Why Isn't Everyone Doing It? A Comparison of Antecedents to Following Brands on Twitter and Facebook

Abstract

This research compares the antecedents to young adults’ intentions to follow brands on Twitter and Facebook. The three conclusions that emerge from this study are not entirely consistent with planned behavior or technology acceptance models. First, perceived ease of use directly affects users’ intentions to follow brands on Facebook and Twitter. Second, peer pressure is an important factor in the decision to follow brands on Facebook and Twitter. Third, consumers’ attitudes toward following brands on Facebook and Twitter do not directly affect their intentions to follow brands. In addition, following brands on Facebook and Twitter appears to satisfy different user gratifications.

Keywords:

• social media advertising
• Twitter
• Facebook
• Technology acceptance model
• Theory of planned behavior

Social media have revolutionized the role of media in consumers’ lives by disintegrating the fourth wall between media providers and media users and facilitating genuine dialogue. Social media are defined by Boyd and Ellison (2007) as “web-based services that allow individuals to (1) construct a public or semi-public profile within a bounded system, (2) articulate a list of other users with whom they share a connection and, (3) view and traverse their list of connections and those made by others within the system” (211). According to Burst Media (2013), 65% of online adults have at least one personal social media account and 59% use their accounts at least once a day. The growing penetration of social media has spurred advertisers to get involved in the medium. Nearly 90% of advertisers use social media free tools, such as Facebook and Twitter, and 75% of them use paid social media advertising, such as paid ads on Facebook or sponsored blogs. While advertisers have opportunistically jumped on the social media bandwagon, 70% of them have invested less than 10% of their advertising budgets on the new medium (Nielsen Company 2013).

Advertisers’ reluctance to commit significant budgets to social media advertising appears to be attributable to the lack of metrics to assess advertising effectiveness versus other online and offline alternatives (Nielsen Company 2013). It could be argued, however, that the development of “media-agnostic” advertising metrics is greatly complicated by the lack of meaningful measures related to the effects of interactivity. Currently, social media campaigns are measured on the basis of “likes/pins,” “click-throughs,” “views,” and “shares/reposts.” These metrics are analogous to the measures used to assess broadcast media audience size. While measures of site traffic certainly provide a means to compare audience size across media, they will not provide insight regarding the advantages of social media compared to other, traditional media alternatives in terms of message effectiveness. This study investigates why consumers follow brands on Facebook and Twitter, providing insight regarding the factors that are most important to young adults (ages 18 to 34) when forming the intention to follow brands. These insights are crucial to the development of measures that can, ultimately, provide media-agnostic comparisons of advertising effectiveness.

SOCIAL NETWORK SITES

The Interactive Advertising Bureau (IAB) divides social media into three categories: social media sites, blogs, and mobile social media (IAB 2009). Facebook, an online social networking service, is categorized as a social media site. Facebook functions as a proprietary walled garden, limiting what developers can do with the application interface (API) and thus allowing the site to retain user data. Users must register before using the site and can subsequently add other users as friends. In addition, a “Follow” button allows users to subscribe to public groups or organizations without adding them as friends. Facebook is used primarily to maintain personal connections.

Younger users (ages 15 to 24) are more likely to use Facebook to manage their social lives (Exact Target 2010b). Facebook offers two categories of ads: premium ads and marketplace ads. Premium ads appear in the newsfeed, on the right-hand side of the page, in the mobile newsfeed, and on the logout page. Marketplace ads also appear on the right-hand side of the page. All premium Facebook ads begin as content posted to a brand page.

Twitter, an instant messaging system that allows users to send text messages up to 140 characters in length to a list of followers, is categorized as a microblogging site. Users can also share videos and links to other content in their tweets. Tweet broadcasts can be public or private. Registered users can read and post tweets, but unregistered users can only read them. In addition, Twitter's flexible API streams tweets to other online forums, such as blogs and corporate sites. Compared to Facebook, Twitter functions more as an open platform, allowing developers to build upon it. In an effort to monetize the site, however, Twitter is also beginning to restrict the terms of its API, moving toward a proprietary walled garden. Twitter provides users with unprecedented access to influencers across many markets, including celebrities and thought leaders in various industries. On Twitter, brands can advertise using direct advertising (promoted tweets, trends, and accounts), third-party network advertising (sponsored tweets), and publisher direct advertising, wherein brands hire influential individuals to publish brand tweets. Both Facebook and Twitter can be accessed on mobile applications.

Industry research has measured the demographics and behavior of social media users. In terms of demographics, social media skew female, young, and White. Approximately 54% of social media users are female (Edison 2012). The slight female skew is reflected across social media accounts. Specifically, 56% of online women have Facebook accounts compared to 49.5% of online men and 16.9% of online women have Twitter accounts compared to 15.5% of online men (Burst Media 2013). Social media users are also young. Over half of social media users are between ages 12 and 34 (Edison 2012). Among 12- to 24-year-olds, approximately 80% maintain a personal profile on Facebook and 18% have a Twitter account. In terms of race and ethnicity, however, the results reveal an interesting skew. While non-Whites account for only 15% of the total social media users, social media penetration among the non-White population exceeds that of the White population. Specifically, 75% of the Black Internet users and 80% of Hispanic Internet users use social networking sites compared to 70% of White Internet users (Pew 2013). In the case of Twitter, penetration among non-White Internet users is almost twice as high as penetration among White Internet users. Specifically, Twitter users account for 27% of Black Internet users and 28% of Hispanic Internet users compared to only 14% of White Internet users.

Following brands on Facebook and Twitter has doubled from 2010 to 2012 (Vision Critical 2013). Specifically, 33% of social media users (age 12 and up) reported that they followed companies or brands during 2012 compared to 16% in 2010. The primary reason provided for following brands on Facebook and Twitter was “Sales/discounts/coupons.” About 43% of social media users reported that they purchased products after interacting with the brand on social media (Vision Critical 2013).

In contrast to the industry focus on social media user demographics and behavior, academic research has focused on consumer perceptions. This study builds on a growing body of scholarly research that explores how individuals perceive social media and what motivates their involvement with Facebook and Twitter. Specifically, this study probes the factors that are most important to young adults (ages 18 to 34) when forming the intention to follow brands on Facebook and Twitter. The remainder of this article provides a review of the relevant literature, the research hypotheses, method, results, and conclusions. Finally, research limitations and opportunities for future research are reviewed.

LITERATURE REVIEW

Motivations for Social Media Use

Because social media is relatively new, academic research has sought to define both the medium and the user. There are studies identifying dimensions of uses and gratifications (Bonds-Raacke and Raacke 2010; Langstedt 2013; and Raacke and Bonds-Raacke 2008), the psychology of social media use (Phillips 2008), the determinants of user engagement (Chu and Kim 2011), and the personalities of the users (Ross et al. 2009). Research suggests that although social media use varies according to the users’ personalities and familiarity with the technology, motivation to use social media is not associated with any personality variable (Ross et al. 2009). Rather, the use of social media appears to relate to the gratifications sought. Social networks satisfy the need for information, friendship, and connection (Bonds-Raacke and Raacke 2010; Raacke and Bonds-Raacke 2008). In their study regarding sports fans’ Twitter use, Witkemper, Lim, and Waldburger (2012) found that information and entertainment were both key motivations among users. Agrifoglio and colleagues (2012) found that entertainment was an important gratification for Twitter users depending upon the context of use. Specifically, intrinsic motivation (enjoyment) better explained Twitter usage than extrinsic motivation (information) unless Twitter was used in a work context. These findings were supported by subsequent research that suggested that consumption of social media was aligned with ritual, passive use and using social media as a source of information was indicative of instrumental, active use (Langstedt 2013). The research found that social media users were more likely to consume information than to communicate information, suggesting that Facebook and Twitter are used primarily in a passive manner to satisfy the need for entertainment.

Brands and Social Media

Advertisers use social media to involve their brands in the social networks of consumers. Unlike other media, social media allows advertisers to create brand-related content that evokes immediate advertiser-consumer interaction. For example, advertisers establish brand profiles on social networks that provide opportunities for consumers to respond in the form of comments, photos, videos, and recommendations. Consumers can also reach out to the brand's other users by responding to their brand-related postings. The advantages of brand presence on social media are not yet reliably calculated, however.

Advertisers appear to benefit from the social network itself. Chatterjee (2011) found, for example, that consumers were more likely to respond to brand postings on their friends’ social network sites if the posting was generated by their friend rather than by the brand. People who had not visited the brand page before were more likely to click on the brand link posted on a friend's site. Therefore, the friend's posting successfully generated brand response among less involved consumers. In fact, brands benefit not only from the exposure to an individual's network of friends but also from trust and peer influence associated with the individual (Chu and Kim 2011). Peer influence may have a direct effect regarding whether consumers engage with specific organizations and corporations on social media. Furthermore, Facebook and Twitter may yield different levels of peer influence. When comparing determinants of Twitter and Facebook use, Lee and Cho (2011) found that use of Facebook was more influenced by social factors than Twitter. This disparity may be explained by the nature of communication on different social media sites. Facebook encourages commentary among a community of users who are quick to exchange remarks on the photos, videos, and experiences posted by others. Twitter provides a far less chatty forum given the brevity required by its headline format. Similarly, branded tweets are viewed differently than branded postings on Facebook. Kwon and Sung (2011) found, for example, that consumers interacted with brands as a collective on Facebook while perceiving their brand interaction on Twitter as more individualistic due to the human characteristics of a tweet.

This explanation aligns with the findings of Smith, Fischer, and Yongjian (2012) in their study regarding the differences in brand-related, user-generated content between Twitter, Facebook, and YouTube. Consistent with the notion that Twitter is more individualistic and Facebook is more collective, they determined that Twitter was the most likely social media to feature prominent branding but generated less positive content about the brands compared to Facebook and YouTube. The more collective nature of Facebook appeals to those who wish to hear about others’ experiences with the brand. Twitter followers, on the other hand, are more interested in hearing from the brand itself rather than the brand followers.

This study investigates factors that affect the decision to follow brands on social media by exploring factors that have been found to predict consumer acceptance of new technology, brands, and advertising.

THEORETICAL FRAMEWORKS AND HYPOTHESES

Two theories that have been applied to the intention to use social media inform the research design for this study. The first, the theory of planned behavior (Ajzen 1991), proposes that attitude, subjective norm, and perceived behavioral control underlie an individual's intentions and actions. Behavioral beliefs affect attitudes toward the act, while normative beliefs affect the individual's perceptions of subjective norms. Perceived behavioral control focuses on the individual's perception of the easiness of the behavior, consistent with Bandura's (1986) concept of self-efficacy. Pelling and White (2009) applied the theory of planned behavior to young people's use of social networking sites. They found only partial support for the model, however, in that perceived behavioral control did not predict behavioral intent or behavior.

The second theory, the technology acceptance model (TAM), suggests that adoption of technology is not influenced by subjective norms at all (Davis, Bagozzi, and Warshaw 1989). Rather, the authors contend that attitudes and intentions to use specific technology are influenced by the perceived usefulness and ease of use (perceived behavioral control) related to the technology. Their study indicated, however, that subjective norms may influence behavioral intent indirectly, through attitudes. In a study regarding factors affecting attitudes toward shopping on social media (Lee and Cho 2011), it was determined that such attitudes were affected by perceptions of usefulness, ease of use, security of the shopping service, and suitability of the items in regard to the social network.

Both theories posit that attitudes have a direct effect on intention to act. While TAM proposes that perceived ease of use affects attitude toward the act, the theory of planned behavior assumes that perceived ease of use has a direct effect on intention to act. Neither model appears to completely explain users’ social media behavior. As a consequence, this study employs variables from both models to determine whether they form new relationships to explain following brands on Facebook and Twitter.

Perceived Ease of Use

Both theories suggest that social media users’ perceived self-efficacy (or perceived ease of use) in terms of navigating social media would affect their continued use of social media. Internet self-efficacy, the need to belong, and collective self-esteem have all been found to be positively related to students’ attitudes toward social networking sites and their willingness to use such sites (Gangadharbatla 2008). Lee and Cho (2011), Agrifoglio and colleagues (2012), and Barnes and Böhringer (2011) proposed that the users’ perceived self-efficacy in terms of navigating social media affected their amount of social media use as well.

Consistent with their findings, Lee and Cho (2011) found that the length of time users had Facebook accounts related positively to continuous usage. Longtime Facebook members would most likely have acquired self-efficacy specifically related to Facebook. They indicated that continuous Twitter use, on the other hand, was positively related to heavy mobile phone use, indicating that Twitter depended on the ease of using mobile technology. Agrifoglio and colleagues (2012) also found that perceived ease of use directly affected continuous use of Twitter. Drawing on expectation-confirmation theory (ECT) from consumer behavior, Barnes and Böhringer (2011) determined that the continuous use of Twitter was at least partially determined by habit, which again would suggest perceived ease of use. The literature suggests, consistent with the theory of planned behavior and TAM, that social media users’ perceptions of ease of Facebook and Twitter use have a direct effect on their continued use of those social media.

H1: Users’ perceptions regarding the ease of following brands on Facebook and Twitter positively affect their intentions to follow brands on Facebook and Twitter.

Perceived Usefulness

McCorkindale, Distaso, and Sisco (2013) determined that Millennials were more likely to engage with organizations and corporations on social media if they were already engaged with the organization offline. In the absence of a previously established relationship, they found that consumers were more likely to engage with organizations and corporations that provided incentives such as discounts or other special offers. This finding conforms to the industry research regarding the consumer motivations for following brands on Facebook and Twitter (Vision Critical 2013). Specifically, industry research indicates that consumers follow brands on Facebook and Twitter in order to receive discounts and promotions (Exact Target 2010a, 2010b). In addition, Facebook and Twitter users follow companies, brands, and associations to learn about updates on future products and to stay informed about the companies’ activities. It would appear, then, that the perceived utility attributed to following brands on Facebook and Twitter relates to discounts, promotions, and information. TAM proposes that the perceived usefulness of a technology directly affects intentions to use the technology.

H2: Users’ perceptions regarding the usefulness of following brands on Facebook and Twitter positively affect their intentions to follow brands on Facebook and Twitter.

Normative Beliefs

Normative beliefs refer to the perceived behavioral expectations of important referent individuals or groups. The theory of planned behavior posits that normative beliefs—perceived approval or disapproval of a particular behavior—will affect individuals’ intentions to engage in the behavior. Research suggests that peer influence varies across social media categories. Lee and Cho (2011) found that use of Facebook was more influenced by social factors than was use of Twitter. Given the collective nature of Facebook compared to the individualistic nature of Twitter, it is likely that normative beliefs are a more important factor in the decision to follow brands on Facebook rather than on Twitter.

H3: Normative beliefs affect users’ intentions to follow brands on Facebook to a greater extent compared to following brands on Twitter.

Subjective Norms

Subjective norms represent individuals’ motivations to conform to social pressure regarding whether to engage in a specific behavior. In the realm of social media, peer pressure is exerted in the form of affiliations as well as comments. For example, Witkemper, Lim, and Waldburger (2012) determined that one of the key constraints to following athletes on Twitter was the belief that peers were not using Twitter to follow athletes, suggesting the influence of subjective norms. Research also indicates that brands benefit from the trust and peer influence associated with social networks (Chatterjee 2011; Chu and Kim 2011). Specifically, consumer ad evaluations were found to have an effect on consumers’ receptivity to social media ads (Steyn et al. 2011). Subjective norms should, therefore, have a direct effect on individuals’ intentions to follow brands on Facebook and Twitter.

H4: Subjective norms affect users’ intentions to follow brands on Facebook and Twitter.

Attitude Toward Following Brands on Facebook

The theory of planned behavior and TAM both posit that behavior is affected by attitudes toward the behavior (Ajzen 1991; Davis, Bagozzi, and Warshaw 1989). Therefore, if consumers have positive attitudes toward following brands on social media they would have positive intentions to follow brands on social media.

H5: Consumers’ attitudes toward following brands on Facebook and Twitter will positively affect their intentions to follow brands on Facebook and Twitter.

Information-Seeking Behavior

Information-seeking behavior is generally associated with consumers’ desires to reduce risk when making purchase decisions. Schiffman, Schus, and Winer (1976) found a positive correlation between perceived risk levels and brand loyalty and information seeking, for example. The role of information-seeking behavior among social media users is not clear. While research suggests that information is a gratification sought by social media users (Agrifoglio et al. 2012; Bonds-Raacke and Raacke 2010; Raacke and Bonds-Raacke 2008; Witkemper, Lim, and Waldburger 2012), there are also indications that social media users are more often passive consumers of information rather than active communicators of information (Langstedt 2013). In fact, followers of brands on Facebook and Twitter are about twice as likely to indicate that they do so to obtain brand information rather than to post brand information (Exact Target 2010a, 2010b).

It is possible, therefore, that social media is used to simplify the information-seeking process. Information-seeking behavior suggests consulting many sources, but social media users may be satisfied with brand updates and comments. This low-intensity version of information seeking is reflected in the fact that while Liang and colleagues (2011) ascertained that a key determinant for consumers’ intent to use social commerce in the future was the quality of information provided by the site, Wolny and Mueller (2013) found that information-seeking behavior did not affect the frequency of consumer engagement with fashion brands on social networks. There are indications, however, that Twitter brand followers are less passive than Facebook brand followers. Among brand followers on Twitter, 20% want to provide feedback (Exact Target 2010a) compared to 13% of Facebook brand followers (Exact Target 2010b).

H6: Information-seeking behavior affects users’ intentions to follow brands on Twitter to a greater extent compared to Facebook.

Online Advertising Skepticism

Research indicates advertising that appears on social networks is evaluated in terms of information and entertainment (Taylor, Lewin, and Strutton 2011), consistent with the way traditional advertising value is assessed (Ducoffe 1995). However, because the primary reasons to follow brands on Facebook and Twitter appear to be transactional in nature (discounts, samples, and promotions), it is unclear whether the act of following brands is perceived as engagement with brand advertising. Advertising skeptics may avoid advertising messages yet remain susceptible to the lure of following brands.

Boush, Friestad, and Rose (1994) suggest that skepticism toward television advertising is a predisposition to reject whatever advertising is shown on television. Their findings indicate that advertising messages are accepted or rejected based on the perceived reliability of the source of information. Consumers have general beliefs regarding the credibility of various kinds of information sources based on their cumulative experience regarding the fairness and factualness of specific sources of information. Consistent with the early work of Hovland and Weiss (1951–52), certain types of information sources are viewed as credible, or trustworthy, and other sources are viewed as untrustworthy. Ultimately, consumers’ acceptance of information is mitigated by the credibility of the source. Consumer attitudes toward social media advertising should, therefore, reflect perceptions regarding the trustworthiness of social media as well as attitudes toward advertising in general.

Recent scholarship indicates that the credibility of social media is enhanced by frequency and recency of postings (Westerman, Spence, and Van Der Heide 2014; Xu 2013). In addition, postings that required cognitive—or careful—processing were positively correlated with overall credibility (Westerman, Spence, and Van Der Heide 2014). Xu (2013) found that the sheer quantity of social recommendations was a powerful influencer because it allowed users to expend little effort when determining which news articles to access. A study of source effects in consumer-generated advertising (Steyn et al. 2011) determined that consumer comments significantly affected perceptions of online advertising, indicating the interrelationship of comments and source credibility. Brands that maintain good social media relationships in terms of frequency, recency, and promotional value of postings may be perceived as more credible sources than traditional advertising and “following brands” may be perceived as more authentic brand communication compared to traditional advertising. Therefore, individuals’ skepticism regarding advertising may be inversely related to their likelihood to follow brands.

H7: Online advertising skepticism positively affects users’ intentions to follow brands on Facebook and Twitter.

Brand Consciousness

Because price promotion appears to be a primary motivation to follow brands on social media (Vision Critical 2013), the role of brand consciousness may also be important. Research indicates that brand consciousness is a means to reduce risk when purchasing an item (Donthu and Gilliland 1996). Brand-conscious consumers were found to be less willing to try new or different brands than consumers who were not brand conscious. In other words, brand loyalty is motivated by the desire to predict consistent, desired product performance. Research also suggests that those who follow brands on social media are likely to be familiar with the brands they follow. Wolny and Mueller (2013) determined that users who were involved with particular brands were most likely to engage with the brands on social media. Given the fact that the primary motivation for following brands on social media is to receive discounts and samples (Vision Critical 2013), it would appear that brand loyalists are using social media to reduce purchase price. Brand consciousness may, therefore, influence the decision to follow brands on Facebook or Twitter as a means to ensure that the individual is aware of all price promotions and product news about certain trusted brands.

H8: Brand consciousness positively affects users’ intentions to follow brands on Facebook and Twitter.

METHOD

Procedure

A 90-item online questionnaire was developed for users of Facebook and Twitter. Respondents were asked about their beliefs and attitudes regarding social media in general and their usage of Facebook and Twitter in particular. Respondents were also asked about their intentions to follow brands on Facebook and Twitter. In addition, demographic data were gathered from all participants. All data were collected electronically by an online research company during a one-week period (March 26 to April 1, 2012). Roughly 72 respondents were collected each day to ensure an even distribution of respondents on each day of the week.

Participants

The final sample consisted of 502 current social media users. Qualified respondents reported using Facebook and Twitter at least once a month. According to the Pew Internet Project (Duggan and Brenner 2013), the age segment representing the greatest participation in social media is teens and young adults (ages 18 to 29). For this study, participants were recruited between the ages 18 and 34. Of the participants, 51.6% (N = 259) were male and 48.4% (N = 243) were female. The sample was also balanced in terms of age distribution in that 49.2% (N = 247) were between the ages of 18 and 24, and 50.8% (N = 255) were between the ages of 25 and 34.

Participants were screened for social media use, age, gender, race, ethnicity, geographic representation, and income. In all, 87% of participants reported daily use of Facebook and 56% of participants reported daily use of Twitter. There were no significant differences in frequency of use based on age or gender. Participants were also screened to reflect the demographics of the U.S. adult Internet users. The racial, ethnic, geographic, and income composition of the sample was roughly comparable to those of U.S. Internet users, although it was less skewed to the Southern region and included a greater percentage of Asian participants.

Measures

Established scales were adapted to measure the variables (see appendix, Table A1). Specifically, perceived ease of use and perceived usefulness (Davis, Bagozzi, and Warshaw 1989), information seeking (Raju 1980), brand consciousness (Shim and Gehrt 1996; Donthu and Gilliland 1996), normative beliefs (Karahanna, Straub, and Chervany 1999), and advertising skepticism (Obermiller and Spangenberg 1998) were measured using seven-point, Likert-type scales (1 = Very strongly disagree, 7 = Very strongly agree). Subjective norms (Fishbein and Ajzen 1975) and attitude toward the act (Donthu and Gilliland 1996) were measured using seven-point semantic differential scales. The dependent variable, behavioral intent (Bauer et al. 2005), was measured using a seven-point, Likert-type scale.

Data Analysis

Confirmatory factor analysis (CFA) using AMOS 21 was used to establish construct reliability and validity. Chi square (χ²), degrees of freedom (df), the ratio of chi square to degrees of freedom (χ²/df), the p value, comparative fit index (CFI), Tucker-Lewis index (TLI), and root mean square error of approximation (RMSEA) are reported. SPSS statistical software was used for all statistical analyses. Pearson correlations and variance inflation factor tests were analyzed to assess multicollinearity issues. Multiple regression analyses were used to examine how the variables influenced intentions to follow brands on Facebook and Twitter.

RESULTS

Measurement Validation

A first-order CFA was conducted to evaluate the appropriateness of the measurement model for the latent constructs. The initial model fit for the Facebook and Twitter CFA models was poor and the standardized regression weights were examined for each latent variable indicator. Two brand consciousness items, one information seeking item, and two attitude toward the act items were eliminated from each model because they did not meet the recommended.70 threshold weight. The revised models achieved acceptable fit in accordance with the guidelines proposed by Hair and colleagues (2010). Specifically, fit indices for the revised Facebook model (χ² = 1683.07, df = 743, χ²/df = 2.26, p <.05, CFI =.95, TLI =.94, RMSEA =.05) and the revised Twitter model (χ² = 1579.80, df = 743, χ²/df = 2.13, p <.05, CFI =.95, TLI =.95, RMSEA =.05) suggested satisfactory fit for the data. The significant p values are most likely attributable to the large sample size and therefore were not considered problematic.

Measures of latent variables achieved satisfactory reliability levels. Construct reliability was assessed using two measures of internal consistency: Cronbach's alpha (α) and composite reliability (CR). Values for both measures should be above 0.70 to indicate an acceptable reliability (Chin 1998). Table 1 provides measures of construct reliability and validity. The constructs were also determined to meet the necessary criteria for validity according to Hair and colleagues (2010). Convergent validity was achieved when the CR assumed values greater than the average variance extracted (AVE), and if the AVE was greater than 0.5. Discriminant validity was achieved if the values of the maximum squared shared variance (MSV) and the average shared squared variance (ASV) were less than the AVE.

TABLE 1 Measures of Construct Reliability and Validity

	α	CR	AVE	MSV	ASV
Information Seeking*	0.89	0.88	0.59	0.55	0.36
Online Advertising Skepticism*	0.95	0.96	0.71	0.57	0.31
Brand Consciousness*	0.91	0.91	0.66	0.57	0.28
Normative Beliefs*	0.91	0.93	0.69	0.57	0.40
Subjective Norms*	0.71	0.75	0.52	0.39	0.12
Attitude Toward Following Brands on FB	0.91	0.91	0.77	0.39	0.14
Perceived Usefulness of Following Brands on FB	0.93	0.94	0.78	0.78	0.41
Perceived Ease of Following Brands on FB	0.94	0.94	0.83	0.72	0.32
Intent to Follow Brands on FB	0.93	0.93	0.81	0.78	0.40
Attitude Toward Following Brands on TW	0.91	0.92	0.79	0.38	0.13
Perceived Usefulness of Following Brands on TW	0.94	0.94	0.79	0.78	0.43
Perceived Ease of Following Brands on TW	0.94	0.94	0.84	0.64	0.32
Intent to Follow Brands on TW	0.93	0.93	0.81	0.78	0.40

Note. α = Cronbach's alpha; CR = composite reliability; AVE = average variance extracted; MSV = maximum squared shared variance; ASV = average shared squared variance; FB = Facebook; TW = Twitter.*Results are reported for the Facebook measurement model because there were negligible differences between Facebook and Twitter models.

Pearson correlations revealed significant, strong correlations between all of the nine observed composite variables for each model, suggesting possible multicollinearity issues (see appendix, Tables A2 and A3). Variance inflation factor (VIF) tests revealed that the multicollinearity issues were not a problem, however. Specifically, the VIFs were less than 5 in the worst cases (Hair et al. 2010).

Hypotheses Testing

Observed variable multiple regressions were utilized to analyze the net effects of each variable on intent to follow brands on Facebook and Twitter. Initially, all eight independent variables (information seeking, online ad skepticism, brand consciousness, subjective norms, normative beliefs, attitude toward the act, perceived usefulness, and perceived ease of use) were included for the multiple regressions.

For intention to follow brands on Facebook, the eight predictors explained 76% of the variance, R² =.76, adj. R² =.76, F (8, 493) = 198.39, p <.001. Perceived ease of use (t = 13.25, p <.001), perceived usefulness (t = 10.15, p <.001), and subjective norms (t = 2.94, p <.01) each significantly predicted behavioral intent. For intention to follow brands on Twitter, the eight predictors explained 74% of the variance, R² =.74, adj. R² =.74, F (8, 493) = 176.76, p <.001. Perceived usefulness (t = 13.16, p <.001), perceived ease of use (t = 8.72, p <.001), information seeking (t = 2.50, p <.01), subjective norms (t = 2.43, p <.01), online ad skepticism (t = −2.22, p <.05), and brand consciousness (t = −2.02, p <.05) each significantly predicted behavioral intent. Table 2 displays the unstandardized regression coefficients (B), standard error, and standardized regression coefficients (β) for each variable.

TABLE 2 Multiple Regressions for Facebook and Twitter: Full Models

	Facebook			Twitter
Predictor variable	B	SE B	β	B	SE B	β
Information Seeking	.01	.04	.01	.10	.04	.09*
Online Ad Skepticism	−.04	.04	−.04	−.09	.04	−.09*
Brand Consciousness	.02	.04	.02	−.07	.04	−.07*
Subjective Norms	.09	.03	.08**	.07	.03	.07*
Normative Beliefs	.08	.05	.07	.01	.05	.01
Attitude Toward Following Brands	.02	.02	.02	−.04	.02	−.04
Perceived Usefulness of Following Brands	.39	.04	.38***	.55	.04	.03***
Perceived Ease of Use of Following Brands	.44	.03	.43***	.30	.03	.29***

*p <.05; **p <.01; ***p <.001.

Reduced models were achieved by iteratively removing the variables with the lowest levels of significance while maintaining the adjusted R². The reduced model for Facebook consisted of five independent variables (online ad skepticism, subjective norms, normative beliefs, perceived usefulness, and perceived ease of use) and explained 76% of the variance, R² =.76, adj. R² =.76, F (5, 496) = 318.31, p <.001. Perceived ease of use (t = 13.79, p <.001), perceived usefulness (t = 10.43, p <.001), and subjective norms (t = 3.73, p <.001), each significantly predicted behavioral intent.

The reduced model for Twitter was achieved by iteratively removing the two variables that did not significantly predict behavioral intent while maintaining the adjusted R². The reduced model consisted of six independent variables (information seeking, online ad skepticism, brand consciousness, subjective norms, perceived usefulness, and perceived ease of use) and explained 76% of the variance, R² =.74, adj. R² =.74, F (5, 496) = 235.14, p <.001. All of the variables significantly predicted behavioral intent. Specifically, perceived usefulness (t = 13.88, p <.001), perceived ease of use (t = 8.79, p <.001), online ad skepticism (t = −2.5, p <.01), information-seeking (t = 2.47, p <.01), subjective norms (t = 2.06, p <.05), and brand consciousness (t = −1.99, p <.05) each significantly predicted behavioral intent. Table 3 displays the unstandardized regression coefficients (B), standard error, and standardized regression coefficients (β) for each variable.

TABLE 3 Multiple Regressions for Facebook and Twitter: Reduced Models

	Facebook			Twitter
Predictor variable	B	SE B	β	B	SE B	β	Hypotheses
Perceived Ease of Use of Following Brands	.44	.03	.43***	.30	.03	.29***	1	Supported
Perceived Usefulness of Following Brands	.40	.04	.38***	.54	.04	.52***	2	Supported
Normative Beliefs	.08	.04	.07*	—	—	—	3	Not supported
Subjective Norms	.10	.03	.09**	.05	.03	.05*	4	Supported
Attitude Toward Following Brands	—	—	—	—	—	—	5	Not supported
Information Seeking	—	—	—	.10	.04	.08**	6	Supported
Online Ad Skepticism	−.06	.03	−.05	−.10	.04	−.09**	7	Not supported
Brand Consciousness	—	—	—	−.07	.04	−.07*	8	Not supported

*p <.05; **p <.01; ***p <.001.

Perceived ease of following brands was one of the strongest indicators of intent to follow brands for both Facebook and Twitter, supporting hypothesis 1. Individuals were more likely to follow brands on Facebook and Twitter if they believed that the activity was easy to accomplish.

Perceived usefulness of following brands was also a significant indicator of intent to follow brands for both Facebook and Twitter, supporting hypothesis 2. Individuals were more likely to follow brands on Facebook and Twitter if they believed that the activity was useful.

Normative beliefs represent the individuals’ beliefs regarding the opinions of people they hold in esteem regarding following brands on Facebook or Twitter. Although normative beliefs were included in the reduced Facebook model they were not a significant predictor in regard to following brands on Facebook. Normative beliefs were also not significant in regard to following brands on Twitter and, furthermore, were not included in the reduced Twitter model. Hypothesis 3 was not supported.

Subjective norms were significant indicators of intent to follow brands for both Facebook and Twitter, supporting hypothesis 4. The more peer pressure individuals experienced to follow brands on Facebook and Twitter, the more likely they intended to do so.

The results indicated that individuals’ attitudes toward following brands on Facebook or Twitter were not significantly predictive of their intentions to follow brands on Facebook or Twitter. Hypothesis 5 was not supported. This result was unexpected due to the fact that attitudes toward the behavior directly affect intention to act in the theory of planned behavior and TAM.

Information-seeking behavior was not a significant predictor of following brands on Facebook and was not included in the reduced Facebook regression model. However, information-seeking behavior was a significant predictor of following brands on Twitter, supporting hypothesis 6. Information-seeking behavior is a more significant factor in the decision to follow brands on Twitter than to follow brands on Facebook.

Online advertising skepticism did not significantly predict following brands on Facebook. Online advertising skepticism was, however, a significant, negative predictor of following brands on Twitter. Hypothesis 7 was not supported. Contrary to expectations, following brands on social media was not more likely to occur among those who were skeptical of online advertising. On the contrary, following brands on Facebook was not related to advertising skepticism at all, and following brands on Twitter was negatively related to advertising skepticism. In other words, the more skeptical individuals felt about online advertising, the less likely they were to follow brands on Twitter.

Brand consciousness was not a significant predictor of following brands on Facebook and was not included in the reduced Facebook regression model. Although brand consciousness was a significant predictor of following brands on Twitter, when controlling for the other predictors the net effect of brand consciousness on following brands on Twitter was negative. In other words, the more positively individuals felt about branded products, the less likely they were to follow brands on Twitter. Hypothesis 8 was not supported. Brand consciousness does not positively affect individuals’ intentions to follow brands on Facebook and Twitter.

DISCUSSION AND CONCLUSION

The results suggest that the decision to follow brands on Facebook or Twitter may be more impulsive than the process suggested by the theory of planned behavior and TAM. The fact that users’ intentions to follow brands were most strongly affected by their perceptions of the usefulness and easiness of the activity in combination with the strong, positive effect of peer pressure indicates that following brands may resemble the act of an impulse purchase.

Impulse buying has been defined as a spontaneous decision that is not based on thoughtful consideration (Rook 1987). Impulse-buying research appears to fall into three streams of study focusing on consumption impulsivity, social and cultural factors, and individual-level constructs that explain differences in perception, motivation, and behaviors or “normative beliefs” (Xiao and Nicholson 2011, 2519). Viewed through the prism of this research, an individual's perception of “ease of use” when deciding to follow a brand suggests a conducive shopping environment similar to the enticement of a candy display near the checkout counter (Rook and Fisher 1995). The perceived “usefulness” of following brands suggests the magnetic pull of a desired object and the power of peer pressure—“subjective norms”—overwhelms the normal decision-making process (Lin and Chen 2012). In fact, those who purchase on impulse do so despite normative influences (Rook and Fisher 1995). In other words, respected wisdom (“normative beliefs”) does not always prevail when the individual succumbs to immediate gratification.

Therefore, the significant, predictive nature of subjective norms, ease of use, usefulness, and the unimportance of normative beliefs when following brands conforms to explanations of impulse-buying behavior. It would appear that attitude toward the act is overwhelmed by those factors. In fact, this study did not reveal compelling information linking attitudes toward advertising and brands to intentions to follow brands on Facebook and Twitter. While online advertising skepticism was negatively related to intent to follow brands on Twitter, suggesting that Twitter usage corresponds to a favorable predisposition to advertising, attitudes toward online advertising were not at all predictive of intent to follow brands on Facebook. Furthermore, attitudes toward brands were not predictive of intent to follow brands on either Facebook or Twitter. The lack of influence attributed to attitudes may suggest that—in the realm of social media—attitudes are less important than peer pressure and perceived gratifications.

The unimportance of brand consciousness, for example, may reflect the discount orientation of those who follow brands on social media. Specifically, 40% of those who follow brands on Facebook and 30% of those who follow Twitter are most interested in brand discounts and promotions (Exact Target 2010a, 2010b). This study suggests that brand followers may be indiscriminately pursuing bargains regardless of the brands. Following brands on Twitter also appears to satisfy a need for information. The study indicated that information-seeking behavior was an important factor for Twitter users’ intentions to follow brands but not significant for following brands on Facebook.

Overall, three main conclusions emerge from this study. First, consumers who follow brands on Facebook and Twitter assume that the activity is easy and useful. While both the theory of planned behavior and TAM propose that perceived usefulness has a direct effect on intention to act, both also posit that perceived ease of use indirectly affects intention to act through attitudes. This study proposes that perceived ease of use directly affects users’ intentions to follow brands on Facebook and Twitter.

Second, peer pressure is an important factor in the decision to follow brands on Facebook and Twitter. The fact that peers are engaged in the activity increases the attractiveness of the activity.

Third, following brands on Facebook and Twitter may satisfy different user gratifications. For example, information seeking was a significant factor for Twitter followers but not for Facebook followers.

For practitioners, these findings indicate that using social media should be an effortless experience for users that provides promotional incentives. Advertisers would be wise, however, to regard branded communication on Facebook very differently compared to branded communication on Twitter.

Those who follow brands on Facebook have involved themselves in a collective activity and should be privy to many comments and responses from brand users. Brand followers on Facebook are more interested in the comments of real people as opposed to corporate messages. Branded messages should, therefore, be used to open conversations, such as Dove's invitations to comment on postings or photos. “Sound Off: Girls Night In or Ladies Night Out?” generated 200 likes and 25 comments within a day, for example. Alternatively, Facebook can be used to publicly involve brand followers in promotions such as the Crate and Barrel “Ultimate Wedding“ sweepstakes. Those who follow brands on Twitter, on the other hand, are more information seeking, tolerant of branded messaging, and focused on one-on-one communication. They should receive frequent, chatty tweets from the brand, and those brand followers who retweet brand news should be rewarded. Pantene ingeniously tweeted real-time commentary about celebrity hairstyles during red carpet arrivals at the 2013 Academy Awards (#WantThatHair), complete with sketches and how-to instructions. Arby's created a sensation simply by commenting on the similarity between an artist's hat and the Arby's logo during the Grammy Awards.

Finally, advertisers should be aware of the lack of brand loyalty among brand followers. Promotional efforts on social media should be designed to create added value rather than discounts. Examples of using Facebook promotions to create consumer involvement include the Dove “selfie” sweepstakes event, the Victoria's Secret announcement of an in-store sweepstakes event, and the Best Buy discounted smartphone activation with a preorder of Game of Thrones season 3.

LIMITATIONS AND FUTURE RESEARCH

A limitation of this study is that it relied on self-reported data. A survey instrument was used based on the belief that a broad, nationally representative sample would provide more useful information than an experiment. Self-reporting could allow for overestimations of respondents’ perceived social media self-efficacy and underestimations of their brand consciousness. Furthermore, it is also conceivable that respondents were more positive regarding their intentions to follow brands on Facebook and Twitter than their subsequent actions would demonstrate.

A second limitation is the fact that participation in the survey was limited to subjects with relatively extensive social media use. Specifically, participants were screened to ensure that they had used a social media site, such as Facebook; a daily-deal site, such as Groupon; and Twitter, a microblogging site. The extensive screening may have resulted in a sample with exceptionally strong social media self-efficacy (M = 5.09, SD = 1.26) and thus skewed the survey findings.

The findings of this study indicate a number of areas for future research. First, given the importance of perceived ease of use and perceived usefulness, it would be beneficial to examine the antecedents for both constructs in terms of uses and gratifications. For example, the concept of usefulness may reflect consumers’ needs for value. While there are academic studies regarding the use of online coupons (Dickinger and Kleijnen 2008; Kang et al. 2006), current social media studies have not addressed the price promotion properties of social media. Industry research indicates that more than half of the women who interact socially with brands do so to access offers and coupons (Burst Media 2013). It is possible that this aspect of social media use is acknowledged by consumers as “usefulness” or an implicit reward for social media use.

Further investigation is required to determine the role of attitudes when consumers form the intent to engage with brands on social media. Given the instantaneous nature of the decision, the attitude toward the act of following brands may actually be incorporated within the notions of usefulness and easiness. A more robust attitude measure may be required.

Finally, the most interesting area for future research is the investigation of impulsive-buying theory to better explain the process of consumer involvement with brands on social media.

APPENDIX

TABLE A1 Indicator Factor Loadings and Construct Descriptive Statistics

Indicators	Loading	M	SD
Information-Seeking Behavior (IS)		4.84	1.26
IS1: I like to browse through mail-order catalogs even when I don't plan to buy anything.	.75
IS2: I often read the information on the package of products just out of curiosity.	.76
IS4: I often read advertisements just out of curiosity.	.80
IS5: I often opt in to follow brands on Facebook or Twitter to obtain brand relevant information.	.80
IS6: I often give my e-mail address to retailers to receive product offers and information.	.73
Online Advertising Skepticism (SK)		4.51	1.32
SK1: We can depend on getting the truth in most online advertising.	.84
SK2: Online advertising's aim is to inform the consumer.	.80
SK3: I believe online advertising is informative.	.80
SK4: Online advertising is generally truthful.	.87
SK5: Online advertising is a reliable source of information about the quality and performance of products.	.86
SK6: Online advertising is “truth well told.”	.82
SK7: In general, online advertising presents a true picture of the product being advertised.	.86
SK8: I feel I have been accurately informed after viewing most online advertisements.	.86
SK9: Most online advertising provides consumers with essential information.	.84
Brand Consciousness (BC)		4.44	1.34
BC1: The well-known national brands are best for me.	.82
BC2: The more expensive brands are usually my choices.	.82
BC3: The higher the price of a product the better the quality.	.82
BC4: Nice department stores and specialty stores offer me the best products.	.79
BC5: I usually purchase name-brand products.	.80
Subjective Norms (SN)		4.78	1.37
“Most people who are important to me probably consider my use of social media advertising to be…”
SN1: Foolish—Wise	.40
SN2: Valuable—Worthless	.82
SN3: Useful—Useless	.86
Normative Beliefs (NB)		4.70	1.29
NB1: Successful people think I should use social media advertising.	.80
NB2: People I work with think I should use social media advertising.	.84
NB3: My peers think I should use social media advertising.	.83
NB4: My friends think I should use social media advertising.	.85
NB5: People who find good deals think I should use social media advertising.	.83
Attitude toward Following Brands on Facebook (ATT_FB)		5.08	1.63
“Following brands on Facebook is…”
ATT 2: Useful—Useless	.84
ATT 4: Valuable—Worthless	.90
ATT 5: Good—Bad	.88
Perceived Usefulness of Following Brands on Facebook (PU_FB)		4.93	1.41
PU 1: Following brands on Facebook would help me find solutions more quickly.	.91
PU 2: Following brands on Facebook would increase my productivity.	.84
PU 3: Following brands on Facebook would enhance my effectiveness.	.89
PU 4: I would find following brands on Facebook useful.	.91
Perceived Ease of Following Brands on Facebook (PEOU_FB)		5.28	1.46
PEOU1: Learning to follow brands on Facebook would be easy for me.	.92
PEOU2: I would find it easy to follow brands on Facebook.	.92
PEOU3: It would be easy for me to become skillful at following brands on Facebook.	.90
Intention to Follow Brands on Facebook (BI_FB)		5.02	1.49
BI1: My general intention to follow brands on Facebook is very high.	.89
BI2: I will think about following brands on Facebook.	.90
BI3: I will follow brands on Facebook in the future.	.91
Attitude toward Following Brands on Twitter (ATT_TW)
“Following brands on Twitter is…”		4.90	1.64
ATT2: Useful—Useless	.87
ATT4: Valuable—Worthless	.91
ATT5: Good—Bad	.88
Perceived Usefulness of Following Brands on Twitter (PU_TW)		4.81	1.41
PU1: Following brands on Twitter would help me find solutions more quickly.	.89
PU2: Following brands on Twitter would increase my productivity	.86
PU3: Following brands on Twitter would enhance my effectiveness.	.90
PU4: I would find following brands on Twitter useful.	.90
Perceived Ease of Following Brands on Twitter (PEOU_TW)		5.06	1.45
PEOU1: Learning to follow brands on Twitter would be easy for me.	.92
PEOU2: I would find it easy to follow brands on Twitter.	.91
PEOU3: It would be easy for me to become skillful at following brands on Twitter.	.92
Intention to Follow Brands on Twitter (BI_TW)		4.90	1.46
BI1: My general intention to follow brands on Twitter is very high.	.89
BI2: I will think about following brands on Twitter.	.91
BI3: I will follow brands on Twitter in the future.	.88

TABLE A2 Pearson Correlations for Facebook Scales

	1	2	3	4	5	6	7	8	9
Information Seeking	1.00
Online Advertising Skepticism	−.67**	1.00
Brand Consciousness	.62**	−.70**	1.00
Normative Beliefs	.67**	−.71**	.63**	1.00
Subjective Norms	.24**	−.28**	.21**	.37**	1.00
Attitude Toward Following Brands on FB	.28**	−.25**	.21**	.28**	.54**	1.00
Perceived Usefulness of Following Brands on FB	.60**	−.53**	.49**	.71**	.40**	.36**	1.00
Perceived Ease of Following Brands on FB	.53**	−.37**	.35**	.56**	.33**	.32**	.70**	1.00
Intent to Follow Brands on FB	.57**	−.49**	.45**	.66**	.43**	.37**	.80**	.79**	1.00

Note. FB = Facebook.*p <.05; **p <.01; ***p <.001.

TABLE A3 Pearson Correlations for Twitter Scales

	1	2	3	4	5	6	7	8	9
Information Seeking	1.00
Online Advertising Skepticism	−.67**	1.00
Brand Consciousness	.62**	−.70**	1.00
Normative Beliefs	.67**	−.71**	.63**	1.00
Subjective Norms	.24**	−.28**	.21**	.37**	1.00
Attitude Toward Following Brands on TW	.28**	−.24**	.21**	.28**	.53**	1.00
Perceived Usefulness of Following Brands on TW	.63**	−.59**	.55**	.72**	.35**	.36**	1.00
Perceived Ease of Following Brands on TW	.55**	−.44**	.43**	.59**	.31**	.30**	.71**	1.00
Intent to Follow Brands on TW	.61**	−.56**	.47**	.66**	.36**	.30**	.82**	.74**	1.00

Note. TW = Twitter.*p <.05; **p <.01; ***p <.001.