1 TABLE OF CONTENTS

(1)

(2)

AKNOWLEGDGMENT

This research was supported by so many people and I want to acknowledge all of them who have helped me and supported me through this journey. First of all I would like to thank my supervisor for his patience, understanding, guidance and invaluable contributions through all these months of writing.

Second, I would like to thank all the people and the fellow classmates that participate in my questionnaire and especially to those of the respondents that help me through their social network to collect the needed amount of data.

Most of all, I want to acknowledge my mother for being my rock and my inspiration, throughout my entire academic career and especially a person who is always by my side and without his help, this master would not even started. Thank you Sakis, for your support and love during this period of my life. Lastly, I want to thank my beloved grandmother, who may not be present now at my life but her star always guides me.

(4)

1. INTRODUCTION

During the history of the human beings in earth, communication was always a vital and basic need not only for the development of the society but it was also crucial for the self-deployment. Especially, during the last decade, something changed in a dramatic way, the establishment of Internet arrived to change the way that people are able to communicate. The advantages of This technology have resulted in the social media platforms and their explosion. Communication, searching for information and expression of personal ideas are easier through the social media platforms. Following the studies of Constantinides and Fountain (2008) social media is (p. 232) “a collection of open-source, interactive, and user-controlled online applications expanding the experiences, knowledge and market power of the users as participants in business and social processes.”

The presence of social media has changed and even replaced the traditional media (Television, radio) and the opportunities that can offer, seem to be limitless (Bruhn et al., 2012). Furthermore, these platforms can provide advantages to companies which are enabled with. Nowadays, the majority of people and companies use and participate actively in social media platforms and the concept of the engagement with a brand through them is a challenge for companies. The main difference, compare to traditional media is that, social media provides the individual with the opportunity to become the content-creator, and customers are able to engage with brands in terms of sharing, favoring, and retweeting within the individual’s own personal social network.

“Engage or die” is the new marketing moto, which is the result of the development of social media the last decade (Nelson-Field & Taylor, 2012). Social media are considered as a two way communication with the customer, are interactive platforms which provide customers with the ability to have a direct conversation with the brands, in which they are interested. Through these opportunities the journey of the customer can be affected in different and new ways.

For brands and companies, the concept of engagement seems to be a key factor for the success, as they can create deeper and long lasting customer relationships with positive outcomes such as the word of mouth and brand loyalty.

Brand loyalty refers to a consumer-based construct that involves both behavioral response and psychological perception in the current examined period, as well as in past periods (Dick and Basu, 1994).. Additionally, the American Marketing Association provides us with another definition of brand loyalty-characterizes loyalty as a situation in which a consumer prefers to purchase repeatedly from the same manufacturer, instead of buying a product from different suppliers. Lastly, loyalty is considered also as the extent to which a consumer purchases the same brand when he/she has to decide within a product class (Moisescu, 2006).

An important outcome of the engagement on social media is e-Word of mouth. Word of Mouth can be defined as all kinds of informal communications addressed to other consumers about specific features, ways of usage regarding particular goods or services (Westbrook, 1987). E- word of mouth is any statement from a former or present customer which is spread to existing or potential customers via Internet and social media(Hennig-Thurau et al, 2004).

(5)

negative. Positive e-WOM communication is considered as a powerful marketing tool, which has an impact on consumers. It is more likely for the customers to spread the marketing message because they are satisfied with a brand (positive WOM) or because they are not satisfied with it (negative WOM). Both positive and negative e-WOM have different motivations behind it (Anderson, 1998). This research will focus on e-WOM as a potential outcome based on online brand engagement on Twitter.

1.1 TWITTER facts and features

Twitter is currently considered as the most popular microblogging service. Twitter was created in San Francisco by a small start-up business which was called Obvious Corp., and launched in October 2006(Wikipedia, 2015). Twitter motivates its users to share personal postings, called ‘tweets’, which consist of 140 characters and they are publicly available. Moreover, Twitter is optimized in order to be used anytime, anywhere. Users can share their tweets via twitter.com, text messaging, via Twitter’s mobile Website m.twitter.com, or using third party clients. When a user subscribes in Twitter.com, the website generates a unique code and the user is requested to enter personal information like location and a short biography (a description). Apart from creating and sending messages, via Twitter users are able to subscribe to receiving messages from other users by selecting the ‘follow’ a user choice. When a particular user shares a new tweet, all users that ‘follow’ that user, receive a notification of that message on their personal Twitter home page.

The number of people using Twitter has increased by more than 50 million in the past year (2015). The network now counts over 270 million active users. It is expected that by 2015, one fifth of internet users (20,5%) in the US will have Twitter accounts. This figure has grown from 15.2% in 2012, and is set to rise to 24.2% by 2018. 63% of brands have multiple Twitter accounts to engage with its customers. Finally, the average Twitter user follows five or more businesses and over a third (37%) of Twitter users will buy from a brand thy follow (Wikipedia, 2015).

1.2 Relevance: Literature contribution

(6)

explored customer engagement behavior on the social media platform of Facebook (Gummerus et al., 2012; Men & Tsai, 2014, Kabadayi & Price, 2014). However, social media evolve rapidly and new platforms are constantly emerging (Gummerus et al., 2012). Moreover, many researchers support that there is a gap focusing on customer engagement with a brand by using different social media platforms (Gummerus et al., 2012; Men & Tsai, 2014; Hollebeek et al., 2014; Sashi, 2012). In addition, Men and Tsai (2014) support that it is crucial and highly important to explore other social media platforms such as Twitter, Instagram, and LinkedIn, to understand deeply the potentials of social media. Hence, research on more and different platforms, will give us a deeper understanding of the effects of customer engagement and the potential positive outcomes for the brands. Therefore, it is interesting to investigate deeply the concept of engagement in the brand communities of other social media platforms, apart from Facebook.

Brands are interested in interacting with customers and this is a common theme in social media usage (Kaplan and Haenlein, 2010) and Twitter is considered as the most interactive social platform.

Despite that fact, Twitter’s strength and abilities are relatively unexplored, compared to other social media platforms. Moreover, Men and Tsai (2014) argue that it is important to explore other social networking sites other than Facebook.

Sashi (2012) argues that companies both within the private and public sector face difficulties in their attempt to connect with their customers effectively, in order to be engaged with their customers and hence, establish a loyal and long-term relationship with them. Following Bagozzi and Dholakia (2006), consumers who become fans of these brand fan pages tend to be loyal and committed to the company, and are more open to receiving information about the brand. Researches have tested the impact of online engagement to the e-WOM and brand loyalty using the fan pages that the brands have created in the platform of Facebook without examining the possibility of generalizing the results. Thus, little is known whether the findings can be applied to other types of social media.

As result, this study aims to fill this gap in the research in this field, scrutinizing the drivers of online brand engagement and the possible outcomes of e-WOM and brand loyalty in the case of Twitter brands’ accounts.

1.3 Problem discussion and research questions

The concept of engagement can have different meanings and this is the reason why academic literature suggests to approach this construct carefully. A full, understanding and deep communication between a customer and a brand demands personal interaction and physical proximity, which unfortunately results in a limited number of approachable customers. On the other hand, Internet is an open, highly cost-effective and far reaching global network, which helps reducing or even eliminating the barriers of geography and distance (Sawhney, Verona, &Prandelli, 2005). Hence, Internet provides the firms with the ability and the opportunity to overcome these constraints and reach a larger number of customers-taking a full advantage of the richness of the communication.

(7)

engagement (WARC, 2012a). Despite the popularity of the concept among companies, online brand engagement as a construct has not been successful to explain what can offer to the brand and what the important meaning is. The behavioral measures such as the followers, the visits and the retweets provide little information regarding the returns that are being expected (Nelson-Field & Taylor, 2012). As engagement in an online marketing environment is a relatively new term, there has to be a benchmark of how online brand engagement in social media leads to positive brand and marketing related objectives such as the e-WOM and brand loyalty.

As a result, the objective of this research is to identify the antecedents of the online brand engagement and the possible outcomes that can occur, specifically in the form of e-WOM and loyalty.

Based on this objective and in order to identify these relationships, three important questions have to be answered:

 Is possible content type (information and entertainment) and brand post characteristics (interactivity and vividness) affect online brand engagement?

 Is e-WOM a possible outcome of this engagement?

 Is brand loyalty a possible outcome of the brand engagement through Twitter?

(8)

2. THEORETICAL FRAMEWORK

The purpose of this study is to explore the concept of online brand engagement and the outcomes on a social media environment and to test the proposed conceptual model of online brand engagement (Figure 1) constructed for this study.

As can be seen in the conceptual model, this research investigates the drivers of brand engagement in the social media platform of Twitter and specifically the content type (information-entertainment) and the characteristics of the brand posts (interactivity and vividness).

In addition, it will be analyzed the impact of these antecedents on the different parts of brand engagement (namely cognitive, affective and behavioral). The framework exhibits brand engagement on online social media platforms as the central element embedded in the network of other constructs, which are divided into two groups of potential drivers (content type and posts ‘characteristics) and consequences (brand loyalty and e-wom).

Firstly, Jennings (2000) defines aesthetics of a website as the interesting, pleasurable, and enjoyable attributes present on a particular site and O’Brien and Toms (2008) support that the informational content and the aesthetic attributes present on a website capture participants’ attention (cognition) and interest (affect), which “moves them forward into engagement” (p. 943).

Another driver which plays a crucial role in the present research’s model is interactivity – following on Mollen and Wilson’s (2010) work. The researchers suggest that the consumers must interact with a website and support that the interaction is “two-way, controllable, and responsive to their actions” (Mollen & Wilson, p. 921).

According to Fortin and Dholakia(2005), interactivity can be defined as the extent to which a communication system is designed in such way that gives allowance to one or more end users to communicate as senders or receivers with one or many other users or communication devices, either in real time or on a store-and-forward basis or to look for and gain access to information on an on-demand basis where the content, timing and sequence of the communication is under control of the end user, as opposed to a broadcast basis.

Vividness refers to “rich” in content social media posts. Online rich media include a range of interactive methods that display motion and exploit sensory traits, such as video, audio, and animation (Chabrow 2006). The term "rich media" provides an umbrella expression to describe online content that has multimedia elements, such as sound, video, or content that moves when a user clicks on the page that features the content (Shaw 2004).

(9)

engaged in an online environment in situations in which they actively participate by responding and creating conversations and discussions (Atherley, 2011; Evans & McKee).

Online brand engagement as defined in this study can be considered as head, heart, and hands. The “head” represents the cognitive dimension of engagement that is thoughtful and process oriented. The “heart” represents the affective part of engagement that is emotionally driven. And, the “hands” represent the behavioral aspect of online brand engagement that is the act of participating.

Loyalty towards a brand is considered as an outcome of engagement, as previous research has found a positive relationship and link between the engagement and loyalty (Algesheimer, Dholakia, & Herrmann, 2005; Bagozzi & Dholakia, 2006; Shang et al., 2006). In addition, loyalty is considered to be a key factor in achieving company success and long-term sustainability (Casalo et al., 2007; Flavian et al.)

Bowden (2009) states that customer engagement is the superior predictor of customer loyalty compared to other more traditional marketing tools. Lastly, Cheung et al. (2011) studies support that a customer, who is willing to invest physically, cognitively and emotionally into an online platform will also have a higher intention to recommend and spread the news through word-of-mouth communicative process.

Drivers Brand Engagement Outcomes

,

Figure 1: Conceptual model Brand post

characteristics: Interactivity vs Vividness Content type: Information

(10)

2.1 Literature review

2.1.1 Customer brand engagement

Customer engagement is a key term and is in the core of marketing strategy in the online and offline environment. As it is usually happened, important terms are discussed by different perspectives and it is possible to be analyzed in various ways.

Customer engagement is related to the engagement of customers with one or more company or brand (McEwen 2004). This is a relationship between the brand and the customer, indicating emotional and rational bond that has been developed towards a company. Therefore, brand engagement is supported to include feelings of confidence, pride and passion towards a brand. Another definition of engagement comes from Anderson who supports that customer engagement is a long process, which encourages customer loyalty and promotion by generating word-of-mouth actions- resulting in change of statements (Anderson 2006). Precisely, customer engagement has five different stages on which consumers pass through as they interact with a particular brand (McEwen 2004). This process is well-known as customer engagement cycle or customer journey (McEwen 2004) :

1) Awareness, 2) Consideration, 3) Inquiry, 4) Purchase, 5) Retention. 6) Satisfaction

(11)

2.1.2 Online brand engagement

Online brand engagement is qualitatively different from offline engagement as the nature of the customer’s interactions with a brand, company and other customers are different on the internet (Chak 2003). Verhoef et al. (2010) support that customer engagement includes multiple online-behaviors as blogging, customer ratings and e-word of mouth. Furthermore, they state that customer engagement is affected by firm initiatives, customer characteristics, and environmental/contextual factors (Verhoef et al., 2010).

According to Sashi (2012) customer engagement is a topic that has had an emerging interest during the last years, based on the development of Internet and the new tools that have emerged with it – Web 2.0 (Sashi, 2012). Due to the increasingly networking society customers can easily interact with other customers and this non-transactional customer behavior has become more important for companies when developing their strategies (Verhoef et al., 2010). Hollebeek et al. (2014) and van Doorn et al. (2010) mention how important the understanding of online brand engagement is in the social media platforms as well as recognizing the opportunities for brand to extract value from their followers. Van Doorn (2010) supports also, that it is important for companies to fully understand the impact of customer brand engagement since the digital world involves a broad audience with immediacy breadth.

To be more precise, social networks are communities where people are able to communicate and to socialize in ways that cannot be replicated by any offline interactive medium. Online brand engagement activity aims to create, shape and influence customers behavior differently compared to the offline one-way marketing communication. Consumers can download, read, watch, or listen to content provided by a company (Evans & McKee, 2010). Consumers can also sort, filter, rate, or review a company’s content (Evans & McKee). Moreover, consumers are able to comment, respond, provide feedback, and give opinions to brands’ posts and other consumers’ posts. Gummerus et al. (2012) further state that social media could provide customers with platforms where they can co-create value with companies and engage in behaviors like participating in online discussions, search for information, and commenting. Lastly, online brand engagement is considered as a cognitive and affective obligation to an alive relationship with the brand as personified by the website or other computer-mediated entities designed to communicate brand value (Mollen and Wilson, 2010). It is characterized by the features of strong, live and sustainable cognitive processing and the satisfying instrumental value (utility and relevance) and experiential value (Mollen and Wilson, 2010). Cognitive engagement demands high levels of concentration in searching, interpreting, analyzing, and summarizing information. Furthermore, online brand engagement requires affective feelings, which include emotional bonding and connection with the brand, products, and other users that lead to overall satisfaction. Lastly, consumers must invest themselves within the online vehicle by participating through sharing, conversing, and co-creating with the brand and/or other users (behavior).

2.1.3 Brand Loyalty

(12)

customer behavior (Chegini, 2010). Furthermore, researches support that loyal customers to a brand could reduce marketing costs regarding the efforts of approaching new customers (Kotler, Bowen, & Makens, 1998). Additionally, companies’ efforts are focused on enhancing customer loyalty, as customer loyalty results in favorable behaviors, such as customers’ repurchasing, positive word-of-mouth to nudge the competitors’ customer base, and cross-selling (Verhoef et al., 2002; Hur et al., 2010; Stokburger-Sauer, 2010).

Woisetschlager et al. (2008), state that brand loyalty is a deep commitment or attachment to a brand or service, or the desire to buy a certain product or service in favor of the equivalent from competing brands. As repeated purchases may also indicate the temporary acceptance of a brand, a distinction has to be made between attitudinal loyalty and behavioral loyalty (Shang, et al., 2006). Attitudinal loyalty represents a systematically favorable expression of preference for the brand, or a reflection of the emotional attachment that consumers have for brands (Morgan, 1999; Dick & Basu, 1994). On the other hand, behavioral loyalty is considered as the loyalty status of a given consumer from an observation of repeated purchasing behavior (Morgan, 1999; Kahn). Moreover, Algesheimer et al. (2004) supported that a stronger identification and commitment to the brand leads to a stronger member loyalty and intentions to recommend the brand or service to others (word-of-mouth). Lastly, Brodie et al. (2013) found out that participants express loyalty towards a brand and its community by revealing satisfaction and recommending the brand to others.

2.1.4 E- Word Of Mouth

Word of mouth is the natural conversation between communicators and the shared information is passed from one person to another in order to be spread and reach a large amount of people. In marketing, word of mouth refers to the ability of making people talk about a company (Sernovitz, Kawasaki and Godin, 2012).

Word of mouth is considered as an important and strong marketing tool for companies by which information can be easily spread and brands can approach large populations, hence it is more likely to influence consumer decisions through purchase journey ( Bansal & Voyer, 2000), and at the same time affect consumer expectations (Anderson& Salisbury, 2003). Word of mouth is consisted of ways to communicate positive or negative products or services related assessments between consumers (Mazzarol et al. 2007)

Word-of-mouth can be defined as an interpersonal way of communication between consumers about products and services and is a key driver in influencing consumer attitudes and behaviors (Richins, 1984)

The concept of WOM is considered as highly useful tool for marketeers, in order to spread the marketeers’ message and brands’s news, vision and strategy. A positive WOM communication can influence a huge amount of consumers and can result in the spread of the marketing message by the customers as they are satisfied by a brand.

(13)

interpersonal way of communication and it can be hosted through electronic channels in terms of reviews and recommendation sites, social media, online communities, blogs, and chat rooms. Social media users are able to connect and share their enthusiasm about their favorite brands with their friends, personal contacts and other acquaintances (Cheung and Lee, 2012). In addition, the rise of social media, improved the strength and distribution of positive or negative statements, as consumers can express easily their opinions and share them with other Internet users anytime, anywhere (Dellarocas, 2003). However, e-WOM is significantly different from traditional WOM. Despite the fact that e-WoM is less personal than traditional word of mouth, it is more powerful because it has an immediate effect, can reach higher amount of people, is credible, and is publicly available (Hennig-Thurau et al., 2004). Hoskins and Smith (2007) assume that credibility and persuasiveness of a message determine whether or not the message will pass along to others. Social media users who seek information, support that information related to products and services which is provided by other consumers is more valuable than when this information is provided by marketers (Greene, 2009).

Previous research has shown that e-wom may have higher credibility amongst customers than marketeers-based sources of information (Bickart & Schindler, 2001).

2.2 Hypotheses

2.2.1 Content type: Information vs Entertainment and brand engagement

Looking for information is a highly important driver for a consumer to use social media platforms. Information quality refers to a consumer’s assessment of the information presented on a website based on accuracy, relevance, helpfulness, currency (being up-to-date), and unbiasness (Cao et al., 2005; Ou & Sia, 2010; Zhang & von Dran, 2000). Information is considered to be a key website feature that influences consumer behavior. Past research (Cao et al.; Day, 1997; Huizingh, 2000; Iyer, 2001;; Ou & Sia; Zhang & von Dran) has found that information is highly important to the development of a company’s website to draw in and continue to attract online customers. Interacting with information is described as the “process people use in interacting with content” (Toms, 2002). If a brand post contains information about the brand or product, then the brand fans’ motivations to participate or consume the content of the brand post are realized.

Information content is considered to influence the cognitive (seeking, attending, interpreting, and critiquing information), affective (excitement toward the information and ultimately the brand and satisfaction), and behavioral (interacting with the information) components of online brand engagement. Based on these, the following hypotheses are derived:

H1A: Perceived information quality on a Twitter brand’s account that a consumer has “followed” positively predicts cognitive online brand engagement.

H1B: Perceived information quality on a Twitter brand’s account that a consumer has “followed” positively predicts affective online brand engagement.

(14)

The entertaining features of a social networking site is also an important factor for using it (Cheung et al., 2011, Dholakia et al., 2004, Lin and Hsi-Peng, 2011 and Park et al., 2009). In addition, websites that are fun, interesting, exciting and entertaining, has been found to have a positive effect on a customer’s evaluation of a company’s website (Chang et al., 2005; Ou & Sia, 2010).Entertainment leads people to consume, create or contribute to brand-related content online (Muntinga, Moorman, and Smit 2011). Past research has found that consumers who view their time on a company’s website as enjoyable also experience emotional involvement with the brand, resulting in positive brand bonding (Zhang & von Dran) and can increase overall satisfaction (Eighmey, 1997). Furthermore, online experiences that are considered as entertaining can make users feel cognitively involved with a brand (Zhang & von Dran) which can promote branding learning and concentration (Watson et al., 1998). Enjoyable experiences can also increase online participation, based on the idea that consumers are more willing to stay on the site and provide their input (Watson et al.; Zhang & von Dran). Entertaining brand posts influence the cognitive (brand learning and concentration), affective (emotionally involvement, brand bonding and satisfaction), and participative (providing input) components found in online consumer engagement. Hence, the following hypotheses are derived:

H2A: Perceived entertainment on a Twitter brand’s account that a consumer has “followed” positively predicts cognitive online brand engagement.

H2B: Perceived entertainment on a Twitter brand’s account that a consumer has “followed” positively predicts affective online brand engagement.

H2C: Perceived entertainment on a Twitter brand’s account that a consumer has “followed” positively predicts behavioral online brand engagement

2.2.2 Brand post characteristics: Interactivity vs Vividness and online consumer

engagement

According to Mollen and Wilson (2010) interactivity attempts to capture whether a user-in this case a follower perceives that the communication is two-way, controllable, and responsive when the user is present in a mediated environment. In addition, a definition of interactivity is given by Liu and Shrum, and is considered as the situation in which two or more parties communicate with each other on the communication medium, and on the messages and the degree to which such influences are synchronized (Liu and Shrum 2002). Moreover, O’Brien and Toms (2008) support that engagement involves also interactivity , as users are bonded and engaged with a brand when feedback is present and users have the feeling that interaction is under their control (Schneiderman & Plaisant, 2005). There is a great variety of interactivity on brand related posts such as a brand post with a direct link to a website is more interactive (Fortin and Dholakia 2005) as brand followers are able to click on that link and interact with the brand. Lastly, a post which includes a questions, is highly interactive as it enables followers to post an answer (Liu and Shrum 2002).

Interactivity can result in the three online brand engagement components, which are cognition, affect and behavior. Hence, the following hypotheses are proposed:

(15)

H3B: Perceived interactivity on a Twitter brand’s account that a consumer has “followed” positively predicts affective online brand engagement.

H3C: Perceived interactivity on a Twitter brand’s account that a consumer has “followed” positively predicts behavioral online brand engagement.

Vividness reflects the richness of a brand post, it is the extent to which a brand post stimulates the different senses (Steuer 1992). Vividness can be expressed by using animations, (contrasting) colors, or pictures (Drèze and Hussherr, 2003, Fortin and Dholakia, 2005, Goldfarb and Tucker, 2011 and Goodrich, 2011). The degree of vividness can be different in the way that it stimulates multiple senses (Coyle and Thorson 2001). A video is thought to be more vivid than a picture because it includes not only sight, but also hearing.

Vividness can have effect on the online consumer engagement-cognitive, affective and behavioral components leading to the following hypotheses:

H4A: Perceived vividness on a Twitter brand’s account that a consumer has “followed” positively predicts cognitive online brand engagement.

H4B: Perceived vividness on a Twitter brand’s account that a consumer has “followed” positively predicts affective online brand engagement.

H4C: Perceived vividness on a Twitter brand’s account that a consumer has “followed” positively predicts behavioral online brand engagement.

2.2.3 Online brand engagement and outcomes

In this study, two customer based outcomes are chosen: brand loyalty and word of mouth, which refers to brand recommendation.

According to Bowden (2009) customer engagement is the most successful and key way to predict customer loyalty compared to other more traditional marketing constructs.

Based on previous studies, the strength and the effect of brand loyalty is higher in cases in which consumers are psychologically or cognitively bonded with a brand (Tyebjee, 1977). Consumers who are cognitively engaged with a brand, obtain deeper brand related information and the learnings are improved and result in promoting customer loyalty (Shang et al., 2006). Furthermore, Dick and Basu (1994) have found that there is a positive relationship between the loyalty and the affective engagement. A consumer that poses positive feelings about a brand and is affectively engaged, is more loyal toward the brand. Jang et al. (2008) stated that commitment is in the core of term loyalty, as commitment may play an important role in posing positive attitudes for company, which can lead to loyalty toward the mentioned previously. Additionally, Matzler et al. (2008) and Shang et al. (2006) supported that there is indeed a positive relationship between brand affect and loyalty found and affect can have a positive effect on brand loyalty.

(16)

positive relationship between behavior and loyalty (Bagozzi & Dholakia, 2006; Muñiz & O’Guinn) and support the notion that if the consumers participate in an online platform which belongs to a brand, the connection with the brand is enhanced and result in brand commitment and eventually in loyalty.

Word of mouth communication is considered to be one of the consequences of online engagement towards a brand. Kumar et al. (2010) by using a customer oriented approach, revealed that the true core of customer engagement involves four aspects: customer purchasing behavior, customer referral behavior, customer influencer behavior through customers influence on other existing or prospect customers, and customer knowledge behavior, which applies through feedback to the company.

Customer loyalty and word-of mouth have established grounds as potential engagement consequences in the literature, as it is proved based in the past studies. Hence, two new hypotheses are formulated:

H5A: Online brand engagement with a brand “followed” in Twitter predicts positively loyalty by the follower.

(17)

3. METHODOLOGY

This chapter provides a detailed description of the methodology, which will be used to investigate the hypothesized relationships, presented in the previous chapter. In order to test the theoretical model proposed before, primary data will be collected using an online survey. Surveys collect data from a population of respondents as an attempt to understand behavior in a variety of contexts (Negrine & Newbold, 1998). According to Bostrom (1998), surveys are an appropriate research approach to use to understand consumer behavior (Negrine & Newbold). Furthermore, surveys make use of a formal set of questions to “estimate the distribution of characteristics in a sample (Dillman, 2000, p. 9) and are able to provide empirical data to lend support or negate hypotheses or propositions (Negrine & Newbold). Finally, surveys are administered to “describe, find, or validate” specific proposed relationships (Reagan, 2006, p. 92), as the relationships which are presented in the previous section.

The survey was administered online using Qualtrics technology. Qualtrics is a web-based survey tool that offers to researchers the opportunity to create comprehensive surveys for academic purposes (Qualtrics, 2011). Qualtrics generates a unique URL after a researcher has built a survey using software that can be easily distributed and accessed online (Qualtrics). The questionnaire will be launched online and distributed using various web tools such as email and social networking platforms.

The questionnaire is comprised by several parts regarding and contains measurement items for the following variables: perceived information quality, perceived entertainment, perceived interactivity, perceived vividness, online brand engagement cognition, online brand engagement affect, online brand engagement behavior, loyalty, and e-word of mouth.

The first question is a screening question. Lastly, the questionnaire contains demographic information including gender, age, ethnicity, education, as this information could provide a better understanding of the respondents taking the survey. According to Negrine and Newbold (1998), the demographic questions – gender, age, ethnicity, education, considered as standard questions on consumer behavior surveys to “seek out basic socio-demographic data”.

3.1 Variables-measures

According to previous research quantitative methods and questionnaires, have been used in order to measure customer engagement on social media (Gummerus et al., 2012; Men & Tsai, 2014; Kabadayi & Price, 2013; Bunker et al., 2013; Weman, 2011).

Based on previous studies (Cao et al., 2005; Ou & Sia, 2010; Zhang & von Dran, 2000), information quality attempts to examine the consumer’s evaluations regarding product and company information on the brand’s official Twitter account that they “follow”. In this study, information quality is measured with five items created by Cao et al.’s and Zhang and von Dran’s. All five items are measured using a 7-point Likert-type scale ranging from “strongly disagree” (1) to “strongly agree” (7).

(18)

specific features on a brand’s Twitter account that they “follow” and the extent to which followers perceive these features as fun, enjoyable and attractive. Entertainment is measured with five items, which were modified from the instrument developed by Cao et al. and Koufaris (2002). All five items are measured using a 7-point Likert-type scale ranging from “strongly disagree” (1) to “strongly agree” (7).”

Following previous definitions regarding the interactivity (Lee, 2005; Mollen & Wilson, 2010), perceived interactivity expresses the extent to which a consumer perceives the interaction or communication to be two-way, controllable, and responsive on a brand’s Twitter account that he/she “follows”. Perceived interactivity is measured by six items which were created by Lee (2005) and Cyr et al. (2009).These measurements were modified in order to be applicable for this study. All items are measured using a 7-point Likert-type scale ranging from “strongly disagree” (1) to “strongly agree” (7).

Twitter provides users with the choice of sharing (1) status, (2) photo, (3) video, and (4) link. These kinds of media are part of the content’s richness and are commonly known as vividness of online content (Daft and Lengel 1986). This dimension of online posts is related to the breadth and depth of the message. Precisely, breadth represents the features that awake followers’ senses such as cues with vivid colors, graphics. On the other hand depth refers to the quality and resolution of the presentation- following Steuer (1992). Perceived vividness is measured by four self-constructed items, which measured using a 7-Point Likert-type scale are ranging from “strongly disagree” (1) to “strongly agree” (7).”

In the case of this study, the construct of customer brand engagement on online social media platforms has been divided into three components – behavioral, emotional and cognitive. Cognition represents the degree to which a consumer is present within the mediated space and he/she is able to process and focus on the perceived information, which is presented on a brand’s Twitter account that he/she “follows” and at the same he/she forgets about the mediated space (Cheung et al., 2011). Furthermore, cognition attempts to measure the extent to which the consumer managed to deepen his/her knowledge about the company, brand, or product on the company’s Twitter account that he/she follows. In this study, the cognitive engagement scale of Cheung et al. (2011) is used, and the construct is measured with six items. All items are measured using a 7-point Likert-type scale, anchored by 1=”Strongly disagree”, 7=”Strongly agree”.

Affect is another online brand engagement component that is researched in this study. Affect attempts to catch how emotionally present and bonded a consumer is with the shared posts and seeks for further information regarding pleasure, satisfaction with the followed brand. Affect was also measured, using Cheung et al. (2011) scale. The construct is measured using six items. All items are measured using a 7-point Likert-type scale ranging from “strongly disagree” (1) to “strongly agree” (7).

(19)

on Twitter. The items are measured using a 7-point Likert-type scale ranging from “strongly disagree” (1) to “strongly agree” (7).

The consequences of customer brand engagement on online social media platform of Twitter have been measured in terms of behavioral brand loyalty and word-of-mouth. The scale for behavioral brand loyalty consisted of four items regarding future purchase intentions (Chaudhuri & Holbrook, 2001). Word-of-mouth, which can also be defined as the intention to recommend the brand to others, has been measured with three items suggested by Zeithaml, Berry & Parasuraman (1996).

Lastly, demographic variables are measured. The following is the list of demographic variables along with the operationalization definitions for each as defined by Yan (2005):

Gender: Female/Male

Age: Age of the respondent given in years Ethnicity: Ethnic group the respondent belongs to Education: Highest level of formal education completed

In order to overcome any form of discomfort of confusion among respondents regarding the questionnaire, it will be made clear that they will be kept anonymous, that results will exclusively be used for this study, the questionnaire will be kept as succinct as possible and the purpose of the study will be disclosed after completion on demand.

Overview measurement variables-Table 1

Measurement construct Literature Scale

Information quality Cao et al., 2005; Ou & Sia, 2010; Zhang & von Dran, 2000

7 point

Entertainment Cao et al. and Koufaris

(2002)

7 point

Interactivity Johnson, Bruner, and Kumar

(2006)

7 point

Vividness Self-constructed 7 point

Online brand engagement

Cognition Cheung et al. (2011), 7 point

Affect Cheung et al. (2011) 7 point

Behavioral Casalo et al. (2010) 7 point

Outcomes

Brand loyalty Chaudhuri & Holbrook, 2001 7 point

E-wom by Zeithaml, Berry &

Parasuraman (1996).

(20)

4. RESULTS: analysis

A total of 276 online surveys were completed during the two-week data collection period. Out of the 276 questionnaires received, 43 were eliminated as the respondents did not complete the survey, 105 were eliminated because the participants did not obtain a Twitter account. Additionally, 38 more surveys were eliminated because the participants, do not follow any brand on Twitter. As a result, 132 questionnaires were usable for analysis.

Hair et al. state that to analyze whether the sample could lead to reliable results, there must be an “absolute” minimum of 50 participants. Given this, the 132 participants in this study is within the range of appropriate sample size for analysis.

4.1 Respondents’ Characteristics

Respondents ranged from 18 years old to 54 years old, with the majority (67%) between the ages ranges 25-34. 60% of the respondents were female. Additionally, the sample population was well-educated. About 50% had a bachelor’s degree and 33% had a master’s or doctoral degree.

Demographic characteristics of survey respondents-Table 2

Characteristics Sample size

Age 18-24 25-34 35-44 45-54 31% 22% 10% 5% Gender Male Female 60% 40% Education

High school diploma Some college, no degree Associate’s degree Bachelor’s degree

Master’s or Doctoral degree

(21)

4.2 Twitter usage Statistics

In regard to the length of time respondents have been Twitter users, responses ranged from 3 months to eight years with a mean of four years. On average, respondents spend nearly eight hours (M = 7.75) a week on Twitter. Additionally, respondents rated their Twitter experience level as intermediate, with a mean of 4.5 on a scale from one to 7(one = lowest, 7 = highest). Furthermore, on average, respondents “followed” approximately 20 companies on Twitter in a variety of industries. Nearly 57% “followed” apparel/accessories; 14% “followed” automotive; about 56% “followed” entertainment, close to 30% “followed” personal care and food/beverage; 37% “followed” technology; and 10% “followed” other types of industries on Twitter including media-news, design and travel.

Twitter Information-Table 3

Twitter Information Sample size Length of Twitter user

Years (Range: 3months – 8 years)

4 years

Hours Spent on Twitter

Per week (Range: 0 – 42 hours)

7.75 7.75

Twitter Experience Level

On a scale from 1 to 7 (1=lowest; 7=highest) 4.5

Number of brands “followed” on Twitter 20

Types of brands “followed” on Twitter

Apparel and Accessories Automobile

Entertainment Food/beverage

(22)

4.3 Factor Analysis

Exploratory factor analyses using Varimax rotation were conducted on each of the multiple-item scales, including the exogenous constructs ( information quality, entertainment, interactivity and vividness) and endogenous constructs (cognitive engagement, affective engagement, behavioral engagement and the outcomes of this engagement-loyalty and e-WOM) as a way to refine the measures in the study. Exploratory factor analysis was conducted prior to testing the full model to identify items with poor psychometric properties and to purify the measurement model for future testing (Anderson & Gerbing, 1988). A priori designations for each of the proposed antecedents of engagement (information quality, entertainment, interactivity and vividness), for loyalty, and for e-wom were employed in the analyses, given that they have been confirmed in previous studies (Mollen & Wilson, 2009; O’Brien & Toms, 2008; Shang et al., 2006; Shukla, 2009 ).

Eigenvalues larger than one and scree plots were analyzed to assist in determining the number of factors for each construct. Additionally, the strength of factor loadings as well as face validity were analyzed to further assist in determining the items to be included for each exogenous and endogenous variables.

The coefficient reliability analysis revealed that all the scales consisting of more than two items exceeded the recommended Cronbach‟s alpha benchmark of 0.70 (Nunnally, 1978).

4.4 Exploratory Factor Analysis of Exogenous Constructs

4.4.1 Perceived Information Quality

Exploratory factor analysis for this variable resulted in a one-factor solution with all of the original five items in the scale. Factor loadings of these items ranged from .737 to .821. Explained variance for this factor was 61% with a Cronbach’s alpha of .845. Furthermore, the researcher calculated a composite score for this factor using the average score of the five items.

4.4.2 Perceived Entertainment

Exploratory factor analysis revealed a one-factor solution for perceived entertainment that contained the original five items in the scale. Factor loadings ranged from .719 to .852. Internal reliability based on Cronbach’s alpha was .858, and the variance explained by the five items was 63.92%. Moreover, a composite score was calculated for the items of this factor.

4.4.3 Perceived Vividness

(23)

4.4.4 Perceived Interactivity

Factor analysis produced a one-factor solution for perceived interactivity. Factor loadings for perceived interactivity ranged from .598 to .885, with 57.27% of variance explained. Internal reliability based on the Cronbach’s alpha was .847 but according to the analysis, the reliability is improved when the first item (Customers share experiences about products with other customers on the company’s account I “follow” on Twitter) deleted, leading to .851.

4.4.5 Online Brand Engagement (cognitive-affective-behavioral)

As the online brand engagement concept is relatively new in the marketing research literature, scales have to be confirmed. The exploratory factor analysis approach provided the opportunity to analyze all of the proposed engagement components (cognition, affect, and behavior) at the same time to determine the factor structure.

By using the exploratory factor analysis for the construct of online brand engagement and by using all of the proposed engagement components (cognition, affect, and behavior) at the same time, the analysis led to a two factor solution. Two factors emerged from the exploratory factor analysis instead of the three factors as expected (cognition, affect, and behavior). Factor one consisted of five items related to cognitive measures, and factor two consisted of seven behavior and affect measures (referred to hereafter as Behaff). The total variance explained was 73% and the factor loadings ranged between .552 and .832 for factor 1 and .520 to .819 for factor 2. The reliability based on the Cronbach’s alpha test was .944.

Cognition: The one-factor solution for cognition included all the items related to cognitive engagement (mentally involved, learning and understanding about the vision and the goals of the brand). Internal reliability for the cognition variable, based on the Cronbach’s alpha, was .85. 62.3% percent of variance was explained by the items included with factor loadings which range from .64 to .81. In addition, a composite score was calculated for this factor using the average score of the items.

Behaff: The factor analysis of the questions regarding the dimensions of the online brand engagement resulted in two factors instead of three-hypothesized and presented in the first chapter (cognition, affect and behavior). The second factor of the analysis -Behaff- is represented by a combination of all the items, which are relevant to the behavioral dimension and refer to types of behavior, consumers engage in on Twitter’s brand account they follow and items related to the affect dimension of online brand engagement, regarding enjoyment, fun and pleasure. Variance explained by the factor was 63% with factor loadings ranging from .65 to .78. Internal reliability for the factor was .82 based on the Cronbach’s alpha. Finally, a composite score was calculated for this factor.

4.4.6 Expected outcomes: Loyalty

(24)

companies I “follow” on Twitter- was deleted because of low-negative loading),

characterizing the respondents’ loyalty to the brand they “follow” on Twitter. Factor loadings ranged from .605 to .849, and internal reliability was .844 (instead of .558 before the item was deleted), based on the Cronbach’s alpha. 76.61% of variance was explained by this factor.

4.4.7 Expected outcomes: E-word of mouth

The factor analysis of e-wom showed that the variable has a one-factor solution. The e-wom scale, which consisted of three items, attempts to capture the likelihood that the respondents would recommend to friends the brand they “follow” on Twitter. The Cronbach’s alpha showed that the internal reliability was .817 with 73.46% of variance explained. Factor loadings for the four-item scale ranged from .785 to .925.

4.5 Multicollinearity test

Finally, the variables in the model should be tested in order to examine if they are correlated, i.e. if there is a relation between different predictor variables. A multicollinearity test was conducted with the composite scores calculated for every factor in order to determine how correlated the variables were (Hair et al., 1992). To detect multicollinearity, variance inflation factor (VIF) values can be checked, which should be smaller than 5, following Leeflang et al. (2015). Using e-wom as the dependent variable and the proposed antecedents of online brand engagement (information, entertainment, interactivity and vividness), online brand engagement (cognition, affect and behavior) and loyalty as independent variables, the variance inflation factor (VIF) for relevant regression models ranged between 1.2 and 2.4. Based on the fact that the VIF values did not exceed 10, it could be assumed that multicollinearity cannot harm the further analysis.

4.6 Revised model

(25)

Table of correlation coefficients-Table 4 Mean (S.D) Alpha 1 2 3 4 5 6 7 8 1.Online Brand Engagement: Cognitive 5.35 (1.14) .85 1.00 2.Online Brand Engagement: Behaff 5.24 (1.19) .82 .35 1.00 3.Brand Loyalty 5.28 (1.17) .84 .56 .34 1.00 4.E-wom 4.75 (1.21) .82 .42 .23 .75 1.00 5.Perceived Information quality 3.02 (1.43) .84 .67 .35 .42 .32 1.00 6.Perceived entertainment 5.53 (1.16) .86 .81 .43 .47 .38 .74 1.00 7.Perceived Interactivity 5.78 (1.16) .85 .42 .50 .30 .21 .39 .41 1.00 8.Perceived Vividness 5.32 (1.14) .79 .39 .49 .32 .22 .43 .45 .68 1.00

4.7 Test of hypothesis-Multiple regression analysis

(26)

Sets of multiple regression analysis-Table 5

Independent Variable(s) Dependent

Interactivity Cognitive brand engagement

Vividness Cognitive brand engagement

Information Cognitive brand engagement

Entertainment Cognitive brand engagement

Interactivity Behaff brand engagement

Vividness Behaff brand engagement

Information Behaff brand engagement

Entertainment Behaff brand engagement

Online brand engagement Loyalty

Online brand engagement E-wom

The first regression analysis tested the effect of the drivers-information, entertainment, interactivity and vividness on the cognitive component of online brand engagement. The model was overall significant, with F=9.423 for p=.000. In the following table, the results are presented.

Multiple regression 1-Table 6

Independent variables Parameter estimate Standard error P-value Constant 1.895 0.453 0.000* Information quality Xcognitive brand engagement 0.323 0.068 0.000* Entertainment X cognitive brand engagement 0.238 0.057 0.03* Interactivity X cognitive brand engagement 0.193 0.093 0.000* Vividness Xcognitive brand engagement 0.185 0.091 0.000*

(27)

H1A: Perceived information quality on a Twitter brand’s account that a consumer has “followed” positively predicts cognitive online brand engagement.

The hypothesis H1A, which states that the perceived information quality on a Twitter brand account positively predicts cognitive brand engagement was supported. Results based on the revised research model indicated a significant result (γ=.323, t = 4.34 , sig.=.000) between information quality and cognitive engagement.

Based on this, it could be assumed that consumers who perceive a brand’s Twitter account to possess information are likely to become cognitively engaged. This finding is also supported by prior research (O’Brien & Toms, 2008).

H2A: Perceived entertainment on a Twitter brand’s account that a consumer has “followed” positively predicts cognitive online brand engagement.

Hypothesis H2A, which states that perceived entertainment of a brand’s Twitter account positively predicts cognitive online brand engagement, was supported. The model revealed a significant result between entertainment and cognitive brand engagement (γ=.238, t = 5.05, sig.=.000).

The results suggested that an entertaining, funny and enjoyable experience on a brand’s Twitter account can result in the cognitive connect of the consumer with the brand through brand learning, which aligns with past research on consumers’ assessment of entertainment of a company’s website (Zhang & von Dran, 2000; Watson et al., 1998).

H3A: Perceived interactivity on a Twitter brand’s account that a consumer has “followed” positively predicts cognitive online brand engagement.

The hypothesis that perceived interactivity predicts positively cognitive online brand engagement (Hypothesis 3A) was supported. Results confirmed that there is a positive relationship between interactivity and cognitive online brand engagement (γ=.193, t = 4.5, sig=.000).

This positive relationship confirms past research (Mollen & Wilson, 2010), which suggests that consumers who process and interpret the brand’s Twitter account in their own controlled environment, may result in online brand engagement.

H4A: Perceived vividness on a Twitter brand’s account that a consumer has “followed” positively predicts cognitive online brand engagement.

Hypothesis 4A, that suggested that perceived vividness of a Twitter brand’s account has a

positive impact and predicts the cognitive online brand engagement, was supported. The research model revealed a significant result (γ=.185, t = 4.069, sig.=.000). The results suggested a vivid and rich experience can help consumers to cognitively related to the brand-which concurs with past research (Fortin and Dholakia, 2005, Goldfarb and Tucker, 2011 and Goodrich, 2011).

(28)

Multiple regression 2-Table 7 Independent variables Parameter estimate Standard error P-value Constant 2.214 0.502 0.000* Information quality Xbehaff brand engagement 0.193 0.072 0.000* Entertainment X Behaff brand engagement 0.209 0.066 0.000* Interactivity X Behaff brand engagement 0.185 0.087 0.015* Vividness X Behaff brand engagement 0.191 0.069 0.008* regression model - N =132

R2 = .347, , Adj. R2 = .278, F-stat = 7.842, P-value = .000 * Results are significant at p < .1

H1BC: Perceived information quality on a Twitter brand’s account that a consumer has “followed” positively predicts affective and behavioral online brand engagement. The revised hypothesis H1BC, which supports that the perceived information quality on a Twitter brand account positively predicts affective and behavioral brand engagement was supported.

Results based on the revised model are significant(γ=.193, t = 2.985, sig=.03), implying that there is a positive effect on the level of behavioral and affective engagement of customers and the perceived information quality that the Twitter account provides to the followers. Given the results that are found, it could be assumed that when a consumer believes and perceives that the brand’s Twitter account includes content that is exciting and interesting, then it may be more likely for the consumer to become affectively engaged with the brand. This result aligns with O’Brien & Toms’ (2008) study, which suggests that information which evokes a sense of excitement, pleasure and satisfaction may have an effect on consumers’ engagement.

H2BC: Perceived entertainment on a Twitter brand’s account that a consumer has “followed” positively predicts affective and behavioral online brand engagement.

(29)

Based on this, Hypothesis H2BC was supported. When a brand’s Twitter account is perceived to be entertaining, enjoyable, funny and pleasurable, it may be more likely for the consumers to be engaged via participation and emotionally related behaviors. This finding is also supported by previous studies of Watson et al. (1998) and Zhang & von Dran (2000). Moreover, this result proposed that websites with aesthetically appealing features have a positive impact on consumers’ engagement and bonding with the brand, which also concurs with past research (Aboulafia & Bannon, 2004; Jennings, 2000; O’Brien & Toms, 2008).

H3BC: Perceived interactivity on a Twitter brand’s account that a consumer has “followed” positively predicts affective online brand engagement.

The prediction that perceived interactivity of a brand’s Twitter account influences behavioral and affective online brand engagement-Hypothesis 3BC was supported, given the fact that the revised research model tested behavioral and affective dimensions as one.

The findings showed that there was a positive path between the hypothesized relationship (γ=.185, t = 2.473, sig=.015). This finding aligns with past research (Cyr et al., 2009; Lee, 2005; Mollen & Wilson, 2010), which suggested that a website’s interactive capabilities can positively influence a user’s participation behavior. Furthermore, when a consumer interpret the interactive features as satisfying to him/her, it more likely to lead to online brand engagement.

H4BC: Perceived vividness on a Twitter brand’s account that a consumer has “followed” positively predicts affective and behavioral online brand engagement.

The revised hypothesis H4BC, which supports that the perceived vividness on a Twitter brand’s account positively predicts affective and behavioral brand engagement was supported.

Results based on the revised model are significant(γ=.191, t = 2.708, sig=.008), implying that there is a positive effect on the level of behavioral and affective engagement of customers and the vivid elements of a brand’s Twitter account. It could be assumed, that when consumers interpret the content of a Twitter brand’s account as vivid and rich, it is more likely to become engaged and moves theme towards participation. The results align to previous studies (De Vries., Gensler, Leeflang, 2012).

H5A: Online brand engagement with a brand “followed” in Twitter predicts positively loyalty by the follower.

As online brand engagement consists of three dimensions- cognition, affect and behavior, the model tested separately the impact on brand loyalty. Based on factor analysis, that showed that behavioral and affective components should be measured together, the researcher tested the main effect of cognitive online brand engagement on brand loyalty and the impact of affective and behavioral online brand engagement to brand loyalty, simultaneously.

Hypothesis 5A, which examined the positive relationship between predicted cognitive online brand engagement and loyalty, was supported. The results indicated a significant path between cognition and loyalty (F = 60.16, t = 7.756, sig.=.000).

(30)

Additionally, the updated research model showed a significant path (F= 58.23, t = 6.834, sig=.000) for the hypothesized relationship between behavioral and affective online brand engagement and brand loyalty, was supported. It could be assumed that when consumers participated and being emotionally engaged with a brand’s Twitter account, it is more likely for the consumers to feel stronger brand loyalty to the brand. This notion is congruent with several past studies (Algesheimer et al., 2005; Bagozzi & Dholakia, 2006; Muñiz & O’Guinn, 2001; Chaudhuri & Holbrook, 2001; Dick & Basu, 1994; Jang et al., 2008).

Multiple regression 3-Table 8

Independent variables Parameter estimate Standard error P-value Constant 2.389 0.156 0.000* Cognitive brand engagement Xloyalty 0.378 0.032 0.000* Behaff brand engagement Xloyalty 0.374 0.028 0.000 Regression model - N =132

R2 = .358, Adj. R2 = .276, stat = 60.16, P-value = .000, R2=.326, Adj. R2=.239, F-stat=58.23, P-value=0.000

* Results are significant at p < .1

H5B: Online brand engagement with a brand “followed” in Twitter predicts positively e-WOM by the follower

Given the fact that factor analysis for online brand engagement resulted in a two factor solution, the researcher tested the effect of cognitive online brand engagement to e-wom and the impact of behavioral and affective online brand engagement to e-wom, separately.

(31)

Multiple regression 4-Table 9 Independent variables Parameter estimate Standard error P-value Constant 1.139 0.215 0.000* Cognitive brand engagement X e-wom 0.178 0.023 0.036* Behaff brand engagement X e-wom 0.137 0.017 0.023* Basic regression - N =132

R2 = .142, Adj. R2 = .187, stat = 4.494, P-value = .036, R2=.132, Adj. R2=.171, F-stat=3.992, P-value=0.023

* Results are significant at p < .1

4.8 Summary of the results

Overall, this paper confirmed all of the proposed relationships, presented in the conceptual model. The revised research model-based on factor analysis, regarding the components of online brand engagement- revealed support for the impact of perceived information quality on cognitive, affective and behavioral online brand engagement (Hypothesis 1A, 1BC), the effect of perceived entertainment on cognitive, affective and behavioral online brand engagement (Hypothesis 2A, 2BC), the impact of perceived interactivity on cognitive, affective and behavioral online brand engagement (Hypothesis 3A, 3BC), the impact of perceived vividness on cognitive, affective and behavioral online brand engagement (Hypothesis 4A, 4BC), the effect of cognitive online brand engagement on brand loyalty (Hypothesis 5A), the effect of behavioral and affective online brand engagement on brand loyalty (Hypothesis 5A), the effect of cognitive online brand engagement on e-wom(Hypothesis 5B) and the impact of behavioral and affective online brand engagement on e-wom (Hypothesis 5B).

(32)

Summary of findings-Table 10

H1A: Perceived information quality on a Twitter brand’s account that a consumer has “followed” positively predicts cognitive online brand engagement.

Supported

H1BC: Perceived information quality on a Twitter brand’s account that a consumer has “followed” positively predicts affective and behavioral online brand engagement.

(Revised based on factor analysis)

Supported

H2A: Perceived entertainment on a Twitter brand’s account that a consumer has

“followed” positively predicts cognitive online brand engagement.

Supported

H2BC: Perceived entertainment on a Twitter brand’s account that a consumer has

“followed” positively predicts affective and behavioral online brand engagement.

Supported

H3A: Perceived interactivity on a Twitter brand’s account that a consumer has

Supported

H3BC: Perceived interactivity on a Twitter brand’s account that a consumer has “followed” positively predicts affective and behavioral online brand engagement.

Supported

H4A: Perceived vividness on a Twitter brand’s account that a consumer has

Supported

H4BC: Perceived interactivity on a Twitter brand’s account that a consumer has “followed” positively predicts affective and

(33)

behavioral online brand engagement.

H5A: Online brand engagement with an account “followed” in Twitter predicts positively loyalty by the follower.

Supported

H5B: Online brand engagement with an account “followed” in Twitter predicts positively e-wom by the follower.