Master Thesis Marketing Management
Benefits or Facts only?
The Moderating Effect of the Type of Message on
the effectiveness of Firm-Controlled Social Media
Posts about Hedonic or Utilitarian Products.
Master Thesis Marketing Management
Benefits or Facts only?
The Moderating Effect of the Type of Message on
the effectiveness of Firm-Controlled Social Media
Posts about Hedonic or Utilitarian Products.
Maja Sogorski
January 2019
University of Groningen
Faculty of Economics and Business
Department of Marketing
PO Box 800, 9700 AV Groningen (Netherlands)
Moselstraße 12
28199 Bremen (Germany)
Mobile: +49 (0) 15786595685
E-mail: m.sogorski@student.rug.nl
Student number: s3391027
Supervisors:
University of Groningen
First Supervisor: Dr. Hans Berger
I
Management Summary
For the past decades, the Internet has taken a central position in our daily lives. Marketing research has therefore increasingly concentrated on answering the questions, why we buy, what we buy and where we buy. And not only research is interested in answering these questions, but marketing managers are too, as online marketing activities have become an important part of complete marketing campaigns (Garbarino & Strahilevitz, 2004; Lamberton & Stephen, 2016).
One of the most recent developments in online marketing is social media marketing. Social media channels like Facebook, Instagram or Twitter have become a convenient way of communication for customers and companies (Lamberton & Stephen, 2016; Kumar et al., 2015; Shankar et al., 2016). But as social media platforms’ main aim is to enable people to communicate and interact with friends, most of the content is random and can’t be controlled by firms (Shankar et al., 2016; Kumar et al., 2015). Being able to control a part of the content in order to activate and sell products and services has become increasingly important for marketing managers (Wedel, 2011). Two common ways to control for content are creating firm-generated content or pay for sponsored recommendation posts (Kumar et al., 2015; Lu Chang & Chang, 2014).
Prior research has investigated the main effects of either one of the social media platforms on consumers’ decision process or only focused on one type of product (e.g. Lu, Chang & Chang, 2014; Kumar et al. 2015). Thus differentiating between different social media platforms, products or types of messages and comparing them were pointed out as important topics for further research (Lu Chang & Chang, 2014; Kumar et al., 2015; Maheswaran & Sternthal, 1990).
This thesis will extent prior research and answer the question if firm-generated content and sponsored recommendation posts are more effective for hedonic or utilitarian products and if the effectiveness differs if the posts are written in a more informative or more persuasive way (Overby & Lee, 2006; Maheswaran & Sternthal, 1990).
II Based on the hypotheses, a quantitative study based on an online questionnaire (n=165) was conducted to examine the effects of different firm-controlled social media posts for different product types, paired with different types of messages on online purchase intention. Therefore, two common social media platforms, namely Facebook and Instagram are used as representatives, as Facebook is often used to publish firm-created content, and Instagram is often used to publish sponsored recommendation posts (Kumar et al. 2015; Woods, 2016). Prior research suggested to differentiate between hedonic and utilitarian products, which is why these studies will concentrate on the mentioned types (Kumar et al., 2015). Besides the main effects between different posts for different product types and online purchase intention, prior research has shown, that different types of messages could cause a moderating effect on this relationship (Overby & Lee, 2006; Maheswaran & Sternthal, 1990). Consequently, the moderating effect of persuasive messaging, which is based on benefits, and informative messaging, which is based on key facts, is measured, too.
The results show, that firm-generated content at Facebook is more successful for utilitarian products than for hedonic products and this relationship is moderated by the type of message. For utilitarian products, persuasive content was more successful, whereas, for hedonic products, informative messaging was more successful.
Whereas, within the context of sponsored recommendation posts at Instagram, no significant difference in the purchase intention between the different types of products or messages was found.
However, comparing the results of the two studies show, that there are significant differences between the two social media platforms. Participants in the Instagram setting show a significantly higher purchase intention than participants in the Facebook setting. Compared to the Facebook setting, the purchase intention in the setting of the sponsored recommendation post (Instagram) was significantly higher for hedonic products paired with persuasive messaging.
III
Preface
Social media marketing has become a very important field of marketing management. Every company, I worked for in the past years had to make important decisions in this field as social media marketing offered great advantages due to relatively low costs and its wide scope. But not every social media campaign was as successful as expected.
Many researchers have studied consumer behavior in the online environment in the past years but there are still many relationships and effect variables that have to be further examined. Within two studies, I examine the impact of two types of firm-controlled product posts for either hedonic or utilitarian products on consumers’ online purchase intention. As different types of firm-controlled posts, firm created content on Facebook and sponsored recommendation posts on Instagram are used. Additionally, I investigate the influence of different types of messages on this main relationship. Hereby, the types of messages will be differentiated between fact-based (informative) and benefit-based (persuasive) messages. The outcomes offer valuable insights both for prospective research and for marketers. Prospective research will get insights into which variables and statistical relationships to examine further, and marketers will get guidance about where to invest, having a certain type of product, and which type of messaging technique to use. This knowledge will further contribute to the success of firms’ social media campaigns.
I would like to express my gratitude to my friends and family, my line manager at work, as well as to my supervisor Dr. Hans Berger who offered me great support, especially in the past semester.
I hope you enjoy reading my thesis! Maja Sogorski
IV Table ofContents Management Summary ... I Preface ... III 1. Introduction ... 1 2. Theoretical framework ... 3 2.1. Firm-generated Content ... 3
2.2. Sponsored Recommendation Posts ... 4
2.3. Influence of Type of Message ... 5
3. Methodology ... 8
3.1. Research Method ... 9
3.2. Data Collection ... 9
3.2.1. Participants ... 9
3.2.2. Procedure and Measurements ... 10
3.3. Plan of Analysis ... 13
4. Results ... 14
4.1. Study I ... 14
4.2. Study II ... 18
4.3. Comparing the two studies ... 20
5. Conclusions & Recommendations ... 22
5.1. Conclusion ... 22
5.2. Recommendations ... 24
5.2.1. Managerial Implications ... 24
5.2.2. Limitations and implications for further Research ... 25
References ... 26
Appendices ... 31
Appendix I: Description of stimulus and its sample size ... 31
Appendix II: Measures for Constructs ... 31
Appendix III: Profile Plot study I ... 33
Appendix IV: Profile Plot study II ... 34
1
1. Introduction
Mobile influences get more and more important in our lives every day. It affects the way we communicate, decide and purchase. Shopping motivations and goals have become more dynamic due to new online reference groups and new shopping and browsing possibilities, the digital world offers. And not only consumers benefit from the new digital environment, but online marketing also offers retailers and manufacturers cost efficient possibilities to influence consumers along the consumer decision process (Lamberton & Stephen, 2016). Thus it is not surprising, that there has been a strong trend towards implementing new digital marketing activities, like relationship-, one-to-one- and Social Media Marketing (Wedel, 2011).
In their social media marketing strategies, firms profit from efficiently interacting with consumers via firm-generated-content (FGC) (Kumar et al., 2015) or the increasingly stronger influence of new reference groups through electronic Word-of-Mouth (eWOM) in social networks which can also be controlled by firms (Shankar et al., 2016). With all those new possibilities the online world offers, it has become increasingly important for e-retailer or manufacturer to gain a better understanding of the underlying mechanisms, which influence consumers along the decision process (Garbarino & Strahilevitz, 2004). Due to the rising awareness of the importance of this topic, different researchers have conducted studies, in which they take a closer look at the influence of FGC and parts of eWOM, such as Sponsored Recommendation Posts (SRP), on consumers’ purchase intentions and they measured different moderating variables (e.g. Lu, Chang & Chang, 2014; Kumar et al., 2015). Their research shows the significant positive effect of FGC and SRP on consumers’ willingness to purchase. While the effectiveness of FGC and SRP as firm-controlled social media marketing activities were proved, the researchers proposed the importance of further investigating different variables, which could strengthen the effect on consumers’ decision making. But this prior research only focused on one type of product or one type of social media communication and did not take their interaction or effects of different message types into account. Thus differentiating between different types of social media communication, product characteristics (hedonic vs. utilitarian) or types of messages (persuasive vs. informative) seems to be an important extension for future studies, as prior research has measured separate direct effects of these variables on consumers purchase intention (Overby & Lee, 2006; Maheswaran & Sternthal, 1990).
2 messages influence consumers’ decision process among different types of social media posts (Kumar et al., 2015) in the online social media environment. This leads to the following research question:
3
2. Theoretical framework
To examine these influences mentioned in the introduction, previous research will further be summarized and findings of the impact of FGC and SRP for different product types on consumers’ online purchase intention and possible moderating effects of the type of message will be analyzed. Based on literature and prior research hypotheses to address in this study will be created.
2.1. Firm-generated Content
Firms increasingly use social media to interact with customers. During the past years, different aspects of firm’s engagement through social media have been studied (Kumar et al., 2015). One widely used initiative of a firm’s engagement with consumers on social media is FGC. FGC is a tool with which firms can “tell customers about their current product offerings, prices, and promotions” and which has an advertising function that influences the consumer decision process (Kumar et al., 2015 p. 3; Vakratsas & Ambler 1999). Interacting with consumers via FGC (e.g. Post on Facebook Fan page) can have a positive effect on brand equity and building or retaining relationships, especially with loyal customers, brand aficionados or fans (Schivinski & Dabrovski, 2016; Kumar et al., 2016). Thus, FGC has the ability to create or capture incremental shareholder value (Gensler et al., 2013; Keller, 1993). Another study by Lea (2012) at Inc.com supports this argument as customers’ and firms’ interaction on Social Media turns out to be beneficial for both parties. In line Kumar et al. (2015) who found, that FGC has a positive effect on purchase intention of experimental products, and prior mentioned research results, it can be hypothesized independently of the type of product, thatFGC, in general, has a positive effect on online purchase intention.
4 utilitarian and hedonic value perceptions, such as the product type. Products can offer both utilitarian and hedonic value, whereas one of these dimensions mostly dominates consumers’ product evaluations for certain types of goods (Batra & Ahtola, 1990). Utilitarian goods, like minivans, raincoats or makeup remover, are primarily functional, goal-oriented and instrumental products, whereas the consumption of hedonic goods (e.g. sports cars, designer handbags or designer lipstick) primarily provide consumers with emotional, exciting, fun and pleasurable experiences (Hirschman & Morris 1982).
As mentioned before, using FGC, firms can inform customers about current product offerings, prices, and promotions and reinforce favorable brand attitudes by interacting with consumers (Kumar et al., 2015). This functional information about current offerings, prices and promotions offers utilitarian value to consumers and matches the primarily utilitarian reasons, consumers shop online (Overby & Lee, 2006). Consequently, these utilitarian posts are expected to be more positively processed by consumers when they are about products that offer functional, utilitarian value, namely utilitarian products (Overby & Lee, 2006; Maheswaran & Sternthal, 1990). Based on these findings, it can be hypothesized, that
H1) if the firm-generated post is about a utilitarian product, consumers’ purchase intention is expected to be higher than if the post is about a hedonic product.
2.2. Sponsored Recommendation Posts
5 therefore have the ability to spread information widely among the social network societies (Watts & Dodds, 2007). Companies make use of Influencer as part of their social media strategy and pay them to write a recommendation post about their product or service which they will then spread across their followers to create awareness and influence consumer decision process towards consideration and purchase of the presented brand, product or service (Lu, Chang & Chang, 2014; Bouhlel et al., 2010).
Since many online marketers differ in products and target segments, it would be useful to understand how the effect of SRP on purchase intention differs among product types and types of messages. This knowledge could then be used to improve eWOM tools within targeted digital marketing strategies. And previous research has indeed shown, that the nature of hedonic products makes a recommendation harder than for utilitarian products since emotional experiences are rather subjective (Hsieh et al., 2005). For products that are focused on functional benefits, like utilitarian or search goods, consumers might find it easier to obtain reliable information on product quality and functional benefits prior purchase (Mudambi & Schuff 2010). SRP’s might therefore be more helpful and credible for consumers who want to purchase utilitarian products than for hedonic products. Accordingly, it can be hypothesized, that
H2) if the Sponsored Recommendation Post is about a utilitarian product, consumers’ purchase intention is expected to be higher than if the post is about a hedonic product.
2.3. Influence of Type of Message
6 Different researchers have found, that the type and quality of message moderates the effect of the message in the consumer decision journey. Park, Lee & Han (2007) for example examined the influence of information characteristics like relevance, understandability, sufficiency, and objectivity. They found, that messages that give sufficient reasons based on facts, are logical and persuasive and have a positive effect on purchase intention whereas messages that are emotional, subjective and without any factual information have a negative effect on purchase intention. Additional research has found, that preferences of consumers differ due to their level of expertise. Based on their findings Alba & Hutchinson (1987, p. 426) assume, that "physical features may be meaningless to novices, so advertisements directed at them are structured around easily comprehended benefits". Being more understandable, they can be perceived as informative more easily and could further be progressed in the decision process of novices. Experts, on the contrary, prefer messages with attribute information only as they have enough knowledge to process the meaning and value of certain product attributes (Maheswaran & Sternthal 1990).
As FGC in this study consists of a post about product offers, prices, and promotions at the Facebook Fan page of a brand, which sounds more fact-based, it is highly interesting to measure, if consumers actually prefer the post to be more informative (fact-based) or more persuasive (benefit-based). When using FGC, previous literature has characterized the addressed consumers as loyal consumers, aficionados or fans (Schivinski & Dabrovski, 2016; Kumar et al., 2016). Accordingly, it can be assumed, that they are likely to have a higher level of brand expertise, as the consumers are more likely familiar with the brand when intentionally following it on Facebook. Consequently, they can be rather seen as experts than novices (Maheswaran & Sternthal, 1990). As clear informative messaging seems to be more appealing for experts than persuasive posts and FGC overall seems to have a positive effect on online purchase intention for both utilitarian and hedonic products, it can be hypothesized, that
H3a) the type of message moderates the relationship between FGC and online purchase intention for both types of products
H3b) and the moderating effect is stronger if the type of message is informative.
7 to be more understandable and easier to progress for novices, they are likely to be more effective when used for SRP than informative messages (Maheswaran & Sternthal, 1990). When taking the product types into account, research has shown, that persuasive messaging based on benefits coming from a strong tie (influencer) is more successful for novices when the advert is addressing hedonic products, as it is more affective and sensory. Utilitarian products, on the other hand, are goal-oriented and problem-solving, which is why more fact-based informative messages would be more successful. Due to the lack of information processing power of novices, persuasive messages would consequently be more successful in moderating the effect of the post on online purchase intention (Kelman, 1961; Maheswaran & Sternthal, 1990; Wen, Tan & Chang, 2009). Consequently, it can be hypothesized, that
H4a) the type of message moderates the relationship between SRP and online purchase intention for both types of products
8
3. Methodology
Based on prior research and the resulting research hypotheses, the moderated conceptual model (see Figure 1) has been developed. The dependent variable consists of a late stage in the consumer decision process, the online purchase intention. This variable displays if the consumer is willing to purchase the product or service mentioned in the post.
Figure 1: Conceptual Model
The independent variables display the use of FGC versus the use of SRP. For both social media communication types, two characteristics namely utilitarian and hedonic products are chosen to differentiate between the product types presented in the posts. The different types of social media communication (FGC and SRP) will be measured in two separate studies in order to examine separate effects. But to give managers and researchers further valuable insights, the two studies will be made comparable by setting a comparable experimental setting with comparable participation.
As the literature research has resulted in the hypothesis, that the type of message would have an effect on the effectiveness of advertising, it has been added to the model in a moderating role. The type of message differentiates purely informative messages based on facts and attributes from more persuasive messages based on benefits.
9 In the setting of SRP (study II) online purchase intention is hypothesized to also be higher for utilitarian products than hedonic products and the effect will be moderated by the type of message such, that the influence is stronger for persuasive content than informative content.
3.1. Research Method
In order to test the hypotheses a self-administered online questionnaire with 30 questions (see Appendix II) was conducted at qualtrics.com to collect quantitative primary data (Mooi & Sarstedt, 2011). Before the main quantitative survey, a pre-test with five persons was conducted to ensure timing and understandability. To evaluate the hypotheses, stimulus material of the manipulation variables of prior conducted research was used, as they have already been validated.
3.2. Data Collection
As for quantitative surveys it is important to get gain a large data set, the online survey was spread via Facebook, WhatsApp and E-Mail and it was appealed to share the survey with further networks to make use of the snowball procedure (Mooi & Sarstedt, 2011). By spreading the survey across different networks, a more heterogeneous sample could be gained, which could result in more generalizable outcomes. This way, a quantitative sample size of 165 participants could be achieved. Additionally, by spreading the survey via social media, it was controlled for the actual usage behavior of social media for most of the participants, which is of importance for the subject of this study.
3.2.1. Participants
10 Table 1 shows, that 64,2% of the participants were females and 35,8% were men, so the gender participation of this study was not proportional. Most respondents were between 25 and 34 years (49,1%) or 18 and 24 years (41,8) old, which supports the validity of this study as these age groups have a high
amount of social media affinity and will be able to realistically imagine the situation and answer the questions (Hernández et al., 2011). About 62,4% of the participants were from Germany, 16,4% from the Netherlands, 17% from other EU countries and about 4,2% from countries outside the EU. The participants had a high level of education, as about 79,4% had a Bachelor degree or higher and most of the participants had a good to very good command of English (91,5%) which also contributes to the validity of the answers due to the resulting understandability.
3.2.2. Procedure and Measurements
To test the hypotheses, an experimental methodology with a 2 (type of product: utilitarian vs. hedonic product) x 2 (type of message: informative vs. persuasive) between-subjects factorial design has been chosen for each of the two studies. Experimental settings offer a high internal validity and higher precision of measurements and a between-subject factorial design has the advantage over a within-subject design, that participants are not influenced by order effects like practice effects, fatigue effects or carryover effects (Aronson et al., 1998; de Winter & Dodou 2017).
11 assigned to one of two situations, namely either sitting at a couch visiting Facebook (study I) or Instagram (study II). After presenting each of these two conditions, the participants were equally randomly assigned to one of four different stimuli, consisting of two different types of products (hedonic and utilitarian) paired with two different types of messages (persuasive and informative). Consequently, the two studies each consisted of a total of four different conditions (see Appendix I). Depending on the social media setting, the participants were assigned to, they had to read one of four different messages. After carefully reading the posts, the participants had to answer different questions, containing manipulation checks, which will be described in detail in the next section (see Appendix II). The measurements were based on 7-point scales (1= strongly disagree – 7= strongly agree) to ensure validity and easiness of use for the participants (Preston & Colman, 2000).
3.2.2.1. Manipulation of Type of Post and Product Type
The two studies do not only measure the general effects of the two types of firm-controlled posts FGC and SRP on online purchase intention, but they also take different types of products into account. To better predict the output when differentiating between product types, this study concentrates on hedonic and utilitarian product characteristics. Recent experiments tested the strength of utilitarian or hedonic characteristics of different products and found, that designer clothes are strongly perceived as hedonic and a vacuum cleaner as utilitarian (Chang, Chen & Tan 2012). These products are used as manipulation checks to measure the effects of hedonic and utilitarian product types mentioned in FGC and SRP’s on online purchase intention. In order to eliminate confounding effects, which could be caused by already known brand names, fictitious brand names were designed for the products: Luxionista Designer Clothes & Dirtaway Vacuum Cleaner (Kent & Allen 1994).
For each study with one of the two types of firm-controlled posts, namely SRP and FGC, each of the two product types will be measured. Based on the before mentioned 2 (Type of product: utilitarian vs. hedonic product) x 2 (Type of message: informative vs. persuasive) experimental stimulus, four versions of posts were created for each study to observe the reaction of the participants in the social media environment. Besides the text description of the post, pictures of the products were added to create a realistic situation.
12
3.2.2.2. Manipulation of Type of Message
In order to test the moderating effect of the type of message, the previously mentioned posts were either written in an informative or persuasive way. As manipulation checks, the informative post listed short, functional attributes about the product, whereas the persuasive post presents extensive benefits the product offers (see Appendix I). In order to control the effectiveness of this manipulation, characteristics of informative and persuasive posts validated by Maheswaran & Sternthal (1990) and Braverman (2008) were chosen to create informative and persuasive posts (see Appendix I)
3.2.2.3. Online Purchase Intention
To measure the treatment effect of the dependent variable, online purchase intention, three items developed by Mackenzie et al. (1986) were used (see Appendix II). The stronger the participants agree on each item, the stronger the purchase intention. To measure the agreement with each item, the participants were instructed to rate each item on a 7-point semantic differential scale.
3.2.2.4. Control Variables
13
3.3. Plan of Analysis
For the analysis, both studies were a 2x2 between-subject factorial design with two categorical independent variables. For the four groups per study, a medium efficiency (f=0.25) with a 0.05 significance level and a power setting of 0.8 was expected (Bortz & Döring, 2007). With 83 participants for study I and 82 participants for study II, which were equally assigned to one of the four different conditions per study, the required observation minimum per study was concisely exceeded and the analyses could be conducted.
In order to test the two studies with 2x2 between-subjects factorial designs with two categorical independent variables, two-way analyses of variance (ANOVA) were conducted. This method offers the possibility to involve two factors simultaneously as well as the moderating or interaction effect of these two factors (Malhotra, 2010).
14
4. Results
For each of the two studies, the manipulation checks were conducted first, to ensure, that the variables were perceived the way they were intended to. After the manipulation checks, robustness checks were conducted to see whether other variables mentioned by prior research cause external influence on the variation in the dependent variable and should be accounted for in the model or if they could be ignored. Finally, the hypotheses were tested for each of the studies.
4.1. Study I
4.1.1. Manipulation Checks
In order to check the reliability of the scales, Chronbach’s Alpha was checked for the items. For the type of product, Cronbach's Alpha for the items belonging to the utilitarian product characteristics and Cronbach's Alpha for the items belonging to the hedonic product characteristics were both above 0,86. Consequently, the scales were highly reliable and could be summed and averaged to compose an evaluation score. After composing an average score for both utilitarian and hedonic product categories, a dummy variable was created with 1 = utilitarian perception and 0 = hedonic perception to check whether the participants who were randomly assigned to the utilitarian or the hedonic product stimuli perceived the given product correctly. After conducting the Pearson's
Chi-Square test with a stimuli * type of product perception cross-tabulation it was clear, that the products were perceived correctly and the highly significant asymptotic Pearson Chi-Square significance indicated that the observed differences between the groups did not arise by chance (see Table 2).
Cronbach's Alpha for the items belonging to purchase intention also had a very high score with 0,89, which also proved a high reliability of the scales and they were summed and averaged as well.
4.1.2. Robustness Check
15 advertising (FGC). After measuring the Chronbach’s Alpha for the three items (α = 0,878) it was proved, that there was a high reliability of the scales for the different treatments so they could be summed and averaged to one control variable “Attitude.factor”.
In a separate model, the influence of the variable on online purchase intention was measured and in line with the study of
Chang, Chen & Tan (2012) a significant linear effect (ß = 0,456, t(81) = 4,615, p < 0,05) could be observed (see Table 3).
To be able to conduct an ANCOVA, there are different assumptions that have to be met, in order to make sure, that the data which has to be analysed can actually be analysed using an ANCOVA and that the ANCOVA offers valid results. To check if the data can be analysed by an ANCOVA, 10 assumptions have to be met (Lund & Lund, 2018). But praxis has shown, that it is not uncommon, that a few of the 10 assumptions might be violated. In this case, the researcher has to decide, whether to proceed (if a few assumptions are not met) or to stop the analysis at this point and exclude the variable from the analysis (if too many assumptions are not met).
When testing the assumptions for the Attitude.factor, a high amount of assumptions, e.g. variation of covariate values across different levels of the independent variable, homogeneity of regression, normal distributions and homogeneity of variances could not be met. This indicated, that the variable could not be used as a control variable as the outcomes of the ANCOVA would most likely not offer valid results. Consequently the variable was excluded and not used as a control variable in the main model.
Another variable, prior research mentioned to influence the impact of different types of messages on purchase intentions, was product expertise (Maheswaran & Sternthal, 1990). When looking at the main linear effects of product expertise on online purchase intention in Table 4, a strong significant effect can be observed (ß = 0,401, t(81) = 3,941, p < 0,05). In order to test whether the
variable can be added to the model as a covariate and whether analysing it in an ANCOVA would offer valid
16
4.1.3. Hypothesis testing
To test the hypothesis H1, a one-way ANOVA was conducted to measure the main effect of the type of product on the dependent variable because once an interaction effect exists in the two-way ANOVA, the interpretation of the main effects might be misleading (Malhotra, 2010). Table 5 shows, that the
main effect of the type of product on online purchase intention is significant at the 90% level (F (1,81) = 2,755, p < 0,1).
Consequently, the purchase intention differs between hedonic and utilitarian products and it is higher for utilitarian products (Meanutilitarian = 2,65) than for hedonic products
(Meanhedonic = 2,17) (see Table 6). Accordingly,
H1 is supported. But it has to be added, that the effect size is low, as only 3,3% of the variance in purchase intention can be attributed to this main effect (see Partial Eta Squared in Table 5).
To test hypothesis H3a and b, a two-way ANOVA was conducted. Table 7 shows, that the interaction effect of the type of product and the type of message on online purchase intention is statistically significant (F (1,79) = 5,318, p < 0,05).
17 highly versatile variable, it could be of high value for managers and researchers to be able to predict 10% of the purchase intention of consumers with this model.
When looking at the profile plot (see Appendix III), the crossing lines indicate, that a significant disordinal interaction is present. A significant disordinal interaction in an ANOVA indicates, that the main effects of the model should not be interpreted, as the results depend on the treatment condition (Malhotra, 2010).
The output of the ANOVA shows, that there is a statistically significant interaction between the independent variables. But to further investigate this interaction, the simple effects have to be analyzed. The univariate test of the type of message indicates, that there is a significant difference between the types of products at the level of persuasive content but no significant difference between the types of products at the level of informative content (see Table 8). Consequently, the
moderating effect of type of message is stronger for persuasive content than for informative content, which means, that H3b can’t be supported.
4.1.4. Discussion
18
4.2. Study II
4.2.1. Manipulation Checks
Just like in study I, the reliability of the scales were checked for all the mentioned items. The outcomes were similar to the manipulation checks under 4.1.1., which is why they will not be discussed in detail at this point.
4.2.2. Robustness Check
Similar to study I it was analyzed, if the attitude of the participants towards the social media platform Instagram causes external influence on the variation in the dependent variable. As Table 9 shows, that this variable
has a significant main effect on the dependent variable
(ß = 0,303, t(80) = 2,844, p < 0,05), it had to be controlled
for in the main model. Consequently, the ANCOVA assumptions were tested, in order to examine if this variable could be used as a control variable in the main model and if the ANCOVA would offer valid results. As almost all assumptions were met (e.g. variation of covariate values across different levels of the independent variable, homogeneity of regression, normal distributions and homogeneity of variances), the Attitude.factor was added to the main model ANCOVA.
Additionally to the Attitude.factor, significant linear effects of product expertise on the dependent variable could be observed (ß = 0,244, t(80) = 2,247, p < 0,05) (see Table 10). But similar to
19
4.2.3. Hypothesis testing
To test the hypothesis H2, a one way ANOVA was conducted to measure the main effect of the type of product on the dependent variable. Table 11 shows, that the main effect of the type of product on online
purchase intention is not significant (F (1,80) = 0,262, p = 0,610). Consequently, the purchase intention does
not differ between hedonic and utilitarian products in the Instagram setting. Accordingly, H2 is not supported.
To test hypothesis H4a and b, a two-way ANCOVA was conducted. Table 12 shows, that the interaction effect of the type of product and the type of message on online purchase intention is also not significant
(F (1,77) = 0,904, p > 0,1), even after controlling for the Attitude.factor. In fact, after looking at the profile plots (see Appendix IV) it appears, that the purchase intention for utilitarian
products is higher with informative messaging and for hedonic products, it is higher with persuasive messaging, but as the differences in the results are too small and therefore not significantly different, H4a and b cannot be supported. But the table also shows, that the
Attitude.factor is significantly related to the dependent variable online purchase intention (F (1,77) = 8,216, p < 0,05), which indicates, that this variable is valuable to the decision process and should be analysed in detail in further studies.
When comparing the model that includes the covariate (ANCOVA) to the model that does not include the covariate (ANOVA), the ANCOVA better explained the relationship between the variables. The amount of variation accounted for by the model (see Partial Eta SquaredCorrected model) has increased from 0,016 (ANOVA) to 0,111 (ANCOVA) and the independent variables
20
4.2.4. Discussion
The goal of study II was to investigate consumers’ purchase intention in a social media environment that is often used for SRP, namely Instagram. Similar to study I, the aim was to measure the hypotheses that the purchase intention differs among types of products and that this relationship is moderated by the type of message. Additionally, the robustness check has shown, that the attitude of the participants toward Instagram also has a significant impact on purchase intention. Consequently, it was added as a control variable in the main model, to extract extraneous influences on the dependent variable. In contrast to study I, the results did not support the hypotheses H2, H4a and H4b, as the purchase intention did not differ significantly among the type of product and there was no statistically significant interaction effect of type of product and type of message, even though it was controlled for consumers’ attitude towards the social media platform. In summary, it can be said, that sponsored recommendation posts at Instagram are equally successful for utilitarian and hedonic products with either persuasive or informative messaging.
4.3. Comparing the two studies
The conceptual model presented in chapter 3 implied the comparability of the two studies. In order to see if the two studies could be compared with each other, the groups differences based on demographics were observed first. Pearson’s Chi² for the demographic variables of the different social media settings was not significant (Chi² = 0,911), as can be seen in Table 13. Consequently, H0 could not be
rejected and it was proven, that there were no significant differences between the group's demographics, which also contributes to the comparability of the two samples.
As a second step an independent samples t-test for the independent samples was conducted. Levene’s test was not significant, with F(163) = 0,366, p = 0,546 (see Table 14), so the assumption H0 for homogeneity of
21 The t-test for the Equality of Means was significant on the 90% level (see Table 15). T (161) = -1,662, p = 0,098 indicates, that the second mean was larger than the first mean. So the participants in the Instagram condition were more willing to purchase the products, than in the Facebook
condition independently of the type of product or message.
When looking at the different means of the interactions, the t-tests displayed in Table 16 show, that there is no significant difference between both the participants who were assigned to the condition with a hedonic product and an informative content (t (35) = -0,188, p = 0,852) and the participants who were assigned to the condition with a utilitarian product and persuasive content (t (45) = 0,761, p = 0,451). There were significant differences in the groups assigned to the condition with a hedonic product and persuasive content (t (34) = -2,126, p = 0,041) at the 95% level, and the groups assigned to the condition with a utilitarian product and informative content (t(43) = -1,692, p = 0,098) at the 90% level.
22
5. Conclusions & Recommendations 5.1. Conclusion
Since social media marketing has become an important topic in marketing management and research, investigating success factors in this field has become increasingly important. But firm’s control over social media communication is limited to a few activities, such as FGC or SRP. Prior research has shown, that the success of marketing communication activities differs among products or message types (Overby & Lee, 2006; Maheswaran & Sternthal, 1990), but there were no studies that included both variables in one model. This thesis presents two studies, one for each type of firm-controlled social media communication technique, measured by a 2 (type of product) x 2 (type of message) between-subject experimental design. The valid sample consisted of 165 participants, who were randomly assigned to one of the two studies with four treatments each. Their purchase intention was measured with regard to the different combinations of product type (hedonic vs. utilitarian) and message type (informative vs. persuasive).
23 product, but these results are not significantly significant. This stands in contrast with the outcomes of Maheswaran & Sternthal (1990), as it indicates, that product expertise is not a good predictor of which type of message to use for a successful online communication in the Facebook setting.
Within the treatment of SRP at Instagram, consumers’ purchase intention did not differ significantly between types of products and types of messages and there is no significant interaction effect (see Figure 2: Study II). This result stands in contrast to the study of Lu, Chang & Chang (2014), who found significant differentiating effects between different types of products (search good and experimental good), which have similar characteristics as the products presented in this study. But different to this study they measured the influence on the attitude towards the sponsored recommendation post and as a next step, they examined the overall positive effect of the attitude on purchase intention. They did not measure the direct effect of the post about a product on purchase intention, which could be one reason for the different outcomes.
Figure 2: Overview Results
24 had a slightly higher purchase intention of utilitarian products when informative messaging was used than in the Facebook condition when persuasive messaging was used.
These outcomes are new to the field of marketing research, as there is no existing research about the comparison of the success of FGC and SRP for different products and types of messages.
5.2. Recommendations 5.2.1. Managerial Implications
Based on the findings of the two studies, several contributions can be made to the field of social media marketing. This study offers assistance to marketing managers when they have to make decisions about where to invest in social media communication for which type of product and how to communicate successfully.
Within the Instagram condition, there were no statistically significant differences between the groups which indicate, that the overall purchase intention in the Instagram condition is not statistically dependent on the type of product or type of message. This output indicates, that marketing managers do not have to pay too much attention on how to communicate which type of product in the setting of SRP’s, as each of the combinations appear to be similarly successful. But this is just partly true.
When taking the outcomes of the comparison of the two social media platforms into account, there are significant differences between types of product and types of messages. The comparison of the two studies shows, that consumers’ overall purchase intention is higher for SRP than for FGC, and especially for hedonic products in combination with persuasive content. This highlights the importance of SRP in form of influencer posts at Instagram as part of a successful social media marketing campaign.
25
5.2.2. Limitations and implications for further Research
26
References
Aronson, E., Wilson, T. D., & Brewer, M. B. (1998), Experimentation in social psychology. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), The handbook of social psychology (pp. 99-142). New York, NY, US: McGraw-Hill.
Batra, R., and Ahtola, O.T., (1990), Measuring the Hedonic and Utilitarian Sources of Consumer Attitudes. Marketing Letters, 2, 159-170.
Bortz, J., and Döring, N. (2007), Forschungsmethoden und Evaluation: für Human- und Sozialwissenschaftler [Research Methods and Evaluation for Social Scientists]. Berlin: Springer. Bouhlel, O., Mzoughi, N., Ghachem, M. S., & Negra, A. (2010), Online purchase intention: Understanding the blogosphere effect. International Journal of e-Business Management, 4(2), 37–51.
Braverman, J. (2008), Testimonials versus informational persuasive messages: the moderating effect of delivery mode and personal involvement. Communication Research, 35, 666–694. Chang, C.C. & Chin, Y.C. (2010), The Impact of Recommendation Sources on Online Purchase Intentions: The Moderating Effects of Gender and Perceived Risk. International Journal of Economics and Management Engineering, Vol. 4, No. 6.
Chen, Q., & Wells, D. (1999). Attitude toward the Site. Journal of Advertising Research, 27-37.
De Winter, J.C.F & Dodou, D. (2017), Human Subject Research for Engineers. A Practical Guide. Springer International Publishing.
Retrieved from https://www.springer.com/de/book/9783319569635
Dhar, R., & Wertenbroch, K. (2000). Consumer Choice Between Hedonic and Utilitarian Goods.
Journal of Marketing Research, 60-71.
Garbarino, E. & Strahilevitz, M. (2004), Gender differences in the perceived risk of buying online and the effects of receiving a site recommendation. Journal of Business Research 57, 768– 775.
27 Gilly, M. C., Graham, J. L., Wolfinbarger, M. F., & Yale, L. J. (1998). A dyadic study of personal information search. Journal of the Academy of Marketing Science, 26(2), 83–100.
Goldsmith, R.E. (2006). Electronic word-of-mouth. In Khosrow-Pour, M. (Ed.), Encyclopedia of E-Commerce, E-Government and Mobile Commerce, Idea Group Publishing, Hershey, PA, 408-12.
Harmon, R. R., & Coney, K. A. (1982). The persuasive effects of source credibility in buy and lease situations. Journal of Marketing Research, 19(2), 255–260.
Hernández, B., Jiménez, J., & Martín, M. J. (2011). Age, gender and income: do they really moderate online shopping behaviour? Online Information Review, 35 (1), 113–133.
Hirschman, E.C. & Holbrook, M.B. (1982), Hedonic Consumption: Emerging Concepts, Methods and Propositions, Journal of Marketing, 46 (Summer), 92-101.
Hsieh, Y. C., Chiu, H. C., & Chiang, M. Y. (2005), Maintaining a committed online customer: A study across search-experience-credence products. Journal of Retailing, 81(1), 75–82.
Keller, K.L. (1993), Conceptualizing, Measuring, and Managing Customer-Based Brand Equity,
Journal of Marketing, 57, 1, 1-22.
Kelman, H. C. (1961). Processes of Opinion Change. Public Opinion Quarterly, 57-78.
Kent, R. J., & Allen, C. T. (1994), Competitive Interference effects in consumer memory for advertising: The role of brand familiarity. Journal of Marketing, 97-105.
Kumar, A., Bezawada, R., Rishika, R., Janakiraman, R., & Kannan, P.K. (2015), From Social to Sale: The Effects of Firm-Generated Content in Social Media on Customer Behavior, Journal of Marketing, 80 (January), 7–25.
Lamberton, C., & Stephen, A. T. (2016), A Thematic Exploration of Digital, Social Media, and Mobile Marketing: Research Evolution from 2000 to 2015 and an Agenda for Future Inquiry.
Journal of Marketing, 80 (November), 146-172.
Lea, W. (2012), The New Rules of Customer Engagement, Inc.com, (accessed September 29, 2018).
28 Lee, J., Park, D. H., & Han, I. (2008), The effect of negative online consumer reviews on product attitude: An information processing view. Electronic Commerce Research and Applications, 7, 341–352.
Lu, L.C., Chang, W.P. & Chang, H.H. (2014), Consumer attitudes toward blogger’s sponsored recommendations and purchase intention: The effect of sponsorship type, product type, and brand awareness. Computers in Human Behavior, 34 (2014) 258–266.
Lund, A. & Lund, M. (2018), One-way ANCOVA in SPSS Statistics. Blog post at Lund Research Ltd (online). (Accessed December 14, 2018).
Retrieved from https://statistics.laerd.com/spss-tutorials/ancova-using-spss-statistics.php Mackenzie, S. B., Lutz, R. J., & Belch, G. E. (1986). The Role of Attitude toward the Ad as a Mediator of Advertising Effectiveness: A Test of Competing Explanations. Journal of Marketing Research, 130-43.
Maheswaran, D. & Sternthal, B. (1990), The effects of knowledge, motivation, and type of message on ad processing and product judgments. Journal of consumer Research, 17 (1990), 66-73.
Maheswaran, D., & Chaiken, S. (1991). Promoting Systematic Processing in Low-Motivation Settings: Effect of Incongruent Information on Processing and Judgment. Journal of Personality and Social Psychology (61:1), pp. 13-25.
Malhotra, N.K. (2010), Marketing Research. An applied orientation. Pearson Education, Inc. New Yersey.
Mooi, E., & Sarstedt, M. (2011), A Concise Guide to Market Research. The Process, Data, and Methods Using IBM SPSS Statistics. Berlin Heidelberg: Springer-Verlag.
Mudambi, S. M., & Schuff, D. (2010), What makes a helpful online review? A study of customer reviews on Amazon.com. MIS Quarterly, 34(1), 185–200.
Muehling, D. D. (1987). An Investigation of Factors Underlying Attitude Toward Advertising in General. Journal of Advertising, 16 (1), 32-40.
29 Park, D. H., Lee, J., & Han, I. (2007), The effect of on-line consumer reviews on consumer purchasing intention: The moderating role of involvement. International Journal of Electronic Commerce, 11(4), 125–148.
Preston, C.C. & Colman, A. (2000), Optimal number of response categories in rating scales: reliability, validity, discriminating power, and respondent preferences. Acta Psychologica, 104, pp. 1–15.
Putrevu, S., & Lord, K. R. (1994). Comparative and noncomparative advertising: Attitudinal effects under cognitive and affective involvement conditions. Journal of Advertising, 23(2), 77– 91.
Román, S. (2010). Relational Consequences of Perceived Deception in Online Shopping: The Moderating Roles of Type of Product, Consumer’s Attitude Toward the Internet and Consumer’s Demographics. Journal of Business Ethics, 95:373–391.
Schivinski, B. & Dabrowski, D. (2016), The effect of social media communication on consumer perceptions of brands. Journal of Marketing Communications, 22:2, 189-214.
Shankar et al. (2016), Mobile Shopper Marketing: Key Issues, Current Insights, and Future Research Avenues. Journal of Interactive Marketing, 34, 37–48.
Srinivasan, T.J. (1987). An Integrative Approach to Consumer Choice. In Melanie Wallendorf and Paul Anderson, Eds., Advances in Consumer Research, 14, Provo, Utah: Association for Consumer Research, 96-101.
Vakratsas, D. & Ambler T. (1999), How Advertising Works: What Do We Really Know? Journal of Marketing, 63 (January), 26–43.
Watts, D.J. & Dodds, P.S. (2007), Influentials, Networks, and Public Opinion Formation. Journal of Consumer Research, 34 (4), 441–58.
Wedel, M. (2011), Is segmentation history? In Wieringa, J.E. et al. (eds). Liber Amicorum in honor of Peter S.H. Leeflang,University of Groningen, 99-110.
30 Woods, S. (2016). #Sponsored: The Emergence of Influencer Marketing. Honors Thesis, University of Tennessee, Knoxville, USA. (Accessed December 29, 2018)
Retrieved from http://trace.tennessee.edu/utk_chanhonoproj/1976
31
Appendices
Appendix I: Description of stimulus and its sample size
Appendix II: Measures for Constructs
Construct Source Scale Measure Type of Product Dhar & Wertenbroch
2000
7-point Likert Scale 1. The product is functional. 2. The product is practical. 3. The product is useful.
4. The product helps achieve a goal. 5. The product is pleasant.
6. The product is fun. 7. The product is enjoyable. 8. The product appeals to senses. Strongly disagree (1) / strongly agree (7) Product Expertise 7-point Semantic
Differential Scale
1. How high is your expertise on this type of product? (not the special brand)
32
Purchase Intention Mackenzie et al. 1986 7-point Semantic Differential Scale
1. How likely are you going to purchase the product?
Extremely unlikely (1) / extremely likely (7) 2. How willing are you going to purchase the product?
Extremely unwilling (1) / extremely willing (7) 3. Do you plan to purchase the product? Absolutely not (1) / absolutely yes (7) Experience with Social
Media Platforms
Kent & Allen 1994 7-point Semantic Differential Scale
1. How experienced are you with the Social Media Platform Facebook?
Very inexperienced (1) / very experienced (7) 2. How experienced are you with the Social Media Platforms Instagram?
Very inexperienced (1) / very experienced (7) 3. How often do you visit the Social Media Platform Facebook?
Not at all (1)/ very often (7)
4. How often do you visit the Social Media Platform Instagram
Not at all (1)/ very often (7) Attitude towards Social
Media Platforms
Chen & Wells 1999 7-point Likert Scale 1. Facebook makes it easy for me to build a relationship with my friends.
2. I would like to visit Facebook again in the future.
3. I am satisfied with the service provided by Facebook.
4. I feel surfing through Facebook is a good way for me to spend my time.
5. Compared to other Social Media Platforms, I would rate this as one of the best.
____________________________________ 1. Instagram makes it easy for me to build a relationship with my friends.
2. I would like to visit Instagram again in the future.
3. I am satisfied with the service provided by Instagram.
4. I feel surfing through Instagram is a good way for me to spend my time.
5. Compared to other Social Media Platforms, I would rate this as one of the best.
Strongly disagree (1) / strongly agree (7) Attitude towards Social
Media Advertising
Muehling 1987 7-point Semantic Differential Scale
1. What is your attitude towards Social Media Advertising through Firms Post on Facebook? a) Bad (1) / Good (7)
33
________________________________ 2. What is your attitude towards Social Media Advertising through Influencer Posts on
Instagram?
a) Bad (1) / Good (7) b). Negative (1) / Positive (7) c) Unfavorable (1) / Favorable (7)
34
Appendix IV: Profile Plot study II
