A sender’s motivation perspective
Master Thesis Completion Date October 21, 2009 University of Groningen Faculty of Economics and Business Master Business Administration Department Marketing Management
Student Karlijne Bos Willem Barentszstraat 18bis 3572 PJ Utrecht 06 52640683 karlijnebos@hotmail.com s1256653 Supervision 1st supervisor: Dr. M.A. Tuk M.A.Tuk@rug.nl 2nd supervisor: Dr. S. Gensler S.Gensler@rug.nl
Company: Schwarzkopf & Henkel Company supervisor: Drs. A.N. Toering
Word-of-Mouth Communication
A sender’s motivation perspective
Karlijne Bos
University of Groningen, The Netherlands
Abstract
The purpose of this research was to empirically test whether a company could actually activate customer referrals by designing WoM campaigns that appeal to the basic motives underlying WoM behavior. Results could not confirm the hypothesis that messages sent by a company to customers based on underlying motives to engage in positive WoM behavior, significantly enhance the WoM communication of these customers. Although earlier research has confirmed that customers may talk about products and/or services for a variety of reasons, this research suggests that triggering these motives by making them more salient to the customer, does not enhances a customer’s of-Mouth intention and Word-of-Mouth behavior.
Keywords: word-of-mouth, motives
Stranger: “Do you know a band called Mcfly?
You: “Mcfly, hm I think I heard of it sometime before” You: “W8 a sec, google…”
Stranger: “haha lol”
Stranger: “I’m trying to convert as many people as possible into Mcfly fans” Stranger: “They are awesome”
You: “Why do you want to convert me?”
Stranger: “Because Mcfly is a great band and it deserves as many fans as it can gets☺”
Conversation at Omegle April 8, 2009
1. Introduction
Abundant research has demonstrated that Word-of-Mouth (hereafter WoM) is one of the most powerful channels of communication in the marketplace (e.g. Allsop, Bassett & Hoskins 2007; Derbaix & Vanhamme 2003). To take advantage of this, and of other benefits of WoM, marketers should attempt to build communication channels with interpersonal communicators to reach and influence a wider target market.
services (Biyalogorsky, Gerstner & Libai 2001). To encourage current customers to engender referrals, marketers offer exceptional value through either attractive prices or excellent quality to them. In addition, offering rewards to customers for referring other customers is commonly used to motivate referrals (Biyalogorsky et al. 2001).
The production of WoM in general is thought to be an outcome of customers’ experiences with a product or service. But it has been suggested by Wirtz & Chew (2002) that behavioral outcomes, such as WoM, only emerge when the satisfied customer feels motivated to do so. The existing literature on WoM reveals several underlying motives for customers engaging in WoM communication (Dichter 1966; Sundaram, Mitra & Webster 1998). However, this literature is primarily explorative in nature. In addition, given the unique characteristics of internet communication, such as, for example, anonymity, and directed at multiple persons, Hennig-Thurau, Gwinner, Walsh & Gremler (2004) suggested a typology for motives of consumer online articulation through web-based consumer- opinion platforms. What is lacking in existing research is an answer to the question whether WoM campaigns that appeal to these basic motives driving WoM behavior, could be utilized effectively to encourage customers to engender referrals. Nevertheless, the use of motives in guiding strategy formulation is not new. For example, Westbrook & Black’s (1985) taxonomy of shoppers is based on people’s underlying shopping motivations. This taxonomy has been shown very rewarding in guiding retail strategy formulation. Furthermore, designing campaigns in a more customer-oriented way corresponds to a more recent trend that has been receiving marketers’ attention, namely customer relationship management (CRM) (Hogan et al. 2004; Rust, Lemon & Zeithaml 2004).
By identifying the motives driving customers’ WoM behavior, marketers could design their WoM campaigns in a more customer-oriented way. Therefore, the purpose of this article is to empirically test whether a company could actually activate customer referrals by designing WoM campaigns that appeal to the basic motives underlying WoM behavior. In an effort to provide input to marketers interested in stimulating positive WoM, a field experiment is conducted in which consumers’ motivations to participate in WoM behavior are used to develop different messages that are sent to customers to motivate referrals. The experiment adds new insights to the literature by considering explicitly if and how different messages, which appeal to the basic motives that drive WoM communication, could influence WoM behavior. For the purpose of this study, the focus was on positive WoM creation from a sender’s point of view.
In this article it is hypothesized that messages that appeal to the basic motives that drive WoM behavior can significantly influence positive WoM communication. Furthermore, it is tested which message triggers consumers the most in their WoM behavior.
In pursuit of the objectives of this study, an introductory review on WoM communication, its antecedents, and its underlying motives will be given first. The methodology used to examine the effect of different messages, which are based on consumers’ motivations to engage in WoM behavior, will subsequently be discussed. Then, the results will be given and evaluated. This article will end with a conclusion, recommendations for future research and the limitations of this research. Theoretical and practical relevance
This article contributes to the knowledge of WoM processes from a sender’s perspective. For the most part, exploratory research designs have been used in existing literature on WoM motives. This research, however, uses a field experiment to examine the effects of motivation based messages on WoM communication of the consumer.
The article has practical relevance as well; it provides marketers with a better understanding of the different motives underlying WoM behavior of WoM sources. A better understanding of these motives can help marketers better harness WoM as a promotional tool.
2. Theoretical framework
It has generally been accepted that WoM is the most important communication source between consumers (Allsop et al. 2007; Derbaix & Vanhamme 2003). Furthermore, WoM can have a substantial influence on product choice (Kiel & Layton 1981; Price & Feick 1984). It is said that interpersonal communication is more effective in terms of influencing consumer behavior than commercial communication and that it is the cheapest manner to increase your customer base without great investment (Sheth, Mittal & Newman 1999), which is particularly attractive in times of economic downturn. As such, marketers should attempt to build communication channels with interpersonal communicators to reach and affect a wider target market. The practice of firms stimulating and accelerating WoM is also known as buzz marketing (Rosen 2009).
The popularity of buzz marketing is increasing. Obviously, internet and technology have played a major role in the rise of buzz through social networking sites (Rosen 2009; Hampton 2006), such as Facebook and Hyves. Another issue contributing to the rise of WoM, or buzz, is increasing skepticism toward more traditional forms of advertising. It has been found that consumers are skeptical of information that comes from companies (Rosen 2009) and that WoM is generally seen as more credible (Herr, Kardes & Kim 1991; Murray 1991; Gremler, Gwinner & Brown 2001; Allsop et al. 2007) because we trust people like ourselves (Rosen 2009). Yet another factor contributing to the increased attention paid to WoM marketing by marketers is the increase in media fragmentation and media clutter (Rosen 2009; Hampton 2006) Moreover, traditional media advertising is found to decline in effectiveness (Godes & Mayzlin 2004).
expected that WoM through strong ties is more persuasive than information transmitted through weak tie relationships (Cialdini & Sagarin 2005). Marketers should take this into account when designing programs to stimulate WoM communication about their products and/or services. In practice, this will not turn out to be so easy, however. Godes and Mayzlin (2009) have selected customers for a WoM campaign based on loyalty, because loyal customers are expected to be more satisfied with a product and/or service than less loyal customers (Anderson & Sullivan 1993), and, therefore, are expected to create more WoM than average (Swan & Oliver 1989). However, Godes & Mayzlin (2009) have argued that for the firm targeting less loyal customers for the WoM campaign might have a greater impact than targeting the most loyal customer. This can be explained by the fact that the network of the highly loyal customer is already saturated, putting forward fewer incremental gains from a WoM campaign. This is indeed more consistent with Granovetter’s (1973) theory of weak ties, which has been explained earlier.
2.1 Types of influencers
As information from personal, noncommercial sources has been found to be particularly influential in consumer decision making, research on interpersonal influence has focused on different types of influencers, namely opinion leaders, early adopters (Price & Feick 1987; Rogers 1995), and market mavens (Feick & Price 1987; Walsh, Gwinner & Swanson 2004).
Watts and Dodds (2007) defined opinion leaders on the basis of the number of influence relations they have. Using a series of computer simulations of interpersonal influence, they found that opinion leaders are not as effective as previously thought. However, opinion leadership implies more than just having a central position (Van Eck, Jager & Leeflang - in press). Many studies have added to our understanding of the roles and attributes of opinion leaders ever since Katz & Lazarsfeld (1955) first introduced the concept of opinion leadership (e.g. Weimann, Tustin, Vuuren & Joubert 2007). Characteristics such as interpersonal influence and innovativeness are linked to opinion leaders as well, and for this reason may have a significant effect on the influence opinion leaders have (Van Eck et al. in press). Using an agent based model Van Eck et al. (in press) conducted an experiment with the variables product knowledge, innovativeness, and interpersonal influence and demonstrated that opinion leaders increase the rapidity of the information flow and the adoption process. However, little is known about what actually motivates an opinion leader to disseminate information (Godes, Mayzlin, Chen, Das, Dellarocas, Pfeiffer, Libai, Sen, Shi & Verlegh 2005).
Substantial research attention is given also to early adopters (Rogers 1995). It is found that early adopters, or innovators, can exert either an active, through product-related conversations, or passive, by product use, influence on later adopters or purchasers (Feick & Price 1987). Early adopters’ expertise with the product arises from product usage or purchase experiences. As with opinion leaders, the implicit assumption about early adopters’ product conversations is that they talk about the product for product-related reasons, and as such early adopters are being viewed product class specific (Robertson & Myers 1969).
potential motives to pass on information are the desire to help others, pleasure in sharing information, and the obligation to share information (Walsh et al. 2004). Nevertheless, Goodey & East (2008) have found that the motivational differences between market mavens and non-mavens are not substantial. Furthermore, Phelps, Lewis, Mobilio, Perry & Raman (2004) have found that ‘viral mavens’ passed along emails out of desire to connect and share with others. Unfortunately Feick & Price (1987) did not find a clear demographic and socioeconomic profile of these information disseminators. Targeting the maven with communication may therefore be difficult.
WoM intentions and –behaviors are said to be consequences of other constructs. The next section describes the research that has been performed on the antecedents of positive WoM communication. In other words, what conditions precede the intention of WoM and actual WoM behavior?
2.2 Antecedents of positive WoM
As WoM behavior can have a significant influence on, for example, consumers’ product choices, the use of WoM as a promotional tool has more potential than previous estimates have suggested. Firms can adopt different roles in managing social interactions, as will be explained in the next paragraph. 2.3 WoM as a promotional tool
WoM campaigns are increasingly being used as it is believed by marketing scholars that at least some effects of WoM are partially within a firm’s control (e.g. Biyalogorsky et al. 2001; Godes et al. 2005; Keller & Lehmann 2006). In addition, as mentioned before, traditional media advertising is declining in effectiveness (Godes & Mayzlin 2004) and WoM is generally seen as more credible than communications initiated by marketers (Herr, Kardes & Kim 1991; Murray 1991; Gremler, Gwinner & Brown 2001; Allsop et al. 2007).
Most of earlier research has involved endogenously created WoM. That is, WoM that the consumer has decided to transmit based on some set of past experiences with the firm and/or its products and services (Godes & Mayzlin 2004). However, little research has been done on how marketers may design programs that proactively support, encourage and manage WoM behavior. Godes et al. (2005) and Godes & Mayzlin (2004, 2009) have addressed this issue in their papers, however. Using field data, Godes & Mayzlin (2009) have empirically demonstrated that a firm can drive up sales by creating WoM. Subjects who agreed to participate in their campaign were given a package of products and information, and specific suggestions for producing WoM for a particular firm. In their research, little economic reward was offered to the subjects, suggesting a certain level of intrinsic motivation of the subjects to engage in the activity. Godes & Mayzlin (2009) note that this intrinsic motivation might be the result of, for example, the aspiration to be “in the know” or, possibly, altruism.
Chew (2002) have tested the use of incentives as a motivator for WoM as well, and found that that incentives can effectively be employed to reinforce the likelihood of WoM. By conducting laboratory experiments, Ryu & Feick (2007) have investigated the impact of referral rewards on the likelihood to recommend a product. It was found that tie strength and brand strength are both important factors in explaining the impact of offering rewards on referral likelihood. Offering rewards was found to be particularly effective in enhancing the number of referrals for weaker brands and to weak ties. The explanation for this can be found in equity theory (e.g. Oliver & Swan 1989). Concerning weak ties, this theory holds that a person who recommends a product considers a referral as a favor done for the company and/or the receiver of the referral. Because no simultaneous increase in output comes with the increase in input, referrals yield inequity (Walster, Berscheid & Walster 1973; Walster, Walster & Berscheid 1978). This level of inequity can supposedly be reduced when the recommender receives a reward for recommending the product. In addition, Ryu & Feick (2007) stated that the occurrence on the recommender’s part of a concern about “being bought” is unlikely, unless the reward is very high. Besides tie strength, Ryu & Feick (2007) found that brand strength results in rewarded referrals. The existence of a reward has been found to increase the likelihood of recommending a product more for weaker brands than for stronger brands. The strong brand commitment of stronger brands’ consumers is caused by the fact that the choices of consumers of strong brands are more influenced by preferences than by financial boundaries (Blattberg & Wisniewski 1989). More confidence in making referrals results from this strong commitment, and, as a consequence, the incremental impact of rewards is limited. On the contrary, weak brands’ consumers are likely to be less motivated to recommend a product because of lower choice confidence (Heath, Ryu, Chatterjee, McCarthy, Mothersbaugh, Milberg & Gaeth 2000). Rewards have been found to compensate for this lower choice confidence, thereby increasing the referral likelihood of these consumers (Ryu & Feick 2007).
benefiting from low prices but not referring the product or service to others - could occur then. By contrast, the most favorable strategy when customers are easily delighted is not to offer rewards but to decrease the price. The goal in this case would be to attract more consumers by selling at lower prices. A lower price motivates consumers to buy the product and it increases referrals. Consequently, the range to motivate referrals with reduced prices is limited. In the situation of an in-between level of customer delight threshold a low price strategy should preferably be combined with a reward. To employ such a strategy, however, data is needed on the customers’ willingness to pay and the delight threshold (Biyalogorsky et al. 2001).
Little research has been done about how people react to the incentive to disseminate information (Godes et al. 2005). Verlegh, Verkerk, Tuk & Smidts (2004), however, have researched the credibility of WoM as a function of the sender’s incentives and the recipient’s evaluation of this credibility. Concluding that increased access to ulterior motives, such as, for example, the motive to make money, decreased the impact of WoM recommendations, these scholars found that a heightened awareness of a sender’s motives is likely to have a negative impact on the credibility of the recommending individual. The strength of the social ties between source and receiver of WoM has been found to moderate these effects (Verlegh et al. 2004). Although referral reward programs are popular tools for increasing WoM about products and services, marketers need to have a good understanding of the side effects and potential dangers of these incentive programs (Wirtz & Chew 2002).
dissemination, that is, the company decides on how and to whom information will be passed on. In the least passive strategy, the company participates directly in the conversations by creating WoM itself. Since strategically planned manipulations of online forums have been shown to either boost or reduce the information value of a forum to customers (Dellarocas 2006), an important issue for companies is how to participate in this kind of information dissemination such that is enhances the value of it.
In the next section different motives underlying WoM behavior will be explained. In other words, it will be clarified why people talk about their experiences with products and/or services. It is the belief of the author that these motives should guide companies in their management of interpersonal communication. The use of motives in guiding strategy formulation is widespread. For example, Westbrook & Black’s (1985) taxonomy of shoppers is based on people’s underlying shopping motivations. This taxonomy has been shown very rewarding in guiding retail strategy formulation. 2.4 Motives underlying traditional WoM communication
Wirtz & Chew (2002) have pointed out that WoM behavior only emerges when the satisfied consumer feels motivated to engage in it. Motivation has been defined as “the state of being stirred to action” (internet: Babylon October 3, 2009). In this respect, motives are said to direct consumer behavior (Walsh et al. 2004). Developing a deeper understanding of the motives that encourage consumers to act as an information disseminator in the marketplace can help marketers better harness interpersonal communication as a marketing tool.
As Godes et al. (2005) have written, a key difference exists between how information received from a firm and information received from a fellow customer is being interpreted. The latter usually gives unbiased advice, not driven by the desire to sell. This certainly does not mean, however, that the recipient of the information considers the motives of the sender to be unimportant or irrelevant. In fact, particular motivations spurring an individual to transmit information can definitely imply a certain degree of discounting on the part of the recipient, if the recipient in question is aware of these motivations. An example of such a motivation is a financial compensation being earned by the sender by recommending a specific brand to others.
As mentioned by Dichter (1966), consumers feel threatened when they believe that an advertisement is more of a sales tool than information and guidance. Furthermore, Verlegh et al. (2004) found that increased accessibility of ulterior motives decreases the impact of WoM recommendations, and has a negative impact on the perceived credibility of the recommending person. Incentive programs aimed at stimulating WoM carry with them side effects and potential dangers (Wirtz & Chew 2002). It could be interesting, therefore, to examine ways through which to further build on the intrinsic motivations consumers might have to transmit information. To encourage customers to engender referrals without facing the risk of negative side effects, such as a negative impact on the perceived credibility of the recommending person caused by, for example, referral rewards, marketers could attempt to build communication channels with interpersonal communicators based on the intrinsic motives customers have to talk about a product and/or service. Within this research it is tested whether such a WoM program that takes into account the basic motives driving WoM communication could significantly influence consumers’ WoM intention. A number of different potential motives underlying WoM have been identified in scholarly literature. A short outline of the scientific discussion on these motives will follow first.
Human behavior in general is regarded as the product of both internal states and external stimuli (Westbrook & Black 1985). McGuire (1974) has identified 16 basic human motives in literature that guide human behavior (see table 1). According to McGuire (1974), motives are passive or active in character, internal or external in goal orientation, self-preservation or self-growth in purpose, and affective or cognitive in representation. People’s fundamental motives for talking about their experiences are rooted in these psychological needs (Dichter 1966).
namely focus-related utility, consumption utility, and approval utility (Balasubramanian & Mahajan 2001). While focus-related utility refers to “the utility the consumer receives when adding value to the community through his or her contributions” (Balasubramanian & Mahajan 2001), consumption utility concerns the value that a consumer obtains through the “direct consumption of the contributions of other community constituents” (Balasubramanian & Mahajan 2001). The third utility type, approval utility, deals with a consumer’s satisfaction that results from other constituents’ consumption and approval of the constituent’s own contributions (Balasubramanian & Mahajan 2001). In addition to these three utility types, Hennig-Thurau et al. (2004) have extended the framework to include moderator-related utility and homeostase utility, two additional utilities that focus on the distinct characteristics, such as anonymity and directed to multiple individuals, of online platforms. Although it has been shown that negative WoM can have a powerful impact (Arndt 1967), the focus of this study is on positive WoM creation. Therefore, motivations for negative WoM communication are ignored in the present theoretical framework. The intrinsic motives underlying positive WoM communication that follow from previous research will be discussed in the next subparagraphs.
Table 1: A Structuring of 16 general paradigms of human motivation. Source: Mc Guire (1974) p. 172.
of the product or service. Because of its importance in influencing consumer behavioral responses, product involvement has received considerable attention from marketing scholars. It is defined as “an internal state that indicates the amount of arousal, interest, or drive evoked by a product class” (e.g. Dholakia 2001). Within the framework of Hennig-Thurau et al. (2004) this motive is being referred to as “expressing positive emotions” and it is associated with homeostase utility. This utility is based on the belief that people are striving for balance in their lives (e.g. Zajonc 1971), and therefore will attempt to restore any unbalanced state into an equilibrium (Heider 1946, 1958; Newcomb 1953). Balance can be restored through for example writing comments on web-based opinion-platforms (Hennig-Thurau et al. 2004) or expressing positive emotions by talking to other consumers (Sundaram et al. 1998). An example of product involvement can be found in the coffee category (internet: H. Stijger September 21, 2009). Since the introduction of Nespresso and Starbucks, coffee has become a high involvement product. More and more people are buying expensive coffee machines and talk about the best way to prepare coffee on parties. The product involvement motive can be considered active in character, internal in goal orientation, self-preservation in purpose and affective in representation.
Scholars (e.g. Bloch & Richins 1983) have made a distinction between enduring involvement and situational involvement, whereby enduring involvement is a constant concern for a product class that is self-determining, i.e. not dependent of specific purchase situations, and situational involvement is the raised level of concern arising from a specific situation, usually a purchase occasion. In case the product itself serves to satisfy the continuing, self-identity-related desires of consumers enduring involvement becomes the result. Situational involvement, on the other hand, occurs when a product’s importance is based on an aspiration on the consumer’s part for the accomplishment of certain goals, which may have been derived from the purchase of a product and making use of it. The two-component model of product involvement is widely accepted and it has been shown by Richins, Bloch & McQuarry (1992) that a simple additive model in which enduring involvement and situational involvement are combined, best represents the level of product involvement experienced by the consumer.
(Hennig-Thurau et al. 2004). People are hunting for ego enhancement by adding satisfying roles to existing self-concepts, which is related to what McGuire (1974) called identification theories. One way to signal a level of social status or connoisseurship that can become important to a consumer’s self concept is by being viewed by others on web-based opinion platforms, as mentioned by Hennig-Thurau et al. (2004). The self concept, or the individual’s image of himself, is considered an import construct in the consumer behavior literature (e.g. Grubb & Hupp 1968). Based on the theory of individual self enhancement (Rogers 1951), Grubb & Grathwohl (1968) stated that the consuming behavior of an individual would be directed toward the protection and improvement of an individual’s self concept. If the use of goods could communicate symbolic meaning, than behavior of individual consumers would be directed towards enhancing the self concept through the utilization of goods as symbols, as reasoned by Hogg & Michell (1996). That brands can serve as symbolic devices, allowing consumers to project their self-image, is also being recognized within brand management literature (Keller 2003). By extension, self-confirmation has been found to play a major role in motivating people to talk about their experiences with products (Dichter 1966). Without being aware of it, people use products to reassure themselves in front of others and as a manner to carry them carefully through their own insecurity as can be read in a consumer’s quotation from Dichter (1966) “I recommended it to my sister-in-law Ruby and to Rosie Gibson, my friend on the hill, … . And I feel rather good that something I recommended was so well liked because it makes me feel that my judgment is good, and I know that people respect my judgment.”
2.4.4 Helping the Company. Consumers share their positive experiences with others also because of a desire to help the company. Although this is in fact an act of altruism, Sundaram et al. (1998) created a separate category because the objective here was to help the firm instead of another consumer. Although Sundaram et al. (1998) created a separate category for this motive, helping the company is being viewed as another form of focus-related utility (Hennig-Thurau et al. 2004). Keller (2003) talks about brand relationships as the last step in his “branding ladder”. Within this model building a strong brand can be thought of in terms of a sequence of steps: (1) create a brand identity, (2) create brand meaning, (3) obtain customer responses to the brand identity and brand meaning, and (4) try to convert the responses to the brand in loyalty relationships. The final step, which can be referred to as brand resonance, is divided into four substeps (Keller 2003), namely (1) behavioral loyalty, which refers to the share of category volume attributed to the brand and repeat purchases, (2) attitudinal attachment, (3) sense of community, whereby consumers feel affiliated with other people associated with the brand, and (4) active engagement in which the consumer is willing to invest time and energy in the brand. In this last step, consumers become ambassadors or brand evangelists and help the company to communicate about the brand.
2.4.5 Message involvement. The last motivation referred to by Dichter (1966) is message involvement, which encompasses the dissemination of information mainly stimulated by intriguing advertisements, commercials or selling appeals. Consumers motivated by message involvement do not necessarily have experience themselves with the product or service. Although consumers are often skeptical toward sales messages, it is also found to be accompanied by other consumer thoughts, which give rise to WoM, making them “sell” indirectly (Dichter 1966). An example of such an attitude is verbal play, consumers like to apply slogans and verbally ad lines. Furthermore it is found that the amusement value of ads and commercials have become topics of talk themselves (Dichter 1966).
of the positive effects that WoM can have as a promotional tool. In other words, is it possible for companies to increase WoM communication by capitalizing on the intrinsic motives consumers have by making them more salient to them? In the next session, a conceptual model is introduced in which the potential role of the intrinsic motives is being visualized.
2.5 Triggering events
Why wait for natural contagion to occur if you can accelerate buzz by ‘teasing’ the consumer? It is believed by the author that a small feature can act as the stimulant for inducing satisfied and motivated people to share their experiences with other persons. Instead of waiting until Customer A tells Customer B, who will tell Customer C etcetera, marketers need to trigger their satisfied and motivated customers, thereby enhancing the chance of WoM occurring. The conceptual model below shows the WoM process as suggested in this research.
Figure 1: Conceptual Model
As discussed in the literature, several antecedents precede WoM before occurring. For example, consumers need to be satisfied with the product or service and they must feel committed to the product or service in question. Wirtz & Chew (2002) have cited that satisfaction with a product or service alone is not enough for WoM to occur. More precisely, consumers need to feel motivated to pass on information to their peers. The conceptual model suggests that these motives underlying consumers’ WoM behavior can be used by marketers to stimulate satisfied and motivated customers to share their experiences with others. An example from practice is the use of incentives to trigger a consumer to talk about a product. Incentives then function as an extrinsic motivator. As is believed by the author, other triggers than incentives exist, which correspond to other underlying motives of WoM behavior. The aim here is to find out whether making the different motives more salient, can effectively trigger consumers to share their positive experiences with a product.
appeal to the basic motives driving WoM behavior can significantly influence positive WoM communication.
Hypothesis 1: Messages sent by a company to customers based on underlying motives to engage in positive WoM behavior, significantly enhance the WoM communication of these customers.
Within this research both WoM intention and WoM behavior are supposed to be part of the total WoM communication concept1.
Furthermore, it is examined which message (motive) triggers consumers the most in their intention to recommend and their actual WoM behavior. It is decided to create messages for the product involvement-, helping the company-, and altruism motive, because it is believed that making these more common motives salient for the customer is more easier for a company than adding into motives that are predominantly individualistic in nature (i.e. the self involvement motive). In addition, message involvement is not experimented with because consumers motivated by this motive do not necessarily have experiences themselves with the product or service. A connection with other constructs, such as satisfaction and commitment to the product or service, which are considered to precede positive WoM communication (Mittal, Kumar & Tsiros 1999; Swan & Oliver 1989), is therefore in some cases not possible. Furthermore, it is difficult to elicit behavior based on a particular execution of the message instead of the information in the message, especially in the product category that is used in this experiment (the product category - skin care within the fast moving consumer goods market - used in the experiment will be explained in the research design session). While the self concept is considered an important construct in consumer behavior (e.g. Grubb & Hupp 1968), the self involvement motive is not accounted for in the experiment because of the difficulty of creating a message that triggers this particular (more individualistic) motive.
Within the research of Sundaram et al. (1998) the product involvement motive is mentioned by 33% of the respondents as a motive for engaging in positive WoM. Altruism and helping the company follow with 28,7% and 17,9% respectively. Although product involvement is cited most often as the reason for recommending a product or service, a conclusion about the effectiveness of the motives cannot be reached based on this information. More customers seem to be motivated by product involvement than by helping the company or altruism, but this does not provide any evidence that
1
messages based on product involvement will be more effective in enhancing WoM communication. Hennig-Thurau et al. (2004) have analyzed the ability of these and of other motives that are considered unique to eWoM, to predict eWoM communication though. The number of comments written by the consumer on opinion platforms and their frequency of visiting the platform have been used to measure eWoM behavior. It has appeared that helping the company did not have an impact on the two dependent variables. Although no positive effect was found for the helping the company motive via recommending a firm’s products and/or services using opinion platforms, it could be that this motive will be of importance when articulated differently, i.e. not by using virtual communities. Positive impact was, however, found for extraversion (i.e. product involvement) and concern for other consumers (i.e. altruism) on both the number of comments written by the consumer on opinion platforms and their frequency of visiting these platforms.
As mentioned by Dichter (1966), people’s motives for talking about their experiences are rooted in several psychological needs. The product involvement motive can be considered active in character, internal in goal orientation, self-preservation in purpose and affective in representation. On the other hand, the helping the company- and altruism motive are predominantly external in goal orientation and self-growth in purpose. It is believed by the author that an individual’s need to maintain an equilibrium will be of less importance in case of low involvement products and/or services. Therefore, in case of low involvement products and/or services, it is thought that making the helping the company- and/or altruism motive salient will be more effective than triggering the product involvement motive. The product category used within this experimental research (skin care in the fast moving consumer goods market) can be considered a low involvement category because of the high observed switching behavior and low loyalty. For that reason, compared to making salient the helping the company- and/or the altruism motive, triggering the consumer using the product involvement motive is expected to be less effective in enhancing the WoM intention and/or behavior of this consumer.
Hypothesis 2: Messages sent by a company to customers based on the helping the company- and/or the altruism motive are more effective in enhancing the WoM communication of these customers compared to messages founded on the product involvement motive.
In the next chapter the methodology used to examine the effectiveness of these messages is explained.
3. Research design
For this research data has been used from Henkel Netherlands. With more than 55.000 employees worldwide, Henkel’s ultimate purpose is to develop their competitiveness and to permanently establish a winning culture. The business is organized into three globally operating sectors, namely Adhesive Technologies, Laundry & Home Care, and Cosmetics & Toiletries.
Henkel’s Diadermine, a brand within the Cosmetics & Toiletries division, was founded in France in 1904. In 1981, Diadermine was introduced in several European markets, including in 2000 the Netherlands. Internationally, Diadermine emerged as a major player in the face care market. In the Netherlands, however, Diadermine has a low level of consumer brand awareness and a market share of 2,9% (YTD09 wk28). The Dutch fast moving consumer goods market of skin care is being characterized as highly competitive in nature. Customers are not loyal to brands as expressed by the high switching behavior within this category.
It is believed by Henkel’s management that Word-of-Mouth could be used in a cost efficient (bottom up) way to stimulate trial, create more (top-of-mind) awareness, and ultimately improve the market share of Diadermine. Therefore, a brand ambassadors program has been initialized as a means to facilitate, influence, and encourage WoM about Diadermine among women in the Netherlands. As working with brand ambassadors is a new marketing tool for Henkel, the program is used to experiment with this type of marketing. Furthermore, a new website has been launched where information about the products and the brand can be found. Moreover, consumers could register themselves with the aim of receiving information about the newest introductions and the possibility to send information about Diadermine to their friends and/or acquaintances.
3.1 Target Population
care industry (see appendix 1, figure 1). Twelve women that were likely to share their opinions were selected for the brand ambassadors program after personal interviews. A Dutch beautician working as a freelancer for Diadermine was hired to give workshops on skin care and Diadermine as a brand to the selected brand ambassadors. The workshops were a means to provide the brand ambassadors with some information worth talking about and tools that make it easier for them to share their opinions. The brand ambassadors are expected to bring Diadermine under the attention of friends, family, colleagues and acquaintances through, for instance, sharing their experiences and sampling. The women that are offered WoM are asked to fill out a short questionnaire about their experience with the product (see appendix 1, figure 2). The women also had to mention the name of the brand ambassador they have met. At the end of July 2009 885 questionnaires are being collected and the information is stored in a database managed by an outside company (Proximity). The email addresses in this database are being used for the experiment, which will be explained in more detail later.
In addition to the database from the brand ambassadors program, 804 email addresses from customers who have enrolled themselves at the website, have been used for the field experiment. 3.2 Methodology
3.2.1 Stimuli. Online pilot tests were conducted to confirm the found effects could safely be attributed to the manipulation. Forty-six respondents, mainly students and women between the ages of 30-55, participated in a first pilot test.
Motivations underlying positive WoM behavior were manipulated in three different email messages. The motivations used are product involvement, helping the company, and altruism. Furthermore, a control group is added to the design. Some sentences differ across the mailings, in this way manipulating the different motives in the mailings. The product involvement mailing triggers a person’s interest in the product and tries to bring forth excitement and positive feelings about Diadermine. The content of the mailing appeals to the fact that experience with the product creates a tension which is not relieved by the use of the product alone, but must be channeled by way of talk, enthusiasm, and recommendation to re-establish the equilibrium. The sentences that were used exclusively in the product involvement mailing are:
“Are you also that satisfied with Diadermine? And can't you get enough of talking about it? Do you want your friends to try the new innovation from Diadermine?”
“Word-of-mouth advertising is very important to convince other women of Diadermine. Would you like to help Diadermine to become better-known in the Netherlands?”
The content of the altruism mailing, on the other hand, plays into the intention of aiding others to make a satisfying purchase decision. But in contrast to the helping the company mailing, this mailing is aimed at helping other consumers. It is based on the idea that people who have been enthusiastic about the product, want others to have it too:
“Do you have friends who you know have to get acquainted with Diadermine?”
The final mailings that were used within the first pilot test can be found in appendix 2 (in Dutch). Participants were randomly assigned to the mailings and were asked to read the email messages thoroughly. After reading them, the email messages were removed and the respondents received a questionnaire. The questionnaire included nine items that correspond to three underlying dimensions to measure the presence of the different motives in the email messages: product involvement, helping the company, and altruism (see appendix 2, figure 5 for the questionnaire and table 1 for the constructs).
The questions used a five-point Likert scale with end points labeled as ‘totally disagree’ (1) to ‘totally agree’ (5). Some items are reverse coded and some may be duplicated for the purpose of assessing reliability. Internal consistency was measured using Cronbach’s alpha. Furthermore, Analyses of Variance (ANOVA), Kruskal Wallis tests and Repeated-Measures ANOVA with one within-subjects factor were run to assess whether differences exist between and within the mailings.
3.2.2 Results pilot test. Cronbach’s alpha indicates acceptable reliability for the altruism construct only (α = 0,693). The low alpha for the remaining two constructs (0,531 vs. 0,563) causes that the questions underlying the constructs instead of the constructs will be used for further analyses (see appendix 2, figure 5 for the questionnaire). A high mean score on a construct or question for a specific mailing (manipulation) means that respondents on average agree with the statements referring to the underlying motive.
When comparing the different mailings, it can be seen that respondents who have received the helping the company mailing, rated helping the company as a more plausible motive (M4 = 4,6; M7 =
4,8; M9 = 4,0)2 than respondents who received the product involvement mailing (M4 = 3,7; M7 = 3,2;
M9 = 3,1), altruism mailing (M4 = 2,9; M7 = 4,1; M9 = 3,2) or control mailing (M4 = 2,9; M7 = 3,1; M9 =
2
Mean scores on a specific question rounded at one decimal, i.e. M4 refers to the mean score on question 4.
3,3). On the contrary, the product involvement- and the altruism mailing do not seem to reflect the correct motivation for WoM, because respondents who have received these mailings, do not rate the corresponding motive as the most likely one. Albeit the content of the product involvement- and the altruism mailing does not match with the underlying motivation it is trying to refer to, all mailings were assessed to establish whether differences between the mailings do exist.
ANOVA is used for the between-mailings analysis of the question “This mailing responds to my involvement with the Diadermine brand” (product involvement motive) and the question “Above all, in this mailing it is assumed that I want to help Diadermine becoming a more familiar brand” (helping the company motive). No significant results were found (p > 0,05), which means that respondents do not rate the mailings differently based on these two questions referring to the product involvement- and helping the company motive (see appendix 2, table 2 for the results of the homogeneity of variances tests).
The altruism construct as a whole and the remaining questions from the product involvement and helping the company constructs were analyzed using Kruskal Wallis tests3 (see appendix 2, table 2 for the results of the homogeneity of variances tests). Based on two questions referring to the helping the company motive, respondents who have judged the helping the company mailing, rated the helping company motive as a more plausible motive than respondents who have evaluated one of the other mailings. The mean ranks on the two questions referring to the helping the company motive (a) “In this mailing it is said that word-of-mouth is important to convince other women about the Diadermine brand” (Chi-Square = 10,222; p = 0,017) and (b) “In this mailing I’m not being called to help Diadermine becoming a more familiar brand” (Chi-square = 11,723; p = 0,008) appear significant. Based on these two results, the helping the company motive seems manifest in the helping the company mailing. No significant results were being found regarding the altruism construct and the remaining questions underlying the product involvement construct (p > 0,05), which means that respondents who have evaluated the altruism or product involvement mailing do not rate one motive as more plausible than another.
It can be concluded that the helping the company mailing reflects the helping the company motive well. However, the product involvement- and altruism mailings do not meet the standards for a
3
successful experiment. The disappointing results can be due to the manipulations, the questionnaire, or the respondents. After some discussions with fellow students, who can be considered experts in this area, it is decided that the product involvement- and altruism manipulation had to be changed. In appendix 3 the adjusted mailings can be found. More emphasis is placed on the motives compared to the mailings use in pilot test 1. Because of the significant results for the helping the company mailing, no changes have been made to this manipulation.
3.2.4 Pilot test 2. A second pilot test has been run to check whether the manipulations have improved after some changes have been made. More emphasis is being placed on the product involvement- and altruism motives:
Product Involvement: “3 million women have already been convinced! Are you also that satisfied
with Diadermine? … Is this a product you like using and you can't get enough of talking about it? All in all, are you convinced?”
Altruism: “3 million women have already been convinced! Would you also enjoy helping friends to get
acquainted with a great product?… Do you have friends who you know have to get acquainted with Diadermine? Would you like to help them get this information?”
A group of 102 people participated and was assigned randomly to the mailings. A new group of participants was selected for the second pilot test because prior knowledge about the mailings is not desirable. It is believed that prior knowledge could actually disturb the results. Furthermore, It should be noted that two extreme cases were deleted because of excessive answers (see appendix 3, figure 3 and 4).
Because Cronbach’s alpha indicates low, but acceptable reliability for the product involvement construct (α = 0,610) and the helping the company construct (α = 0,668), these constructs are used for further analysis. The low alpha for the remaining altruism construct (0,576) causes that the three questions underlying this construct instead of the construct itself will be used for additional analyses (see appendix 2, figure 5 for the questionnaire and appendix 3, table 1 for the constructs).
constructs and questions within the mailings are compared, it can be seen that respondents who have received the altruism mailing, rated altruism as a more plausible motive (M5 = 4,00; M6 = 4,31;
M10 = 4,13) than the product involvement motive (M = 3,43) or the helping the company motive (M =
3,50) (see table 2).
Mean evaluations and standard deviations pilot test 2
Mailing Product involvement (n=16) Helping the company (n=21) Altruism (n=39) Control (n=24) Product involvement construct
Helping the company construct Altruism question 5 Altruism question 6 Altruism question 10 3,94 (0,77)* 3,92 (0,68) 3,43 (0,96) 3,36 (0,83) 3,38 (0,80) 4,27 (0,80)* 3,50 (1,02) 3,08 (1,03) 3,88 (1,31) 4,10 (1,30) 4,00 (1,08) 4,04 (1,08) 4,19 (1,33) 3,71 (1,49) 4,31 (1,06) 4,25 (1,15) 4,44 (0,81) 4,00 (1,00) 4,13 (0,90) 3,83 (1,01)
Table 2: Mean evaluations and standard deviations pilot test 2
*Significant at the 0,05 level (between-mailings analyses)
ANOVA is used in all cases to test whether differences exist between the mailings or manipulations (see appendix 3, table 2 for the results of the homogeneity of variances test). After reading the product involvement mailing, respondents evaluated the product involvement motive as significantly more in line with the mailing, than respondents who received another mailing (F(3, 96) = 2,996, p = 0,035). Significant results are found for the helping the company mailing as well (F(3, 96) = 6,179, p = 0,001); respondents assessed the helping the company motive as significantly more in line with the mailing, than respondents who received another mailing. The three questions underlying the altruism construct, in contrast, do not show significant results on the ANOVA test (p > 0,05).
Tukey’s Honestly Significant Difference (HSD) test, known for its versatility (Bryman & Cramer 2000), is applied to see which mailings are significantly different from one another based on the product involvement and helping the company construct. The mean score of the helping the company construct is highest for the helping the company mailing (M = 4,27) and is found significant higher than the mean scores on this construct for the other three mailings which did not differ significantly from each other (product involvement M = 3,38, p = 0,028; altruism M = 3,50, p = 0,019; control M = 3,08, p = 0,000)
In short, based on the ANOVA tests and post hoc analyses, the product involvement mailing, albeit not significant in the post hoc analyses, and helping the company mailing seem to reflect the motives for WoM communication that were manipulated within the mailings well.
alpha) differently. In situations in which three or more means from the same subjects are compared, a multivariate analysis of variance with one repeated-measures variable is often being used (Bryman & Cramer 2000). The repeated-measures variable in these cases is a factor variable with n levels, since it is repeated n times. Although there is no single variable with n levels in this within-mailings analysis, construct could be considered as a factor variable with three levels, namely the product involvement construct, the helping the company construct, and the altruism construct. Each participant provides a score on every construct, in other words, construct is considered to be repeated three times. The null hypothesis that is tested within each mailing is that the mean scores on the three constructs are equal. In appendix 3, table 3 the mean scores on the constructs can be found. However, in this case, the interpretation of the analysis is only meaningful for the mailings for which the corresponding construct score is higher than the other two scores. Therefore, it is checked whether the construct scores are significantly different from each other in the helping the company- and altruism manipulation only.
Mean scores Mauchly’s test of sphericity
Mailing Construct Mean Mauchly’s W df Sig.
Helping the Company
Product Involvement Helping the Company Altruism 3,92 4,27* 3,94 0,970 2 0,751 Altruism Product Involvement Helping the Company Altruism
3,43 3,50 4,15*
0,882 2 0,097
Table 3: Mean scores and Mauchly’s test of sphericity
*Significant at the 0,05 level
The null hypotheses are being analyzed using Repeated-Measures ANOVA. A basic assumption of a Repeated-Measures ANOVA is sphericity, which is tested using Mauchly’s sphericity test. In table 3 the results of this test are shown. Sphericity cannot be a assumed if the significance level is p < 0,05, which does not hold true for the helping the company and altruism mailing.
The main effect of construct is being found significant within each mailing (helping the company F(2, 40) = 3,220, p = 0,051 and altruism F(2, 76) = 11,256, p = 0,000). In other words, the null hypotheses that the mean scores on the helping the company- and altruism construct are equal within a specific mailing cannot be supported. After reading the helping the company mailing, respondents evaluated the helping the company motive as significantly more in line with the mailing the other motives. In addition, respondents who have evaluated the altruism mailing, rated the altruism motive as a more plausible motives than the other motives.
When the ANOVA results and the Repeated-Measures ANOVA results are taken together, it can be concluded that the three manipulations can be used within the experiment.
3.2.5 Respondents and design of the field experiment. At the end of July 2009 780 email addresses were available for the experiment (88,1% of the email addresses that have been collected with the brand ambassadors program). Email addresses were selected for the experiment based on the opt-in possibility respondents had (“Yes, I want to stay in contact with Diadermine”) and the unique availability of the email addresses (double entries were removed). In addition, 804 email addresses were available from customers who have subscribed themselves at the Diadermine website.
A field experiment is used with random assignment of the respondents to the experimental groups. The experiment contains four conditions, namely (1) the product involvement motive, (2) the helping the company motive, (3) the altruism motive, and (4) a control group. The design and copywriting of the mailings has been updated for the experiment because of the introduction of a new product and a decrease in budget available for the experiment (see appendix 4 for the final mailing). It is expected that this design change has no effect on the results of the experiment because the experimental message remained the same.
questionnaire and the reverse coded items from the pilot test were adjusted in normal coded items. These corrections have been made because some respondents from the pilot test had difficulty understanding the questions. In addition, it is believed that a seven-point Likert scale is better able to yield significant results when available. An incentive is used to enhance the response for the questionnaire. Four Diadermine packages worth €55 each are being allotted among the respondents. Analyses of Variance (ANOVA) and Kruskal Wallis tests were run to assess whether differences exist between the mailings. Furthermore, ANOVA and Independent Samples T-tests were run to assess whether significant differences in consumers’ intention to WoM exist between the manipulated mailings.
4. Results Field Experiment
In this section the results of the hypotheses testing will be discussed. In total 1584 email messages have been sent to female Diadermine users in the age of 18 years old and older. The first mailing has been sent to 780 respondents in the brand ambassadors database, and a second mailing has been sent to 804 respondents who subscribed themselves at the website. Due to opt-out requests and bounced mailings (i.e. out-of-office reports) 267 mailings could not be delivered. The number of bounced mailings after the first mailing (196) is much higher compared with the second mailing, when only 45 mailings bounced. It is expected that this is partially due to the holiday period. During holidays people use their out-of-office option more often and it is expected that more email boxes reach their maximum capacity during this period owing to inactivity. From the 1316 mailings that have been received 33,6% was actually opened. The open rate of the first mailing, 25,5%, has turned out to be lower than the open rate of the second mailing (39,9%). As mentioned earlier, the holiday period is expected to explain for this difference to some extent.
Seventy-four respondents have completed the survey, which is 16,7% of the opened mailings. Due to technical reasons no distinction has been made between mailing 1 and mailing 2 with respect to the survey results. In table 5 mean evaluations of the respondents on the product involvement-, helping the company- and altruism construct are shown.
4.1 Manipulation check
the product involvement construct is being found at the product involvement mailing etcetera). Based on between mailings analysis using ANOVA (product involvement and helping the company) and a Kruskal Wallis test (altruism) it can be concluded that the mailings significantly differ from each other based on the three constructs that have been measured (product involvement F(3, 70) = 2,508, p = 0,066; helping the company F(3, 70) = 8,129, p = 0,000 and altruism Chi-square 20,163, p = 0,000) (see appendix 4, table 3 for the results of the homogeneity of variances test and table 4 for the mean rank of the altruism construct).
Mean evaluations and standard deviations experiment Mailing Product Involvement (n=14) Helping the company (n=22) Altruism (n=14) Control (n=24) Construct Product Involvement 5,43* (0,86) 5,04 (1,24) 4,62 (0,95) 4,54 (1,05) Helping the company 4,52 (1,35) 5,85** (0,97) 4,69 (0,77) 4,56 (0,94) Altruism 4,52 (1,23) 5,65 (1,03) 5,90** (0,53) 4,58 (1,09)
Table 5: Mean evaluations, standard deviations, and results of the between-mailings analysis of the experiment
* Significant at the 0,10 level ** Significant at the 0,05 level
4.2 WoM Intention
The WoM intention within every mailing can be found in table 6. In addition, a distinction has been made between the WoM intention to strong ties and the intention to WoM to weak ties.
WoM intention per motive (mailing) Mailings Product Involvement (n=14) Helping the Company (n=22) Altruism (n=14) Control (n=24) Mean (n=74) WoM Intention (friends
and acquaintances)
4,29 (1,18) 4,43 (1,57) 4,55 (1,13) 4,40 (1,32) 4,41 (1,32)
WoM Intention to friends 4,60 (1,20) 4,61 (1,58) 4,88 (1,26) 4,72 (1,32) 4,69 (1,35)
WoM Intention to acquaintances
3,98 (1,32) 4,26 (1,61) 4,21 (1,24) 4,07 (1,45) 4,14 (1,42)
Table 6: WoM intention per motive (mailing)
based on the helping the company- and/or the altruism motive are more effective in enhancing the WoM intention of customers compared to messages founded on the product involvement motive. Furthermore, no significant effects are being found for the WoM intention to friends and the WoM intention to acquaintances when analyzed separately (p > 0,05) (see appendix 4, table 5 for the results of the homogeneity of variances test).
Nevertheless, the intention to communicate about the brand to friends (M = 4,70; std. deviation = 1,35) is found to be significantly higher than the intention to communicate to acquaintances (M = 4,14; std. deviation = 1,42) (t = 2,458, p = 0,015). This result confirms earlier research about tie strength, i.e. compared to individuals in weak tie relationships, those in strong tie relationships tend to exchange more information and to interact more frequently (Brown & Reingen 1987; Reingen & Kernan 1986; Granovetter 1973). This result, however, does not hold when the intention to communicate about the brand to friends is being compared per mailing to the intention to communicate to acquaintances (p > 0,05).
4.3 WoM Behavior
As described in the research design, WoM behavior is being measured using two separate, but almost equal, mailings. WoM behavior per mailing is being shown in table 7.
WoM behavior Motives
Mailing 1 Product
Involvement
Helping the Company
Altruism Control Total
Number of emails opened 31 46 40 30 147 Click send2friend 1 (3,2%) 3 (6,5%) 3 (7,5%) 2 (6,7%) 9 (6,1%) Send2friend 0 3 3 1 7 Number of friends 0 1/3/15 (total 19) 1/3/12 (total 16) 3 38 Mailing 2 Number of emails opened 78 78 48 91 295 Click send2friend 2 (2,6%) 3 (3,8%) 1 (2,1%) 0 (0%) 6 (2,0%) Send2friend 1 1 1 0 3 Number of friends 1 1 3 0 5
Table 7: WoM behavior
respondents, one from the helping the company mailing and the other from the altruism mailing, have used the send2friend URL to forward information to 15 versus 12 people respectively. Within the second mailing six respondents have clicked on the send2friend URL, which corresponds to 2% of the respondents who have opened the mailing. Only three out of six respondents who have clicked on the send2friend URL have actually used it to forward information to five other people in total. The first mailing has resulted in a higher number of respondents who forwarded information to friends and/or acquaintances. Furthermore, on average more people have been reached by forwarding information with respect to the first mailing than with the second mailing. As an overall result, it appears that the helping the company- and altruism mailing were most effective in enhancing WoM behavior. However, due to the low response no statement can be made with respect to the hypotheses that were drawn up. Furthermore, ANOVA does not yield any significant results with respect to WoM intention. Therefore, a significant result regarding actual WoM behavior would not be very likely.
5. Discussion and conclusion
more forwarding behavior than the second mailing that was sent to respondents who subscribed themselves at the website. An explanation for this could be that customers who have enrolled themselves at the website did this out of a desire to enjoy promotions and receive information about the newest product introductions, and not with the intention to forward information to friends. No analyses could be done to check whether the forwarding behavior displayed by the respondents who received the first mailing was significantly higher compared to the forwarding behavior of the respondents who have received the second mailing.
In this research, forwarding behavior is being used as a metric to measure WoM behavior. The use of a send2friend URL could be one of the reasons for the low forwarding behavior. Respondents could be reluctant to use a send2friend URL to forward information to friends and/or acquaintances for several reasons, i.e. consumers already experience a high level of noise in their day-to-day electronic communication and spam (i.e. unsolicited bulk mailings) is increasingly seen as a problem (De Bruyn & Lilien 2008). Furthermore, forwarding behavior can be considered as too restricted to measure actual WoM behavior, because WoM behavior encompasses more than only forwarding behavior. Indeed, in this research WoM is being defined as “any information about a product or service transferred from one person to another either in person or by the use of some communication tool, whereby the receiver perceives the communicator as non-commercial regarding the information being transferred.” More specifically, respondents talking to a friend, family member or acquaintance about the brand, is not being measured using the send2friend URL, although this is a very easy way for customers to engender referrals. To forward information using the send2friend URL, respondents need the receiver’s email-address, which could be a threshold for actually forwarding the information. This threshold could also explain why more respondents clicked on the send2friend URL than that they actually used it to forward information. A solution could be to use another medium, or a combination of communication tools to encourage customers to engender referrals.
5.1 Scientific Implications and recommendations for future research
be realized through weak ties because it is supposed to reach more people than if the same information is being transmitted through strong ties. This is due to the fact that within social networks people are already likely to have similar or related information. Weak ties, in contrast, often function as the connection between otherwise secluded strong tie networks (Granovetter 1973). Strong ties are considered more important when the main purpose is to persuade consumers instead of enhancing the awareness of a product (Cialdini & Sagarin 2005).
The results of this study contribute to our knowledge about WoM motives. In particular, it is addressed in this research whether a company could activate customer referrals by designing WoM campaigns that appeal to the basic motives underlying WoM communication. The fact that no results have been found with respect to WoM intention and -behavior, challenges current thinking about the design of WoM programs that can proactively support and encourage WoM behavior. This research was based on the idea that being a satisfied or even delighted customer with positive attitudes toward a product or service does not by itself cause one to provide positive WoM, but that customers must feel motivated to engage in WoM before it actually emerges (Wirtz & Chew 2002). Although earlier research has confirmed that customers may talk about products and/or services for a variety of reasons (Dichter 1966, Sundaram et al. 1998), this research suggests that triggering these motives by making them more salient to the customer, does not enhance the customer’s WoM intention and WoM behavior. As a result, the question arises whether these motives that energize or drive a customer’s referral behavior play a significant role in determining WoM at the moment that a customer actually recommends a product or service, or that it is probably more likely that these motives have been constructed in previous research after recommendations have been made, and that the motives are thus not a dominant factor in driving actual behavior? If true, could companies then encourage WoM behavior by triggering the antecedents of positive WoM, such as satisfaction and commitment?
