positive evaluation and higher loyalty. The hypotheses were then tested by providing targets with a choice set in their customized offer and by means of an existing questionnaire for self-concept clarity. Self-concept clarity (SCC) refers to “the extent to which an individual’s self-beliefs are clearly and confidently defined, internally consistent, and temporally stable” (Campbell et. al, 1996, p.141).

It was expected that the provision of a choice set would result in a more positive evaluation and higher loyalty than a single offer would. Moreover, it was expected that adding attractive

alternatives to the choice set would lead to the increased attractiveness of an offer but also to a more difficult choice. The addition of unattractive alternatives, however, was expected to decrease the offer’s attractiveness while making the choice easier. A difficult choice is especially detrimental to people with lower SCC. Therefore, it was proposed that a customized offer with a choice set

including several attractive options would lead to more a positive evaluation and higher loyalty, but only for people with high self-concept clarity. An offer with an easier choice is particularly

beneficial for people with lower SCC, while the unattractive options will be quickly disregarded by people with higher SCC. It was, therefore, proposed that a customized offer with a choice set including several unattractive options would lead to more a positive evaluation and higher loyalty, but only for people with low self-concept clarity.

In the marketing setting, the manipulation of conditions was successful and the offers’

attractiveness was sufficient. The offer evaluation predicted the targets’ loyalty towards a company. However, mixed support was found for the positive effect of freedom of choice on the evaluation of the offer. This finding was not sufficient, however, to confirm the relationship between choice and evaluation. Additionally, no support was found for an interaction effect of SCC and the choice set on evaluation. Subsequently, no mediation effect of evaluation, whether or not moderated, could be significant for the relationship between choice and loyalty.

In the HRM setting, the manipulation of conditions was also successful and the offers’

predictor of loyalty. The hypothesized positive relationship of choice on evaluation within the unattractive alternatives condition, only for people with low SCC, was supported. On a lower reliability interval, with α=.10, support was also found for the positive relationship of choice on evaluation within the attractive alternatives condition, only for people with high SCC. Moreover, these moderated effects on evaluation had an indirect effect on loyalty. This means that choice may lead to higher loyalty through more positive evaluation, but only for people with high SCC when the choice set included attractive alternatives, and only for people with low SCC when the choice set included unattractive alternatives. However, the overall mediation of evaluation on the

relationship between choice and loyalty was not supported.

Especially in the marketing setting, few significant relationships were found. This may be explained by the lower perceived relevance and fewer acceptances of the marketing offer. When targets were asked to indicate the offer’s relevance on a 10-point scale, the marketing offer was assessed substantially less relevant (4.19) than the HR offer (6.70). Moreover, only 41% of respondents in the marketing setting accepted the offer, while 61% of respondents in the HRM setting accepted the offer. The combination of lower relevance and fewer acceptances could have resulted in more mediocre answers. Indeed, the SD for the evaluation was lower in the marketing setting (1.74 versus 1.99), which means that respondents’ evaluations were slightly more univocal, possibly generating fewer relevant results.

One possible explanation for the difference in acceptance rates between the two settings could be the value, whether absolute or relative, of the offer. The absolute difference in monetary value is caused by the subjects of the two offers – a magazine subscription in the marking setting versus the more substantial employment conditions in the HRM setting. In terms of relative value, the

marketing offer was also less attractive because it offered a 25% discount on a subscription as opposed to a 50% upgrade in the monetary value of target’s salary. Moreover, the overall evaluation of the marketing offer was relatively positive and comparable to that of the HRM offer but its relevance was substantially lower. This could reflect the targets’ disinterest in the offer subject, which may have led to inattentive participation in the experiment and, therefore, fewer relevant results.

Offer characteristics, individual characteristics, and the interaction between these two

an unattractive alternative choice set is provided to people with low SCC, their more positive evaluation may lead to higher loyalty.

Implications

An implication for literature on relationship management and, in particular, on customized offers is the influence of freedom of choice. This paper shows that the provision of a choice set is valued by the targets, but only for the HRM setting. Previous research showed that targets were more likely to buy the most attractive option when an offer provided a choice including a less-attractive option than when an offer did not (Simonson and Tversky, 1992). This finding is supported for customized offers by the research in this paper, though only in the HRM setting, which has additionally

demonstrated that including several attractive options also may result in higher sales. More positive evaluation and acceptance may lead to higher loyalty, which may be caused by an unattractive alternative choice set for people with low SCC and an attractive alternative choice set for people with high SCC. Thus, a choice set as offer characteristic may determine the success of a customized offer, although this is rarely mentioned in current literature.

Another theoretical contribution to relationship management is the relevance of SCC. This paper demonstrates that people with lower SCC had greater difficulties in making choices, reacted more positively to unattractive options, and consequently reacted differently to customized offers than people with higher SCC. Lastly, the relevance of SCC proves that personality traits may be determinants of successful customization and therefore successful relationship management; thereby it may lay the foundation for new interesting research questions.

This paper may also have implications for literature about human choice behavior, in particular for the negative effects of choice. It is known that choice can have both positive and negative effects. Negative effects can be caused by choice stress due to a difficult choice, or by the feeling of loss that arises when alternative(s) disappear upon choosing one option. In this experiment the attractive alternatives choice set resulted in a more difficult choice, thus it could have generated choice stress. Besides, targets had to make a choice thus alternatives disappeared, which could have generated feelings of loss. However, it did not result in lower evaluation or fewer acceptances, not even for people with lower SCC. Based on the results in this paper, it may be concluded that the negative effects of choice may be somewhat overemphasized and only occur in more extreme situations. The theoretical contributions of this study can be practically implemented by companies. When sending a customized offer, companies should always provide their targets with a choice set

be possible to increase target’s loyalty with use of the right options in the choice set. Moreover, because there is a substantial relationship between evaluation and loyalty, offering a customized offer should strengthen companies’ relationships with their customers and employees. The subsequent positive effects of loyal customers and loyal employees on a firm’s financial performance have been extensively addressed in this paper.

Limitations and future research

Because the experiment was conducted in a lab, respondents knew that their choice did not have a direct impact on their lives and that they would, therefore, not ‘lose’ anything with the choice they made. Besides, because the company initiated contact, the glass may be seen as half full and the choice would consequently not lead to feelings of loss, only of gain. Because the psychological needs of self-control, self-determination, and consistency were still met in this experimental setting, it can reasonably be argued that the positive effects were not diminished. Therefore, the negative effects of choice could have been overshadowed by the positive effects, resulting in an overly positive evaluation of choice. Finally, due to a specific population there could have been collective areas of (dis)interest, illustrated by a disinterest in the offered magazines.

The limitations of the experimental setting can easily be countered in a field study because a choice does have real impact then. Another advantage is that a field study can provide richer insights into the assumed relationship that SCC has with preference stability and insight. In a marketing setting, it would be valuable to know whether people with unstable (purchasing) behavior can still be successfully targeted using customized offers. Based on the conclusions of this study, unstable behavior is caused by vague preferences and suggests low SCC. If such a customer still regularly buys certain specific products, these can be incorporated into a customized offer, rendering it effective for customers with unstable behavior. This would imply that as long as information about past behavior is available, companies should always send customized offers because it will be effective.

This paper suggests that freedom of choice may be a determinant of a successful customized offer. Therefore, it is valuable to develop a better understanding of why this is the case. As mentioned in the theory section, it is known that receivers of personalized communication may perceive