Keywords

Introduction

The popularity of social media has generated new social platforms for those seeking to connect with others and share their enthusiasm for certain brands or other topics. Blogs—websites that act as a personal diary that the writer (blogger) makes available to the public—are now considered an essential medium for conveying information and opinion, benefiting from the instantaneous communication offered by the Internet (Hsu et al. 2014).

Bloggers use a range of digital tools to reach massive audiences and secure their loyalty as followers. One key variable in the process of achieving loyalty is that of the credibility of the blogger (Cosenza et al. 2015), who must be perceived as having experience, knowledge, and prestige (Kwon 2011). Generating audience loyalty is vital for bloggers as this has positive consequences for the blog itself and also for any brands promoted within it. Blogs are therefore recognized as influencers in the digital sphere (López 2014). Many firms are now becoming interested in understanding how the activities of bloggers can contribute to achieving greater brand loyalty for them (Uzunoglu and Kip 2014).

That said, not all bloggers have the same capacity or effectiveness when it comes to reaching the market (Cosenza et al. 2015). Hence, it is of interest both to bloggers and to firms to (1) identify those activities the blogger can undertake, along with other indicators, to predict the level of performance of the blog in the market and (2) understand the positive effects of such activities on loyalty to the blog and to the brands promoted on it.

Given this scenario, it is fundamental to study the approaches that bloggers can adopt to achieve improved performance in the market, alongside the effect that their activities may exert on user loyalty toward the blog and toward the brands in question. Knowledge of these issues is of tremendous value as it enables bloggers to identify the most competitive approaches to achieving a stronger market position; and similarly, for firms, it provides indicators with which to predict the effect of having their brand exposed via a given blog on its followers.

In light of the above, the aim of the present work is to understand the effect of the blogger’s interactive activities in terms of loyalty formation among the blog’s followers, taking into account the possible moderating effect of his or her social influence. The work seeks to (1) identify the practices of bloggers when interacting with their public; (2) analyzes the effect of these practices on followers’ loyalty toward the blog (intention to recommend) and toward the brands featured in the blog (purchase intention), taking into account one of the key intermediary variables in this context, namely, the blogger’s credibility; and (3) analyzes the moderating effect that blogger social influence may exert on loyalty formation among followers of the blog.

Antecedents of Loyalty Toward Blogs: The Credibility and Practices of the Blogger

The quest for consumer loyalty has become a genuine “strategic objective” for firms using digital social media (Schau et al. 2009). Attitudinal definitions of loyalty lead to measurements based on preferences, intentions, or affective intensity toward the firm. This attitudinal perspective is also relevant for the online context (e.g., Laroche et al. 2012), as e-loyalty has been defined as “the favorable attitude of a client toward the online retailer that translates into loyal behavior” (Srinivasan et al. 2002:42), with users devoting time and effort to their favorite blogs (Dennen 2014).

In the blog context, the notion of loyalty also refers to (a) intention to recommend the blog to others (Cosenza et al. 2015), as the ranking of a blog is determined by the traffic it attracts, which in turn is determined by the recommendations of other readers (Cosenza et al. 2015), and (b) the intention to purchase the brands featured on the blog (e.g., Cosenza et al. 2015).

In the digital arena, credibility is of utmost importance (e.g., Pan and Chiou 2011) and has been studied from the traditional perspective (e.g., Ohanian 1990) and in the blogosphere (e.g., Durukan and Bozaci 2012). In the latter context, credibility is closely associated with that of the blogger (who acts as an influencer) and may be understood as the credibility of the source itself (Cosenza et al. 2015).

From the very earliest days of the literature on opinion leadership, the term “source credibility” has been a key element in the analysis of influencers. To be an influencer, a person must be perceived as credible, on the basis of their appeal, the trust they inspire in the public, and the experience they convey (Ballantine and Au Yeung 2015; Ohanian 1990, 1991). People who are looking for information and opinions typically take the information provided by influencers to be more credible and useful than that of non-influencers (Nisbet 2006).

The previous literature associated the credibility of the influencer with recommendation intention (e.g., Durukan and Bozaci 2012) and with the intention to purchase the brands mentioned in blogs (e.g., Cosenza et al. 2015).

Despite the importance of blogger credibility in achieving loyalty and the works that have evaluated its indirect effect on loyalty, to date, there have been no studies analyzing the direct effect of blogger credibility on (a) the intention to recommend the blog or (b) the intention to purchase the brands mentioned in the blog. We therefore propose that:

H1.

The blogger’s credibility has a significant positive effect on (a) intention to recommend the blog and (b) intention to purchase the brands featured in the blog.

A blog differs from a traditional website in that (a) the blogger’s comments change continually, (b) the blogger has a unique voice, and (c) it is the blogger who has control over the conversation; hence it is their choice whether readers can interact with the blog or not. Hence, the importance of the blogger’s actions in connecting with their public is clear (Hsu et al. 2014). Moreover, it is easy to set up a blog, but maintaining an audience and achieving renown in the online community is much more complex. Interactivity and reciprocity with the public have become key factors in the analysis of blogs (Dennen 2014). As well as publishing updated content regularly, the blogger should interact with the reader, answering questions in a personalized way (Hsu et al. 2014), all of which form part of the practices he or she must implement to involve readers in their blog (Nardi et al. 2004). Such practices can become a source of competitive advantage for the blogger (Li and Bernoff 2008).

The literature highlights the fact that the proactive, interactive practices undertaken on digital social media impact on user behavior (Schau et al. 2009). The more practices bloggers implement to interact at a personal level with their readers, the more credible they appear to their public (Kauffman et al. 2010), as their knowledge and experience are channeled in such a way as to answer the individual questions of each reader and guide them in their choice of brand, according to their needs and preferences (Hsu et al. 2014). In short, it is a great asset to the blogger to achieve a healthy level of credibility. Thus, it is of interest to establish whether the practices they undertake in relation to the public constitute a valid means of further heightening their credibility. In the present work, we therefore propose that:

H2.

A blogger’s interactive practices have a significant positive effect on his or her credibility.

Meanwhile, Hsu et al. (2014) conclude that bloggers’ interactive practices contribute to generating loyalty. When users enjoy a high level of interaction with the blog, they feel more motivated to share posts as they can comment on them with other users, with the author, within their social circle, or with others who simply share their interests—all of which creates loyalty toward the blog in question (Woisetschläger et al. 2008). This, combined with the influencing capacity of the blogger (Ashley and Tuten 2015), can lead to higher levels of intention to follow the blogger’s advice on the brands recommended via the blog (Lim et al. 2006). Although the literature identifies that the blogger’s practices aimed at interacting with readers lead to greater loyalty, to date, no studies have provided empirical evidence showing the effect of such practices on (a) intention to recommend the blog or (b) intention to purchase the brands highlighted on the blog. In light of the above, the following hypothesis is proposed:

H3.

The blogger’s interactive practices have a significant positive effect on (a) intention to recommend the blog and (b) intention to purchase the brands mentioned by the blogger.

The Moderating Effect of the Social Influence of Bloggers

Individuals habitually use reference groups when seeking accurate information to guide them in their responses to the social situations in which they find themselves (Cialdini and Goldstein 2004). Reference groups are defined as those groups of people who have a significant influence on the behavior of an individual (Bearden and Etzel 1982). Hence, social influence may be regarded as a psychological phenomenon from which people seek to generate a favorable reaction in others (Kelman 1958) so as to influence their opinions and recommendations (Yildiz 2014). The literature therefore recognizes the importance of social influence on consumer behavior (e.g., Narayan et al. 2011). In times gone by, social influence was exerted mainly by the people in an individual’s close-knit circle of friends and family. Now, thanks to the arrival of the Internet, it has also become the domain of people who are active online and who have not even had any previous interaction with users (Yildiz 2014). Indeed a person’s capacity to influence online can be examined simply via their activity level on social media, such as their Facebook posts or tweets (Westerman et al. 2012). Social influence, then, is a key variable in explaining the success of blogging and the interest that firms are showing in the use of blogs (Ashley and Tuten 2015). It is positively related to a number of variables linked to the behavior of blog readers, such as source credibility (Edwards et al. 2013), recommendation intention (Kerr et al. 2012), purchase decision-making online (Sridhar and Srinivasan 2012), and loyalty toward the brands featured on the blog (Castle et al. 2014).

Firms seek to identify Internet users with social influence who have the capacity to generate positive mentions for their brands (defined as impressions) (Li and Bernoff 2008). The value of impressions is derived from the scale of the network of the user who makes the comment and from the ease with which the information in question is disseminated (Kerr et al. 2012). Hence, the greater the capacity to generate impressions on social media, the higher the social influence, and vice versa. Recently, objective indicators such as the index developed by Klout (Klout Inc. 2015) have made it possible to quantify an individual’s level of social influence. According to this indicator, the greater the user’s influence on social media, the higher the Klout score (e.g., Ashley and Tuten 2015; Edwards et al. 2013), this being used in the case of blogs, for instance.

It may be the case, then, that the blogger’s level of social influence acts as a moderator in the relationships between their interactive practices (vis-à-vis their followers) and the consequences of these (in terms of credibility, intention to recommend the blog, and intention to purchase the brands associated with it). This is due to the fact that if the blogger has a high level of social influence, the consequences of his or her interactive activities will be magnified, to the extent to which they have a greater capacity to generate impressions, and vice versa.

The literature to date has not provided empirical evidence of this moderating effect of social influence on the relationships between the blogger’s activities and their consequences. Hence, the study of this phenomenon makes an important contribution both to the literature and also to the professional sphere as it highlights the value of using social influence indicators to assess the repercussions of firms’ use of bloggers to promote their brands. The following hypotheses are therefore submitted:

H4.

The effect of the blogger’s interactive practices, on (a) intention to recommend the blog and (b) intention to purchase the brands mentioned by the blogger, is moderated by his or her social influence—this effect being greater, the higher the level of social influence.

H5.

The effect of the blogger’s interactive practices on their credibility is moderated by his or her social influence—this effect being greater, the higher the level of social influence.

H6.

The effect of the blogger’s credibility, on (a) intention to recommend the blog and (b) intention to purchase the brands mentioned on the blog, is moderated by their social influence—this effect being greater, the higher the level of social influence.

Methodology

The Sample

One particular practice recommended by the literature in the development and validation of scales is to work with a homogeneous population (e.g., Cosenza et al. 2015). For the present work, it was important to select a sector in which the use of blogs is well established and that has set the standard for blogging; hence the fashion sector was chosen as the focus of the study as blogs play a major influencing role in the fashion industry (Ruiz-Molina 2013). As Spain is among the countries with the highest rates of adoption of fashion blogs (Ibañez et al. 2012), the present work focused exclusively on the Spanish national fashion blog reader population as the basis of the empirical research.

More specifically, the quantitative empirical study was aimed at habitual readers of at least one fashion blog over the course of the previous 6 months. Data collection was undertaken by means of a self-administered online questionnaire featured on a range of fashion blogs.

The final sample size was therefore 300, which was similar to that of other studies previously undertaken in the context of blogging (e.g., Wang et al. 2015), and comprised entirely of women (due to the nature of the sphere of application). This segment is of interest for the professional field and for the literature (e.g., Kuo et al. 2013) since the role of the personal blogger is fundamentally associated with women as they have been found to be more prone than men to write about their lives as a means of expressing emotions and releasing tension (e.g., Nardi et al. 2004).

Measurement Instrument

The blogger’s interactive practices were measured using the scale proposed by Laroche et al. (2012) which was a broader, adapted version of the original developed by Schau et al. (2009). To measure the blogger’s credibility, a scale was used based on the work of Ohanian (1991) and adapted to the fashion blog context. This scale has been used previously in studies such as those of Ballantine and Au Yeung (2015). To measure recommendation intention, a scale was used that drew on the work of Zeithaml et al. (1996) but, again, was adapted to the blog context, in a similar vein to the work of Cosenza et al. (2015). Intention to purchase the brands promoted via the blog was measured using a scale based on those of Laroche et al. (2012) and Casaló et al. (2007). The items on these scales were duly adapted to the blog context.

Lastly, to measure the social influence of the blogger, the Klout Index was used (Klout Inc. 2015). Klout is computed based on a user’s ability to drive action in social media. A user whose network consists of other influential users who respond to or share the person’s posts would have a higher Klout score. This information is collected on a daily basis, and each morning the user’s Klout score is updated. Klout scores have a range of 1–100, with a higher score indicating a higher level of influence.

The researchers differentiated between each of the two levels of the social influence variable depending on whether the blog in question presented a low Klout score (with values equal to or below 56—considered to represent a low level of social influence) or a high score (with values of over 56—considered to represent a high level of social influence).

The questionnaire in the present study captured readers’ opinions by means of a seven-point Likert scale, on which “one” equaled “totally disagree” and “seven” equaled “totally agree” for the scales relating to the blogger’s interactive practices, recommendation intention, and purchase intention; and a seven-point semantic differential scale was used to measure the “credibility” variable.

Findings

An SEM analysis was undertaken, covering the relationships between “blogger’s interactive practices,” “credibility,” “recommendation intention for the blog,” and “purchase intention for the brands mentioned in the blog.” First, the psychometric properties of the proposed model were estimated and evaluated. Since the test of multivariate normality of the variables included in the proposed model proved significant, it was necessary to undertake the estimation using maximum likelihood combined with the bootstrap method (Yuan and Hayashi 2003). In this case, a valid reference was the value of normed chi-square, which gave a value of 3.73—within the limits recommended by the literature. As regards the overall fit of the model, the GFI value was 0.89, and the RMSEA value was 0.09, slightly above that recommended in the literature. The incremental fit measurements were acceptable as they presented values above the threshold indicated by the literature.

Verifying that all the items used were adequate measures of their respective latent constructs, the composite reliability and variance extracted were found to be above the reference value, at 0.70 and 0.50, respectively (Hair 2012:649–651) (Table 1). Meanwhile, the dimensions considered in the model achieved discriminant validity; as in all cases, the correlation between them did not exceed 0.90 (Kline 2011:72) nor did their confidence interval include “one” (Steenkamp and van Trijp 1991). In general, the results obtained indicated that the set of dimensions proposed to measure the constructs provided an adequate level of reliability and convergent and discriminant validity.

Table 1 Convergent validity and internal consistency of scales used
Full size table

Next, the relationships between “blogger interactive practices,” “credibility,” “recommendation intention for the blog,” and “purchase intention for the brands mentioned on the blog” were analyzed. On the basis of the results, the following aspects are to be noted:

H1 proposed that the blogger’s interactive practices are an antecedent of recommendation intention for the blog (H1a) and of purchase intention for the brands promoted in it (H1b). The results show a statistically significant relationship (p < 0.01). The effect detected was 0.35, with a confidence interval of between 0.15 and 0.54 for the effect on recommendation intention, and 0.48, with a confidence interval of between 0.28 and 0.65, for the effect on purchase intention. Therefore there is statistical support for these hypotheses.

H2 proposed that the blogger’s interactive practices have a positive effect on credibility. The results show a statistically significant relationship (p < 0.01); therefore, there is empirical support for this hypothesis. Furthermore, the effect detected was quite marked (0.78), with a confidence interval of between 0.71 and 0.85.

H3a and H3b proposed that credibility has a positive influence on recommendation intention for the blog (H3a) and purchase intention for the brands showcased in the blog (H3b). The results show a statistically significant relationship (p < 0.01). The effect detected was 0.55, with a confidence interval of between 0.36 and 0.75 for the effect on recommendation intention, and 0.37, with a confidence interval of between 0.20 and 0.57, for the effect on purchase intention. Therefore there is statistical support for H3a and H3b.

With a view to establishing whether the aforementioned results applied across all bloggers or whether, by contrast, the effects depended on the extent of the blogger’s social influence, the moderating effect of social influence on the relationships within the research model (H4) was calculated. A multigroup SEM analysis was undertaken (with factorial invariance) in which the researchers differentiated between the two levels of the social influence variable according to the Klout score of the blog in question (a low score indicating a low level of social influence and a high score indicating a high level).

To this end, first the psychometric properties of the proposed model were estimated and evaluated. In this case, the normed chi-square gave a value of 2.46—within the limits recommended by the literature. As regards the overall fit of the model, the RMSEA value was 0.07, slightly above that recommended in the literature. The incremental fit measurements were also acceptable as they presented values above the threshold indicated by the literature.

Next, as outlined in the proposed model, the effect of social influence on the relationships between the variables was analyzed (H4a, H4b, H5, H6a, and H6b). On the basis of these results, the following aspects are to be noted:

For the group of bloggers with the higher level of social influence, it was found that the effect of the blogger’s interactive practices on credibility was 0.77 (with a confidence interval of between 0.66 and 0.87 and a p < 0.01). In turn, the blogger’s practices and credibility exerted a statistically significant effect on recommendation intention for the blog, with the two achieving a very similar value for their standardized coefficients (0.41 and 0.47, respectively, with a confidence interval of between 0.10 and 0.65 and a p < 0.05 for the blogger’s practices and a confidence interval of between 0.23 and 0.77 and a p < 0.01 for credibility). Finally, it was found that the blogger’s practices exerted a statistically significant effect on purchase intention for the featured brands, presenting standardized coefficients of 0.35, with a confidence interval of between 0.08 and 0.68 and a p < 0.05 and that credibility also influenced purchase intention, presenting standardized coefficients of 0.52, with a confidence interval of between 0.28 and 0.77 and a p < 0.01. These results demonstrate that in the group of bloggers with a higher level of social influence, the blogger’s interactive practices contribute to their credibility and that these practices and credibility can lead to higher levels of recommendation intention and purchase intention among readers.

For the group of bloggers with less social influence, it was found that the effect of the blogger’s interactive practices on credibility was 0.78 (with a confidence interval of between 0.68 and 0.87 and a p < 0.01). However, the blogger’s practices exerted no significant influence on recommendation intention for the blog (with a standardized loading of 0.29, a confidence interval of between 0 and 0.54, and a p-value of 0.23), while the blogger’s credibility did influence recommendation intention, achieving a standardized loading of 0.63, a confidence interval of between 0.37 and 0.89, and a p < 0.01. As regards the effect of the blogger’s interactive practices on purchase intention for the featured brands, a significant effect was identified, presenting a standardized loading of 0.62 (with a confidence interval of between 0.33 and 0.87 and a p < 0.01); and, lastly, the effect of credibility had no significant influence on purchase intention (the standardized loading was 0.20, with a confidence interval of between −0.06 and 0.48 and a p-value of 0.23). These results show that in the case of bloggers with a lower level of social influence, the blogger’s interactive practices contribute to their credibility. They can thus achieve a greater level of recommendation intention only through this credibility and a greater level of purchase intention only through their interactive practices.

These findings provide empirical support for H4a and H6b, as when bloggers progress from having a low level of social influence to a high level, they also progress from a scenario in which their interactive practices do not contribute to generating recommendation intention for the blog to one in which they do contribute to doing so (H4a) and from a scenario in which their credibility does not contribute to achieving greater purchase intention for the brands promoted on the blog to one in which their practices do achieve this higher level of purchase intention (H6b).

However, the rest of the hypotheses on the moderating effect of social influence (H4b H5 and H6a) find no empirical support. This indicates that, regardless of the level of social influence achieved by the blogger, their interactive practices will contribute to purchase intention for the featured brands and to their personal credibility, while their credibility will contribute to recommendation intention for the blog.

Conclusion and Implications

An increasingly competitive, operating environment is leading firms to consider new strategies for achieving customer loyalty, including the use of blogs. This growing interest in the blogging medium calls for indicators to help firms select the most effective bloggers, and at the same time, the bloggers themselves must identify strategies for improving their position within the market.

The present work identifies interactive practices that the blogger can implement to interact with his or her public as a means of securing their loyalty toward the blog (recommendation intention) and toward the brands promoted on it (purchase intention) while achieving greater credibility for themselves. The work also analyzes the moderating effect of the level of the blogger’s social influence on loyalty formation among followers and on his or her credibility. A quantitative empirical study was undertaken, resulting in the following key findings:

When the blogger undertakes effective practices to interact well with their public, these practices have a positive influence on their performance in the market, contributing to enhancing their credibility with their readership and generating greater loyalty toward the blog (recommendation intention) and toward the brands mentioned on it (purchase intention). This finding is in line with those of previous literature that asserts that the blogger’s interactions with their followers have to be perceived and valued by them (Dennen 2014; Hsu et al. 2014); but it offers a new contribution with its empirical measurement on the effects of the blogger’s interactive activities on the behavior of the blog’s followers.

A further contribution of the present work relates to the use of the Klout indicator (Klout Inc. 2015) to objectively measure the social influence of the blogger. The study set out to determine whether the blogger’s social influence, as a variable, is able to magnify the effect of their practices on their credibility and on the loyalty of their followers (to the blog and to the brands featured on it).

Results were mixed, showing that social influence moderates the relationship between the blogger’s interactive practices and intention to recommend the blog and between the blogger’s credibility and purchase intention. It was found that if the blogger wishes to maximize the benefits of the positive effects of his or her interactive practices and credibility on reader behavior, they should seek to achieve a high level of social influence. Should this not be achieved, and the level of social influence remains low, the positive effects of their practices and their credibility on reader behavior will only be partial.

Implications for the Sector

The conclusions reached in this study hold some interesting implications for those actively working in the blogosphere and for firms looking to use blogging as a promotional tool and select bloggers to showcase their brands. Blogging is becoming increasingly professionalized (Ruiz-Molina 2013) and thus requires effective methods for evaluating its value and competitive capacity within the market.

First, a competitive approach is put forward (in the form of blogging practices for interacting with followers, blogger credibility, or a combination of the two). Applying this approach will enable bloggers to achieve better outcomes in terms of loyalty among followers of the blog (recommendation intention) and toward the brands mentioned on the blog (purchase intention).

These interactive practices are based on responding to reader queries, taking a personalized approach to such responses and achieving a good level of social interaction. Thus, if firms wish to bring their brands closer to the public, blogging may be a valid option for generating brand loyalty.

Next, it has been demonstrated that the blogger’s credibility has a positive influence on the loyalty of their followers. Such credibility is linked to the appeal, honesty, and expertise that he or she conveys to their public. It is these characteristics that the blogger must endeavor to foster and communicate if they wish to secure the loyalty of their readers.

Third, the work highlights the relevance of social influence, in this context referring to the capacity of the blogger to reach users and generate impressions online. The Internet offers tools to quantify this capacity, such as the Klout Index (Klout Inc. 2015). The present study affirms that this index is a good indicator of the blogger’s capacity to achieve results such as loyalty formation among followers relative to the blog and also loyalty toward the brands promoted on it. Thus, bloggers with a higher Klout score have a greater capacity to secure the loyalty of their public than bloggers with a lower score.

It is to be noted that the maximum positive effect of blogger interactive practices combined with credibility is achieved when he or she has a high level of social influence as, in this scenario, both the practices and the credibility will exert a positive influence on recommendation intention for the blog and purchase intention for the featured brands. Conversely, when the blogger’s level of social influence is low, the effects of their interactive practices and their credibility on recommendation and purchase intention are only partial.

Limitations and Future Lines of Research

One such limitation is imposed by the variables included in the research model. This study includes variables considered the most relevant for achieving the objectives set for this work. It would be of interest, for example, to incorporate other variables such as a greater number of effects that the blogger’s interactive practices and credibility may exert on other performance measures of the blog (such as revisit rates, continuous participation in the blog, or a higher level of participation in the blog among its followers).

A further limitation of this work lies in its geographical scope of application. A future line of research potentially of interest is the application of the proposed research model to other geographical areas.