Firmâ s Social Media Efforts, Consumer Behavior, and Firm ... · the impact of social media on customers’ purchase behavior, they are based on data from a single firm and do not

Thirty Fifth International Conference on Information Systems, Auckland 2014 1

Firms’ Social Media Efforts, Consumer Behavior, and Firm Performance

Completed Research Paper

Sunghun Chung Desautels Faculty of Management

McGill University 1001 Sherbrooke Street West

Montreal, QC H3A 1G5 Canada [email protected]

Animesh Animesh Desautels Faculty of Management



Kunsoo Han

Desautels Faculty of Management McGill University

1001 Sherbrooke Street West Montreal, QC H3A 1G5 Canada

[email protected]

Alain Pinsonneault Desautels Faculty of Management



Abstract

Despite the increasing attention paid to the business value of social media, it is still not clear how they affect firm performance. This study theorizes and empirically examines how firms’ social media efforts—in terms of intensity, richness, and responsiveness—influence consumer behavior (engagement and attention) and firm performance. Using detailed data collected from the Facebook pages of 63 firms over the 2010-2012 period, we find that the richness and responsiveness of a firm’s social media efforts are significantly associated with the firm’s market performance, captured by abnormal returns and Tobin’s q. Interestingly, the intensity of a firm’s social media efforts is not significantly associated with firm performance. We also find that not only do consumer engagement and attention directly impact firm performance, but they also mediate the relationship between a firm’s social media efforts and firm performance. Unlike prior studies that examine the impact of third-party or consumer-initiated social media, such as blogs and consumer ratings, our study focuses on estimating financial returns to firms’ own efforts on firm-initiated social media, thereby assessing the business value of social media directly.

Keywords: social media, firms’ social media efforts, firm performance, consumer engagement, consumer attention, business value of social media


Introduction

Social media applications have the potential to provide business value by enabling the formation of online customer communities that can support branding, sales, customer service, and product development (Culnan et al. 2010). In order to tap this potential, many organizations are actively utilizing social media, such as Facebook and Twitter, in communicating and interacting with their current and potential customers (Ling 2013). 1 For example, many firms have official Facebook “pages” where they post messages for their customers (e.g., product/service promotions) and respond to messages left by their customers (“fans” of a certain page). Recent industry reports suggest that social media may positively impact businesses’ bottom line and brand image; as a result, firms need to invest in social media and integrate social media into their business strategies (Hardy 2012; J.D.Power 2013; Wessel 2011). However, many practitioners remain skeptical about the returns to investments in social media because of the low reach rate of firms’ social media presence and the difficulty in measuring returns (Dyer 2013; Mueller 2011). Despite the fact that firms are increasingly investing in and using social media, relatively little research has been conducted to examine the returns to firms’ social media efforts. In particular, to our best knowledge, prior studies have neither considered the different dimensions of firms’ social media efforts, nor have they examined the differential effects of these dimensions on firm performance. Further, extant research has not explicated the mechanisms through which firms’ social media efforts can influence firm performance. As a result, there is limited guidance for firms on how to use social media strategically—especially with respect to the particular dimensions of social media efforts they should emphasize— when communicating and interacting with their customers in order to maximize returns to their social media efforts.

We conceptualize Intensity, Richness, and Responsiveness as the three dimensions of a firm’s social media efforts and examine the relationship between these dimensions and the firm’s market performance. In order to uncover the mechanisms through which firms’ social media efforts impact firm performance, we also consider the mediating role of consumer engagement (captured by customers’ various activities, including postings, comments, and “likes” on a firm-initiated social media site) and consumer attention (captured by web traffic on the firm’s website). Based on detailed data collected from 63 firms’ Facebook pages and the firms’ financial data over the 2010-2012 period, we find that the impact of the richness and responsiveness of a firm’s social media efforts on the firm’s market performance is positive and significant. Interestingly, we do not find a significant impact of the intensity of a firm’s social media effort on the firm’s market performance. Further, our results suggest that consumer engagement and attention not only directly impact firm performance, but also mediate the relationship between a firm’s social media efforts and performance.

This study makes several contributions to research. First, while recent studies (Luo and Zhang 2013; Luo et al. 2013) have examined the impact of third-party- or consumer-initiated social media (blogs and consumer ratings) on firm equity value, our study focuses on firm-initiated social media (Facebook pages in our case), and empirically examines whether a firm’s social media efforts matter for the firm’s market performance. Although a number of studies have examined the outcomes of social media (e.g., Goh et al. 2013; Miller and Tucker 2013; Rishika et al. 2013; Tirunillai and Tellis 2012), our study, to our best knowledge, is among the first to quantify the financial returns to firms’ social media efforts. In particular, we contribute to this nascent literature by conceptualizing Intensity, Richness, and Responsiveness as key dimensions of a firm’s social media efforts, by measuring them from the firm’s actual activities on its Facebook page, and by estimating the differential effects of the three dimensions on firm performance. Second, we uncover the mechanisms through which firms’ social media efforts impact firm performance by examining both direct and indirect (through consumer engagement and attention) relationships. Whereas prior studies have mostly focused on the direct relationships between two constructs (e.g., social media – consumer behavior, social media – firm performance), we comprehensively examine the structural relationships among firms’ social media efforts, consumer engagement and attention, and firm performance, collected from multiple data sources. Such a comprehensive examination of the structural relationships among these variables allows us to derive richer insights.

1 As of 2013, there were approximately 18 million business pages on Facebook, and 1 million are being added every month.

Firm’s Social Media Efforts and Firm Performance


Our study also contributes to practice. Our findings help managers gauge financial returns to firms’ social media efforts and underscore the importance of investing in firm-initiated social media to communicate and interact with consumers. In particular, our finding that firms’ social media efforts significantly contribute to both short- and long-term market performance suggest that firms should carefully formulate their social media strategies. Specifically, our results suggest that firms need to put more efforts into providing rich information through social media and actively responding to consumers’ messages.

Literature Review

A handful of studies that have examined the impact of social media can be categorized into two streams. One stream of literature has focused on the impact of social media on firm performance. For example, by employing event study methodology, Yang et al. (2012) showed that a firm’s decision to engage in a social network platform positively affects the firm’s market value. Researchers have also examined the impact of consumers’ actions in social media on firm performance. Notably, Luo et al. (2013) found that consumers’ action in social media (i.e., posting content about a firm on blogs, as well as posting reviews about the firm and its products on the website of CNET.com) can influence a firm’s market value. Although these studies provide valuable implications by highlighting the importance of social media in a firm’s strategy, they do not shed light on the returns to firms’ specific social media efforts per se. Extending this stream of research, our paper proposes three important dimensions of a firm’s social media efforts (i.e., intensity, richness, and responsiveness), and examines the relationship between these dimensions and the firm’s market performance.

Another stream of literature has focused on the impact of social media on consumer behavior. Based on individual-level data regarding social media participation (Facebook page) and transactions, Rishika et al. (2013) find that customers’ social media participation increases their website visit frequency and profitability. In a similar vein, Goh et al. (2013) analyze user- and firm-generated content on a firm’s Facebook page, and find that user-generated content has a stronger effect on purchase behavior, compared to firm-generated content. Although these studies provide important implications regarding the impact of social media on customers’ purchase behavior, they are based on data from a single firm and do not examine the different dimensions of firms’ social media efforts. Further, these studies do not examine the comprehensive relationships among firms’ social media efforts, consumer behavior, and firm performance.

In general, active firm promotion involving word of mouth has been found to be effective in managing firms’ social presence (Godes and Mayzlin 2009). Although previous studies provide several ways of measuring the usefulness of user-generated content to firms (e.g., Sawhney et al. 2005), as well as the theoretical implications of firms’ efforts on online communities (Dellarocas 2006), they have not examined the effect of firms’ social media efforts and the subsequent actions of customers, with the exception of Miller and Tucker (2013). The authors show that active social media management drives more user-generated content in a healthcare setting. Specifically, by using cross-sectional data for 5,035 hospital Facebook pages, they find that firms (e.g., hospitals) who actively manage their Facebook pages through regular posts are likely to induce more postings from internal employees, compared to external customers. Their findings suggest that active management of social media does not lead to the customer-side benefits that firms put forward before adopting social media strategies. Although their study empirically suggests a strategic benefit to firms who decide to actively participate in social media, only the use of social media in healthcare was considered; other metric types of customer-generated content frequency, such as comments and Likes, were not examined to explain customer-side benefits. Our study extends this stream of literature by examining the effect of a firm’s social media efforts on the subsequent actions of the firm’s customers, as well as its financial performance.

Theory

This study builds on prior literature that has looked at the impacts of social media on firm performance and consumer behavior. As depicted in Figure 1, we first examine the direct relationship between firms’ social media efforts (in terms of intensity, richness, and responsiveness) and firm performance (RQ1). Further, we propose two key intermediate processes – consumer engagement (RQ2) and consumer

Social Media and Digital Collaborations

4 Thirty Fifth International Conference on Information Systems, Auckland 2014

attention (RQ3) – that mediate the relationship between firms’ efforts and their financial performance. Finally, we also investigate the relationship between consumer engagement and attention (RQ4).

Figure 1. The Conceptual Model

Value of Firms’ Social Media Efforts

Social media can significantly enhance firms’ ability to interact with customers, which leads to fundamental changes in terms of ease of contact, volume, speed, and the nature of these interactions (Culnan et al. 2010; Gallaugher and Ransbotham 2010). Firm-initiated social media sites allow firms to easily disseminate information on new products/services, promotions, customer relationships and corporate social responsibility. With the growing impact of social media on business, there is a heavy managerial emphasis on studying the returns to investments in firms’ social media efforts. A firm’s efforts to communicate through social media will affect its market value, as the information disseminated can affect the stock price of the firm by providing signals about the marketing strategy, customer service quality, and growth potential of the firm to potential and current investors.

Although firms provide information to investors through financial statements and other disclosures, such as annual reports, numerous studies have suggested the existence of asymmetric information between firms and investors (Healy and Palepu 2001). To overcome this asymmetry, investors seek additional information from multiple sources, such as new product announcements (Chen et al. 2011), expert reviews (Tellis and Johnson 2007), and user-generated content (Tirunillai and Tellis 2012). Communication between a firm and its consumers on social media provides one such source of information to investors, who can infer the firm’s level of involvement with its customer base, and the firm’s commitment to improving its relationships with customers from its social media efforts.

Firms’ social media efforts provide consumers with easy access to various types of useful information, including firms’ offerings, promotions, and actions (e.g., CSR activities). Such information can easily spread to consumers and potential investors, thereby engendering greater trust and customer loyalty (Algesheimer et al. 2005; Thompson and Sinha 2008). Thus, a firm’s social media efforts can affect product/service sales by influencing the awareness, attitudes, perceptions, purchase decisions, and brand loyalty of the potential and current customers who receive the messages. To the extent that active interactions between a firm and customers influence the product awareness and/or likelihood of sales, a firm’s social media efforts serve as a good signal to indicate its sound customer relationship management; therefore, investors’ valuation of the firm’s stock will be associated with the quality and quantity of firms’ social media activity. However, it is an open empirical question as to whether there is valuable information contained in firms’ communication exchanges on social media. Therefore, we examine the following research question:

RQ 1: Does a firm’s social media efforts (in terms of the quantity and quality of information posted by the firm on a firm-initiated social media site) influence its market performance?

Specifically, in order to examine the differential effects involving the different dimensions of firms’ social media efforts, we identify three dimensions of a firm’s efforts on a social media site: 1) the intensity of the firm’s efforts (i.e., the volume of messages posted by a firm); 2) the richness of the firm’s efforts (i.e.,



the information richness of messages posted by a firm); and 3) the responsiveness2 of the firm’s efforts (i.e., the extent to which a firm responds to consumers’ messages).

The first dimension captures the quantity of effort, while the other two represent the quality of the effort. Higher intensity (i.e., a greater amount or volume of a firm’s postings and comments) would mean more opportunities for the firm’s current and potential customers to see and act on the information disseminated by the firm. By actively managing their social media sites through regular posts, firms not only induce more user-generated content (Miller and Tucker 2013), but also identify issues and foster innovation (Gallaugher and Ransbotham 2010). To the extent that the volume of the information provided by a firm can increase customers’ awareness of and engagement with the firm, the intensity of a firm’s efforts is likely to influence its market value. Therefore, we examine the following question:

RQ 1a) Does the intensity of a firm’s efforts (i.e., the volume of messages posted by a firm) on a social media site affect its market performance?

Our second dimension, the richness of a firm’s social media efforts, represents the richness of the messages and comments posted by the firm. In the context of social media, posting messages serves as an important medium to convey information. Messages delivered through different media - texts, pictures, or videos - have varying abilities to deliver information, and accordingly, we can determine the richness of these various types of media (Daft and Lengel 1986). Richer messages are more likely to be noticed and propagated (i.e., through message sharing) by consumers because they are more engaging and informative. Research suggests that messages delivered using pictures are richer than text, as they require less processing effort (Larkin and Simon 1987). Further, it has been shown that video is superior to static pictures because it is more explicit and easier to understand (Park and Hopkins 1992). It can be argued that as one moves from text to pictures, and from pictures to video, more information cues can be communicated. To the extent that such additional information cues lead to engaging and valuable content that can enhance the social engagement of prospects, thereby differentiating the firm from competitors and boosting customer awareness and loyalty (Emerson 2012b), the richness of a firm’s social media efforts will be reflected in the firm’s market value. Therefore, we study the following question:

RQ 1b) Does the richness of a firm’s efforts (i.e., the informational richness of messages posted by a firm) on a social media site affect its market performance?

Finally, firms’ social media efforts can be distinguished in terms of the nature of the participation. A firm generally broadcasts messages to convey information to its subscribers. However, a firm may also choose to interact with its customers by responding to customer comments. Such responsiveness (or interactivity) tends to increase social presence and enhance the depth of information sharing (Miranda and Saunders 2003). Although responsiveness comes at a relatively high cost, as it requires customized messages rather than the same message broadcasted to all subscribers, being responsive (as well as being perceived as responsive) is critical to survival and to building strong social capital on a social media platform (York 2012). However, the value of responsiveness has not been empirically established. Therefore, we investigate the following question:

RQ 1c) Does the responsiveness of a firm’s efforts (i.e., responding to consumers’ messages) on a social media site affect its market performance?

Role of Consumer Engagement as a Mediator

Firms’ actions on firm-initiated social media can be highly visible and contagious because the information posted by a firm is immediately delivered to current subscribers (fans) of the firm’s page through its news feed, and non-fans who are friends of current fans can also observe activities and interactions on the page involving their friends (Gallaugher and Ransbotham 2010). The information that a firm intends to convey

2 Note that in our context (i.e., Facebook pages), all of the messages that originate from firms are positive, and more than 95% of customer comments are also positive in nature (e.g., Rishika et al. 2013). This is because on firm-initiated social media sites, most of the messages originate from the firm, and customer comments are in response to these messages. Thus, they generally take the form of positive comments or general queries. This is different from third-party websites, where customers may post product reviews and recommendations that can be either positive or negative.



to its audience can be diffused through two modes – the subscription mode and the social contagion mode. In the subscription mode, messages are received by the consumers who subscribe to the firm’s news feed (i.e., those who “like” the firm’s page). On the other hand, in the social contagion mode, the message reaches those consumers who are not subscribers, but are directly connected to the subscribers within the social network due to consumer engagement (i.e., specific actions such as posting messages, liking or commenting on a post on a firm’s social media page). Consumer engagement is essential for creating the social contagion effect, whereby the awareness of a firm’s brand and its message reaches non-subscribers or those not participating in a conversation with the firm.

The direct effect of a firm’s social media efforts on its market value, which we focused on earlier, is primarily due to the propagation of information through the subscription mode. However, to the extent that the social contagion effect is significant and leads to the viral propagation of messages to those who otherwise would not have received it, we expect that the effect of a firm’s social media efforts on its market value will be partially mediated by consumer engagement.

In other words, by actively managing its social media site through regular posts (i.e., a high volume of messages posted by a firm), providing more informative information (i.e., focusing on multimedia messages, such as pictures and videos), and frequently interacting with its customers (i.e., responding to customer comments), a firm not only sends a strong signal of its health to its current and potential investors (which should be reflected in its market value), but also increases the possibility that existing (via the subscription mode) and potential customers (via the social contagion mode) engage in the firm’s social media site through their postings, comments, and “Likes.” Specifically, network relations between a firm and customers on a firm-initiated social media site are socially embedded and entail expectations and norms associated with social attachment from the interactions. In these relationships, the social capital (Coleman 1988) generated tends to take the form of mutual trust and goodwill (Adler and Kwon 2002), resulting in a high level of consumer engagement.

In addition to signaling the viral propagation and consequent increase in the market reach of a firm’s message, higher consumer engagement also signals customer satisfaction. Prior studies show that that customer satisfaction is positively associated with firm value (Anderson et al. 2004; Fornell et al. 2006; Gruca and Rego 2005). These studies suggest that customer satisfaction can lead to additional purchases and can be an economic asset that is expected to influence firm value with high returns and low risk (Fornell et al. 2006). In the social media context, comments on a firm’s postings and positive reactions such as “Like” can serve as positive word of mouth (WOM), which is expected to influence firm performance. Consumer engagement in a social media site tends to signal repeat buying, increased market share and long-term growth of the firm, and tends to reduce customer complaints and price elasticity. Therefore, consumer engagement is likely to impact firms’ market value. However, assessing the value of consumer engagement in a firm’s social media site, and mediating the role of consumer engagement between a firm’s social media efforts and its market value is an open empirical question. Therefore, to assess the mediating role of consumer engagement as described above, we examine the following question:

RQ 2a) Does a firm’s social media efforts (i.e., intensity, richness, and responsiveness) affect its consumer engagement (i.e., “Likes,” comments and posts by consumers on the firm’s Facebook page)?

RQ 2b) Does consumer engagement affect firm performance?

Role of Consumer Attention as a Mediator

Consumers generally have to visit a firm’s website to gain information and to process their transactions with the firm. Therefore, one of the primary objectives of a firm having an online presence is to attract the attention of consumers to its website. A firm employs various marketing channels to reach potential consumers, attract them to visit its website, and then convert the website visitors into buyers. The amount of website traffic reflects consumers’ interest in and attention to the firm. Therefore, website traffic acts as a valuable signal to investors about the market share and growth of a firm (Luo and Zhang 2013; Rajgopal et al. 2000; Trueman et al. 2000). In addition, some of the website visitors may be current or potential investors who access the website to obtain information relevant for making investment decisions. Thus, website traffic may also be associated with the number of investors interested in the firm. To the extent that website traffic, also referred as the “eyeball effect,” contains a valuable signal to the investor



community and complements the accounting data (Demers and Lev 2001; Dewan et al. 2002; Luo and Zhang 2013; Trueman et al. 2000), website traffic will be positively associated with market value.

A firm’s effort spent on social media is one of the various ways to increase consumer attention, which is typically captured by website traffic. By providing informative contents and by actively interacting with its customers (i.e., focusing on intensity, richness, and responsiveness), a firm creates opportunities to direct its current and potential customers who subscribe to the social media site (e.g., Facebook page) to its website, thereby increasing consumer attention (Luo and Zhang 2013). Similar to the role of consumer engagement, consumer attention may be influenced by the social contagion effect. Thus, the effect of a firm’s social media efforts on its market value will be partially mediated by consumer attention. Although proponents of social media advocate that firms should spend significant resources and should actively manage their social media channels to increase their website traffic (Emerson 2012a; Vahl 2014), the empirical evidence connecting firms’ social media efforts to website traffic is lacking. Therefore, we examine the impact involving the different dimensions of firms’ social media efforts on website traffic. Specifically, we assess the mediating role of consumer attention by examining the following question:

RQ 3a) Does a firm’s social media efforts (i.e., intensity, richness, and responsiveness) affect consumer attention (i.e., website traffic)?

RQ 3b) Does consumer attention affect firm performance?

Relationship between Consumer Engagement and Consumer Attention

Consumer engagement not only affects firms’ market value directly, but can also indirectly contribute to market value through its impact on consumer attention. As discussed earlier, social media not only create new firm-customer interactions; they also expose these interactions to others, thus creating social contagion (Aral and Walker 2011; Bond et al. 2012; Ugander et al. 2012). As more customers comment on and like a firm’s messages on its social media site, the messages will become more visible to more people through social contagion, which can increase people’s awareness of the firm and may induce more people to visit the firm’s website. In this sense, consumer engagement (e.g., “Likes” and comments) on a firm’s social media site can function as a credible word-of-mouth channel. As the information comes from one’s friends or friend’s friends, people are more likely to trust the information, compared to information from other sources (e.g., TV or Internet ads). This, in turn, can attract their attention to the firm and increase the likelihood for them to visit the firm’s website to gain more information or conduct transactions there. However, there is a lack of empirical research examining the presence and strength of the relationship between consumer engagement and consumer attention. Therefore, we empirically examine the following question:

RQ4) Does consumer engagement affect consumer attention?

Data and Methods

Facebook Pages as Firm-Initiated Social Media

Facebook is an online medium that allows users to interact with one another or with firms by sharing a variety of information. As of March 31, 2014, Facebook is the largest online social network, with 1.28 billion active users, surpassing other online social networks, such as MySpace and Twitter.3 Facebook introduced the ‘‘Page’’ application in November 2007 as one of its basic features, and users share their information by ‘‘liking” firms (or becoming a “fan”) and allowing firms to access their profiles. When someone “likes” a page, they will see updates from that page through the “news feed” feature.

A firm’s Facebook page is set up by the firm, allowing users to share information with the firm and other users. Facebook pages also have the ability to connect individuals who are not “fans” of the firm, yet share common interests. In this sense, Facebook pages are an official firm-initiated online forum for communication and interaction, not only between a firm and its current/potential customers, but also among current/potential customers. Specifically, a firm typically delivers announcements, pricing, and new products/services by posting status updates (i.e., text), photos, or video clips and shares external

3 Source: http://newsroom.fb.com/company-info/



content by using the links function. Customers may post questions through status updates or comment threads and may share their experiences with photos and video clips. The fans of a firm’s page can “like” or comment on a post by the firm or other fans to express their opinions and positive attitudes; then, the action is broadcast to the fans’ Facebook friends through the news feed, thereby increasing the page’s exposure and reach.

Sample and Data Collection

We chose South Korea’s top 100 firms in terms of the total market value, sales, and net income as our sample for two reasons. First, focusing on large firms can ensure active interactions between firms and consumers because larger firms’ Facebook pages are likely to be updated and accessed more frequently. Second, the top 100 firms are likely to invest more in their social media, compared to smaller firms (due to relatively higher levels of advertising expenditures). Our sample firms from diverse industries are publicly traded and have, on average, 17,210 fans who “like” their Facebook pages. Our initial sample contained 8,324 firm-week observations from 100 firms over the period from January 1, 2010 to December 30, 2012. We removed 2,514 observations with missing quarterly R&D spending and 1,602 observations with missing or zero advertising expenditures. Hence, the final sample consists of 4,208 observations from 63 firms spanning 100 weeks.

Our data come from two major sources. First, we collected detailed information on all activities (e.g., posts, comments, and “Likes”) from 100 firms’ official Facebook pages over a period of one hundred and fifty weeks. Our Facebook data were collected by our custom-designed distributed computing platform, the Apache Hadoop Hive-based crawler. By using thirty Linux x86 servers running in parallel, data-mining agents queried the Facebook servers in order to acquire the specific posts and comment information on each firm’s Facebook page.4 Specifically, we collected all available data through the Facebook graph API (Application Programming Interface) for each firm. It is important to note that we could obtain the timestamp of the posts/comments and the anonymized unique user identification numbers. Similar to Da et al. (2011), we aggregated the data at the weekly level for each firm and constructed a panel dataset containing each firm’s weekly social media activities and consumers’ weekly responses. Please note that we do not use daily data because the variation of firms’ social media activities at the daily level was relatively small (e.g., the average number of postings was 3.21, and the standard deviation of postings was 3.57). Because some firms operate several Facebook pages for different purposes (e.g., brand page, CSR page, promotion page, and so on), we manually tracked every official Facebook page for each sample firm. We then aggregated all activities from the multiple Facebook pages for each week of our sample period.

Second, since all sample firms are listed on the Korea Stock Exchange (KSE) or Korean Securities Dealers Automated Quotations (KOSDAQ), we obtained quarterly firm-level financial data from Korea Investors Service, Inc. To match those financial variables with our weekly social media metrics, we followed the data-coding scheme used in prior studies (Luo 2009; Pauwels 2004). Specifically, we assigned the same values of earnings forecasts, cash flows, and other financial data to all 12 weeks contained in a given quarter. We checked the robustness of our results by using VAR-bootstrapping specifications with 5,000 simulated databases (Luo et al. 2013; Statman et al. 2006) and confirmed that our results are not sensitive to temporal aggregation.

Firm’s Social Media Efforts and Consumer Engagement

Recall that we conceptualized three dimensions of firms’ social media efforts: Intensity, Richness, and Responsiveness. To capture the Intensity of a firm’s social media efforts, we use two measures of firms’ Facebook activities: the number of postings and the number of comments. Using principal component analysis (PCA), we extracted one factor from the two measures, and the pattern of factor loadings supports the existence of a single dimension. In the main analysis, we use a weighted average of the number of a firm’s postings and comments in week t to measure Intensity (weighted by their loadings in

4 This study provides an initial step toward analyzing large-scale unstructured (commonly called big data) social media data to investigate the relationship between consumer behavior and its outcomes. By doing so, our dataset about consumer behaviors was constructed and standardized into panel data.



the underlying principle components). We then scaled it by the network size, which is the total number of Facebook fans of a focal firm in week t to adjust for the network effect and the ripple effect.5

The richness of a firm’s social media efforts was measured as the ratio of the number of the firm’s enriched postings (e.g., flash, videos, photos, and music) to the total number of the firm’s postings in week t. A larger value of this measure reflects the firm’s greater efforts spent on social media in terms of the richness of information provided to consumers. The responsiveness of a firm’s social media efforts was measured by the ratio of the number of the firm’s comments to the total number of comments made by both the firm and its fans in week t. It reflects the magnitude of the firm’s reciprocal action toward its customers through comment threads following a certain post. As shown in Table 1, on average, a firm posts or comments about 37.4 times per week; approximately 57% of them contain enriched information; and about 15% of the comments are made by firms. To capture consumer engagement, we use a weighted average of the number of postings, comments, and “Likes” by customers in week t (weighted by their factor loadings in the underlying principle components) as a single-item measure.6

Firm Performance

As our main dependent variable, we use abnormal returns that capture a firm’s market performance beyond the firm’s expected stock market returns. To estimate abnormal returns, we use the extended Fama-French model (Fama and French 1993; Fama and French 1996). The Fama-French model has been recognized as a good proxy for stock market performance by considering market risk, firm size, as well as multiple value factors (Luo et al. 2013; Tirunillai and Tellis 2012). We thus follow the extended Fama-French model:

Rit - Rft = β0i + β1i (Rmt - Rft) + β2i SMBt + β3i HMLt + β4i MOMt + εit, (1)

where Rit is the returns for firm i on week t7; Rft is risk-free rate; Rmt is average market returns; SMBt is the small-minus-big capitalization factor; HMLt is the high-minus-low book-to-market equity factor; and MOMt is the momentum factor in the given time period. Stock price data are obtained from Korea Investors Service, Inc., and data for Fama-French factors and momentum (Rft, Rmt, SMBt, HMLt, and MOMt) are obtained from FnGuide, Inc. (https://www.fnguide.com/SNI/SNI_FactorModel.asp). We ran Equation (1) for a rolling window of 250 trading days prior to the target day.8

The abnormal returns are then calculated as the difference between the actual returns and the expected returns. The mean value of weekly abnormal returns is 0.141%.9

Consumer Attention

As mentioned earlier, we use Web traffic to capture consumer attention to a firm. As similar to Facebook pages, most of firms operate several Websites for different purposes, we manually tracked every official Websites for each sample firm. We then collected Web traffic data from Nielsen Korean Click and aggregated all Web traffics from multiple Websites for each week of our sample period. Specifically, we obtained the following weekly measures of Web traffic for our sample firms: unique visitors (the number of distinct individuals requesting pages from the Website), page views (the total number of page views), reach rate (the ratio of the number of visitors to a firm’s website to the total Internet users in South Korea tracked by Nielsen Korean Click), and average visit duration which can be traffic metrics of time duration. Our use of these measures of Web traffic is in line with prior literature (Demers and Lev 2001; Luo et al.

5 Without scaling, we obtained similar results. 6 Along with the intensity of a firm’s social media effort, the pattern of factor loadings supports the existence of two dimensions corresponding to a firm’s intensity (loadings for # of firms’ postings and comments: 0.891 and 0.827, respectively) and consumer engagement (loadings for # of customers’ postings, comments, and likes: 0.890, 0.810, and 0.870, respectively). Subsequent confirmatory factor analysis (CFA) supports the overall validity of the two-factor model (Chi-square = 167.39; GFI = 0.95; CFI = 0.93; NFI = 0.94). Cronbach’s alpha for the two factors is above 0.6, suggesting acceptable reliability for exploratory analysis. In addition, customers tend to prefer comments (M = 8.75) compared to postings (M = 354.38) on firms’ social media sites. 7 Stock returns are calculated based on the weekly closing price adjusted for corporate actions such as dividends. 8 To avoid computational burden, we use a grid search procedure (e.g., 50, 100, 150, 200, 250, and 300 days) and find that a rolling window of 250 trading days (i.e., 50 weeks) provides the best fit. 9 Later, we use quarterly Tobin’s q as an alternative dependent variable for a robustness check.



2013; Trueman et al. 2000). Based on these metrics, we use page views per user, reach rate, and average visit duration to measure Web traffic for the empirical model. Among different Web traffic measures, reach rate which is more comparable across firms (e.g., Luo et al. 2013) has a best fit in our empirical estimations. We thus report our results with reach rate as consumer attention.10

Control Variables

We include a number of control variables. To control for the effect of Network Size, we include the total number of Facebook fans11 of a focal firm in week t. We include average Post Length of firm in week t (unit: bytes) to control for the amount of information. We also control for the Influence Duration of a firm’s post by including average time gap between the time when a firm posts a message and the time when the last comment to the message is posted in week t (unit: days). On average, a firm has about 4,800 Facebook fans and length of posting is about 200 bytes. Also, the mean value of influence duration of firm’s postings is 0.73 days. Because Web traffic to a firm’s website can be influenced by the intensity of Internet search related to the firm, we control for Internet Search. We obtained Internet search data from Google trend. Following Luo et al. (2013), we measure search intensity by the mean of “firm key words”12 search frequencies at google.com in week t. The mean search intensity is 40.32 a week.

To control for the growth opportunities available to a firm, we include R&D Intensity, computed as the total R&D expenditures in a given quarter, divided by total sales. We also include various firm-level measures: Total Asset (to control for firm size), and Advertising Intensity (the ratio of advertising expenditures to total sales), Operating Margin (ratio of net income to total sales), Leverage (the ratio of long-term debt to total sales), and Liquidity (the ratio of current liability to current asset). At the industry level, we control for industry Concentration by including the Herfindahl-Hirschman Index (HHI) (Han and Mithas 2013). Finally, industry dummies were used to control for time-invariant industry-specific effects. Industries correspond to Korean standard industrial classification (KSIC-9) narrow level. Week dummies were included in our model to control for week-specific effects. Table 1 presents the descriptive statistics for the key variables and the correlations among them.

Table 1. Correlation Coefficient among Key Variables

Mean Std. 1 2 3 4 5 6 7 8 9

1. Abnormal Return 0.14 0.52 1.00

2. Intensity 37.48 50.13 0.03 1.00

3. Richness 0.57 0.29 0.04 -0.01 1.00

4. Responsiveness 0.15 0.16 0.05 0.01 0.02 1.00

5. Consumer Engagement 0.36 3.02 0.03 0.06 0.03 0.07 1.00

6. Consumer Attention 2.90 9.54 0.02 0.06 0.03 0.02 0.03 1.00

7. Network Size 4767.4 15470 0.06 0.12 0.15 -0.04 -0.02 0.04 1.00

8. Posting Length 196.7 111.4 -0.01 -0.05 -0.01 -0.03 -0.03 0.02 0.01 1.00

9. Influence Duration 0.73 0.49 0.01 0.01 0.08 0.19 -0.01 0.03 0.02 0.02 1.00

10. Internet Search 40.32 22.80 -0.01 -0.01 -0.13 0.08 -0.02 0.12 -0.02 -0.03 0.01

Model Specification and Procedure

We estimate three equations to test our research questions. Equation (1) assesses the relationship between a firm’s social media efforts and consumer engagement. Specifically, the equation includes three

10 Note that reach rate is calculated as a percentage in a relative sense. By using alternative measures for Web traffic (i.e., page views per user and average visit duration), we obtained similar results. Details are available upon request. 11 Because Facebook does not provide the weekly trend of fan size (i.e., time information on the number of fans) and we collected cross-sectional Facebook data, and used the number of unique customers who post or comment on firm’s Facebook wall until week t as a proxy for the network size of each firm in week t. 12 The key words for each firm are based on the top 10 query key words from search engines provided by Naver.com, the most popular web portal in South Korea. For example, the top queries driving traffic to Samsung.com include “Samsung,” “Samsung electronics,” “Galaxy,” “Galaxy S3,” and “Samsung TV.”



dimensions of a firm’s social media efforts: intensity, richness, and responsiveness. We then include network size, post length, and influence duration to control for exogenous factors that can affect consumer engagement.13 Equation (2) assesses the relationship between a firm’s social media efforts and consumer attention. Note that this equation includes consumer engagement to test its effect on consumer attention, in line with our research framework (see Figure 1). In addition to the control variables in Equation (1), we include Internet searches to control for web traffic from generic searches.14 Equation (3) assesses the interrelationships among consumer engagement, consumer attention, and firm performance. To predict firm value, this equation incorporates consumer engagement, consumer attention, and the control variables included in Equation (2). Further, we include a comprehensive set of firm-specific exogenous variables, as noted earlier. In sum, our system of equations is specified as follows: i denotes each firm in our sample, t denotes each week, and each equation includes time-period dummies and industry dummies.

Consumer Engagementi,t

= α1 Intensityi,t + α2 Richnessi,t + α3 Responsivenessi,t

+ α4 ln(Network Size)i,t + α5 ln(Post Length)i,t + α6 Influence Durationi,t

+ Σ αs Time Dummiess + Σ αk Industry Dummiesk + εi,t ; (1)

Consumer Attentioni,t

= β1 Intensityi,t + β2 Richnessi,t + β3 Responsivenessi,t

+ β4 Consumer Engagementi,t + β5 ln(Network Size)i,t + β6 ln(Post Length)i,t

+ β7 Influence Durationi,t + β8 Internet Searchi,t

+ Σ βs Time Dummiess + Σ βk Industry Dummiesk + δi,t ; (2)

Firm Performancei,t

= γ1 Intensityi,t + γ2 Richnessi,t + γ3 Responsivenessi,t

+ γ4 Consumer Engagementi,t + γ5 Consumer Attentioni,t + γ6 ln(Network Size)i,t

+ γ7 ln(Post Length)i,t + γ8 Influence Durationi,t + γ9 Internet Searchi,t

+ Σ γj Firm Controlsj,t + Σ γs Time Dummiess + Σ γk Industry Dummiesk + φi,t. (3)

To account for sources of possible simultaneity among certain measures (e.g., consumer engagement and attention), we employed three-stage least squares (3SLS) estimation regressions for our system of equations. The 3SLS procedure is used to derive the parameters of the full system because endogenous variables in some equations of the model are used as explanatory variables in other equations. It combines two-stage least squares (2SLS) and seemingly unrelated regression (SUR) methods to take into account the cross-equation correlation of errors among different dependent regressors. Without loss of generality, we normalize Intensity, Information Richness, Responsiveness, and Engagement to 1. In addition to the 3SLS regression, SUR estimations were also performed for robustness checks. The results were qualitatively similar and are not reported here due to space limitations.

Results

Effects of Firm’s Social Media Efforts

We begin our analysis by estimating equation (1), (2), and (3) simultaneously. Table 2 presents 3SLS results on our models. The estimates reported in Model 1 provide an answer to RQ2a and suggest that the

13 We take the natural logarithm of the network size and post length to control for their skewedness and the different absolute frequency. 14 Unlike Luo et al. (2013) and Luo and Zhang (2013), who consider both web traffic and Internet searches as independent variables to predict firm value, we treat Internet searches as an exogenous variable of web traffic; we include this variable to control for the web traffic caused by generic Internet searches and to tease out the impact of incremental web traffic from a firm’s social media efforts.



impact of Intensity, Richness, and Responsiveness on consumer engagement is positive and significant (α1 = 0.510, p < 0.05; α2 = 0.318, p < 0.05; α3 = 0.537, p < 0.01, respectively). These results show that when firms invest more efforts in their social media with respect to both quantity and quality, customers will be more likely to engage in social media sites through postings, comments, or “Likes.”

Estimates in Model 2, examining RQ3a, suggest that the impact of Intensity and Richness on consumer attention is also significant and positive (β1 = 1.029, p < 0.05; β 2 = 2.184, p < 0.01, respectively), but the impact of Responsiveness on consumer attention is not statistically significant due to a relatively high standard error (β 3 = 0.170, p > 0.1). These results suggest that as a firm increases the amount and richness of information it provides on its social media site, customers are more likely to pay attention to the firm and access its website.

Interestingly, the coefficient of a firm’s social media efforts reported in Model 3 shows that Richness and Responsiveness are important aspects of a firm’s social media efforts, which positively contribute to firm performance (γ2 = 0.305, p < 0.05; γ 3 = 0.713, p < 0.05, respectively), whereas the Intensity of a firm’s social media efforts is not helpful in increasing the firm’s market performance (γ1 = 0.182, p > 0.1). Addressing RQ1a, RQ1b, and RQ1c, these findings suggest that firms’ social media strategy of focusing on information richness and responsiveness pays off more than the strategy of focusing on volume, thereby highlighting the relative inferiority of intensity strategy. This also highlights the importance of considering information richness and responsiveness as effective strategies for managing a firm’s social media.

Table 2. 3SLS Estimates Results

(1) (2) (3) Dependent variable

Consumer Engagement (Facebook)

Consumer Attention (Web Traffic)

Firm Performance (Abnormal Return)

Firm’s Efforts Intensity 0.5102** (0.2624) 1.0287** (0.4507) 0.1824 (0.1233) Richness 0.3176** (0.2055) 2.1836*** (0.2971) 0.3054** (0.1506) Responsiveness 0.5371*** (0.3567) 0.1697 (0.5761) 0.7131** (0.3503)

Consumer Behavior Engagement 1.8472*** (0.3020) 0.2648*** (0.0752) Attention (Web Traffic) 0.2005** (0.0894)

Control variables ln(Network size) 0.2372** (0.0596) 1.0436*** (0.2995) 0.2042*** (0.0762) ln(Post length) -0.0029 (0.0045) 0.0269 (0.0214) 0.0157 (0.0107) Influence duration 0.0143*** (0.0044) -0.0232 (0.0290) -0.0073 (0.0144) Internet search 0.0249*** (0.0040) 0.0075*** (0.0003) Total asset -0.3187** (0.1481) R&D intensity 0.0041 (0.0042) Advertising intensity -0.1134 (0.5749) Operating margin 0.0029 (0.0007) Leverage 0.7312*** (0.2407) Liquidity -0.0137** (0.0057) HHI 1.7920*** (0.5120) Dummies Week/Industry Week/Industry Week/Industry

Number of firms 63 63 63

Observations 4,208 4,208 4,208

Model fit χ2 = 461.62*** Adj. R2 = 0.545

χ2 = 82.88*** Adj. R2 = 0.103

χ2 = 101.64*** Adj. R2 = 0.465

Notes. Standard errors in parentheses. *, **, and *** denote significance at the 0.10, 0.05, and 0.01 levels, respectively. Week and industry dummies are included, but not reported for expositional brevity.



Consumer Engagement, Attention, and Firm Performance

In Models 2 and 3, we examine the relationship among consumer engagement, attention, and firm performance. Our analysis reveals that the impact of consumer engagement on consumer attention is significantly positive (β 4 = 1.847, p < 0.01). Answering RQ4 in the affirmative, these results suggest that consumer engagement on a firm’s social media site increases the probability of consumers visiting the firm’s website. Our results also show that both consumer engagement and attention are important factors that contribute to firm performance (γ4 = 0.265, p < 0.01; γ 5 = 0.201, p < 0.05, respectively). Thus, we provide empirical answers to RQ2b and RQ3b. In sum, our analysis results suggest that a firm’s social media efforts positively affect firm performance, both directly and indirectly through consumer engagement and attention. Interestingly, we find that only the richness and responsiveness dimensions (i.e., qualitative aspects) have a positive effect on firm performance. Further, fans’ active engagement on a firm’s social media site can increase consumers’ attention to the firm through social contagion, thus resulting in more visits to the firm’s website.

Robustness Checks

To check the robustness of our results, we first re-estimate our system of equations separately using Arellano-Bover/Blundell-Bond system GMM models. Our independent variables, except for the week and industry dummies, are instrumented with lagged values of the variables, in both levels and in their own first differences. We use the two-step system GMM estimator, which is asymptotically efficient and robust to any panel-specific autocorrelation and heteroskedasticity. As shown in Table 3, the overall results based on GMM estimations are qualitatively similar to our 3SLS estimation results.

Table 3. System GMM Estimates Results





Firm’s Efforts Intensity 0.7235** (0.3548) 0.8198*** (0.5511) 0.1597 (0.1918) Richness 0.3514** (0.3433) 0.9710*** (0.4215) 0.4958** (0.2201) Responsiveness 0.1469* (0.2113) 0.1492 (0.5042) 0.4320** (0.3431)

Consumer Behavior Engagement 1.0621*** (0.3761) 0.1987*** (0.0790) Attention (Web Traffic) 0.1823** (0.1178)

Control variables ln(Network size) 0.1499** (0.0391) 0.1792*** (0.0532) 0.1008* (0.0643) ln(Post length) 0.0621 (0.0412) 0.0496 (0.0276) 0.0192 (0.0039) Influence duration 0.0686 (0.0904) -0.0292 (0.0845) 0.0062 (0.0083) Internet search 0.0163** (0.0053) 0.0100** (0.0016) Total asset -0.0519** (0.0280) R&D intensity 0.1232* (0.0551) Advertising intensity 0.1220** (0.0526) Operating margin 0.0044 (0.0018) Leverage 0.7052*** (0.2590) Liquidity -0.0134** (0.0061) HHI 0.8163 (0.8916) Dummies Week/Industry Week/Industry Week/Industry

Observations 4,208 4,208 4,208

F-value (df) 9.76 (134,62) (p < 0.01)

10.15 (135,62) (p < 0.01)

25.37 (143,62) (p < 0.01)

Arellano-Bond test for AR(2) z = -0.37 (p = 0.712) z = -1.04 (p = 0.345) z = -1.2 (p = 0.285) Hasen test, chi2 (df) 6.11 (3841) 13.15 (3842) 15.28 (3850) Notes. Standard errors in parentheses. *, **, and *** denote significance at the 0.10, 0.05, and 0.01 levels, respectively. Week and industry dummies are included, but not reported for expositional brevity.



Second, it is likely that a firm’s social media efforts are not random in our setting. For example, when firms choose high levels of their social media efforts, expecting a positive impact of their efforts on firm performance, social media efforts can be endogenous, and the estimates may be biased (Heckman 1976). To control for this potential selection bias, we estimated the relationship between a firm’s social media efforts and firm performance using the Heckman two-step selection model. Specifically, we estimated the combination of the firm performance model in Equation (3) and the selection models. To model firm’s social media efforts selection decision (Efforti), we include several exogenous variables as covariates in the first-stage probit model shown in Equation (4) to (6):

Efforti* = δ1 ln(Network Size)i + δ2 Internet Searchi + δ3 Total Asseti + δ4 R&D Intensityi

+ δ5 Advertising Intensityi + δ6 Operating Margini + δ7 Leveragei + δ8 Liquidityi + πi (4)

Efforti = 1 if Efforti* > 0, and Efforti = 0 otherwise (5)

Prob(Efforti = 1 | zi) = Ф(ziδ), Prob(Efforti = 0 | zi) = 1 - Ф(ziδ), (6)

where zi is a vector of Heckman first-stage model covariates.

Because our social media effort variables are continuous, we dichotomized them and estimated a probit model in the first stage. Then we included an inverse Mill’s ratio in the second-stage models (with consumer engagement, consumer attention, and firm performance, respectively, as the dependent variables). As shown in Table 4, in all models, the relationship between a firm’s social media efforts and firm performance remained highly significant, thus corroborating our main results. As a result, we conclude that selection bias is not a serious concern in our estimations.

Table 4. Heckman Selection Model Results





Firm’s Efforts Intensity 0.1561 (0.1698) 0.1306 (0.1840) 0.1383 (0.1529) Richness 0.2943** (0.1453) 0.1473** (0.1189) 0.1472** (0.1189) Responsiveness 0.4926*** (0.2895) 0.5002*** (0.2583) 0.5015*** (0.2391)

Consumer Behavior Engagement 0.4776*** (0.0911) 0.4654*** (0.0987) 0.4659*** (0.0992) Attention 0.1331** (0.0574) 0.1325*** (0.0512) 0.1324** (0.0583)

Lambda 0.0194 (0.4696) 0.0231 (0.1547) 0.0215 (0.8311) Observations 4,208 4,208 4,208 Notes. This table reports the results of estimating a Heckman two-step selection model in which the first stage regression predicts firm’s social media efforts (intensity, richness, and responsiveness, respectively) as a function of several observable characteristics of firms. In Model 1, we estimated the first stage of the model by predicting the level of intensity using observable characteristics of firms including network size, Internet search, total asset, R&D intensity, advertising intensity, operating margin, leverage, and liquidity. Similarly, in Models 2 and 3, we predicted the level of richness and the level of responsiveness, respectively, in the first stage of the model. We do not report results from the first stage (i.e., “selection”) of the Heckman models for expositional brevity. Robust standard errors are in parentheses. *, **, and *** denote significance at the 0.10, 0.05, and 0.01 levels, respectively.

Third, we used Tobin’s q as an alternative measure of firm performance. Our aim here is to further corroborate the findings from our main estimation approach. Tobin’s q analysis is the most widely used approach to estimate the value of intangible assets and customer satisfaction (Anderson et al. 2004; Hall et al. 2005). Because Tobin’s q offers the advantage of capturing both short-term performance and long-term prospects based on the market value, it has been widely used in prior literature (e.g., Bardhan et al. 2013). We calculated Tobin’s q for our sample firms on a quarterly basis during our sample period; January 2010 - December 2012. We choose a quarterly interval because it is the smallest interval available. We calculate the quarterly averages of all our independent variables except for firm-specific financial control variables. We have therefore 8 quarters and 428 observations in total. Table 5 presents the 3SLS results based on Tobin’s q. The results corroborate our main results based on weekly abnormal returns, and indicate that firms’ social media efforts and subsequent consumer engagement and attention have a



significant influence on both the short-term and long-term performance of firms. It is worthwhile to note that although the impact of the intensity dimension of social media efforts on abnormal returns is not significant, the impact of Tobin’s q is significant; this suggests that although the intensity (quantity) strategy regarding social media is not effective in the short run, it can pay off in the long run.

Table 5. Tobin’s q Analysis

(1) (2) (3) Dependent variable Consumer Engagement

(Facebook) Consumer Attention

(Web Traffic) Firm Performance

(Tobin’s q) Firm’s Efforts

Intensity 0.2201*** (0.0473) 1.1410*** (0.1944) 0.6234** (0.4679) Richness 0.0433 (0.0715) 0.8589*** (0.2151) 0.7644** (0.3250) Responsiveness 0.3797*** (0.1173) 0.6514 (0.4243) 1.6455*** (0.8208)

Consumer Behavior Engagement (Facebook) 4.4251*** (0.6055) 2.1839*** (0.7244) Attention (Traffic) 0.0661* (0.0342)

Control variables ln(Network size) 0.0601 (0.0173) 0.3174*** (0.0637) 0.1254 (0.1359) ln(Post length) -0.0012 (0.0186) 0.0642 (0.1218) 0.2611* (0.3769) Influence duration 0.0102* (0.0084) -0.0127 (0.0717) 0.1321 (0.1002) Internet search 0.0456** (0.0186) 0.1023* (0.0042) Total asset 0.0592 (0.1108) R&D intensity -0.0219*** (0.0078) Advertising intensity -0.1560 (0.6145) Operating margin 0.0212*** (0.0032) Leverage 0.7590 (1.2522) Liquidity -0.0020 (0.0049) HHI -0.7172 (1.8154) Dummies Quarter/Industry Quarter/Industry Quarter/Industry

Observations 428 428 428

Model fit χ2 = 375.58*** Adj. R2 = 0.702

χ2 = 70.72*** Adj. R2 = 0.237

χ2 = 123.70*** Adj. R2 = 0.467

As additional robustness checks, we compared several models with various combinations of lagged values (t-1) of firms’ social media efforts, consumer engagement, and consumer attention in order to address potential simultaneity issues. We first considered the lagged effect of the three dimensions of social media efforts (t-1) on consumer engagement, consumer attention, and firm performance. Then, we considered the lagged effect of social media efforts and consumer engagement (t-1), but not that of consumer attention (t). Finally, we considered the lagged effect of social media efforts, consumer engagement, and consumer attention. The results, based on these different combinations of lagged effects, are qualitatively similar to our main results (not presented here). Thus, we conclude that simultaneity is not a serious concern in our estimations. Further, we estimated our system of equations using a single measure of consumer engagement. Specifically, we used the number of consumers’ comments and the number of consumers’ “Likes,” respectively, in order to address the concern that making comments and clicking on the “Like” button may represent different levels of consumer engagement because the former is a more laborious task than the latter. However, the overall results are qualitatively similar to our main results, based on a composite measure of consumer engagement. This finding indicates that our results are not sensitive to the different measures of consumer engagement.

Discussion and Conclusion

Theoretical Implications

Our study makes several important contributions to the literature. First, we conceptualize and quantify the multiple dimensions of a firm’s social media efforts and the important role they play in predicting firm



value - Intensity, Richness, and Responsiveness. Although considerable research has been conducted on social media, prior research has not considered these multiple dimensions. Therefore, although social media have become indispensable for firms, the financial returns from a firm’s own efforts in social media have been unclear. By conceptualizing Intensity, Richness, and Responsiveness as key strategies in a firm’s social media management (quantity vs. quality strategy), and by quantifying these constructs from firms’ Facebook data, this study validates the conjecture that a firm’s social media efforts contribute to its market performance. Our findings provide a rationale for firms to invest heavily in their social media, while countering those who question the value of social media (Mueller 2011). Importantly, our research complements previous studies on the business value of social media that have focused on blogs and consumer ratings (Luo and Zhang 2013; Luo et al. 2013) by examining firm-initiated social media, which tend to have higher sales elasticity than that of traditional media (Stephen and Galak 2012).

Second, we develop a comprehensive model that considers consumer engagement and attention as intermediate mechanisms through which a firm’s social media efforts influence firm performance. With our research model, our study is the first to uncover the intricate relationships among a firm’s social media efforts, consumer engagement and attention, and firm performance. Specifically, we find that a firm’s social media efforts can improve consumer engagement and attention, which in turn, result in higher market performance. Interestingly, the Intensity of a firm’s social media efforts only has an indirect effect on firm performance through consumer engagement and attention. Thus, firms should put more emphasis on Richness and Responsiveness in order to increase the returns to their social media efforts. Recent reports mention that indiscriminate social media messages can cause digital fatigue in consumers (Fouts 2013; Rosenbloom 2011), suggesting that the Intensity of a firm’s social media efforts may actually have a negative effect. Furthermore, our finding suggests that consumer engagement on a firm’s social media site can increase web traffic to the firm’s website, and this effect is more salient for transaction-oriented firms that make profits from their websites. Thus, managers should consider this complementarity between social media sites and websites for higher returns to investments in social media.

Finally, prior studies have investigated the relationship between web traffic/Internet searches and firm performance (Da et al. 2011; Demers and Lev 2001; Dewan et al. 2002; Trueman et al. 2000), the relationship between web blogs/consumer ratings and firm performance (Luo and Zhang 2013; Luo et al. 2013; Yu et al. 2013), and the relationship between Facebook postings and sales/visit frequency (Goh et al. 2013; Rishika et al. 2013). We extend this stream of research by providing evidence that a firm’s own efforts on social media in terms of richness and responsiveness, which can represent the quality dimensions of social media strategy, can be an important driver of both short- and long-term firm performance; however, the intensity of a firm’s social media efforts, which corresponds to the quantity dimension of social media strategy, can be a driver of long-term performance only. Together, our study provides a rich understanding of the performance impacts of firms’ own social media efforts.

Managerial Implications

The findings of this study provide practical implications for managing social media and achieving higher firm performance. First, our findings can help managers gauge the impact of their firms’ social media efforts in terms of their contribution to firm performance. In particular, a firm’s social media efforts, as captured by intensity, richness, and responsiveness, can improve the firm’s equity value, as well as its market value. Our results provide an explanation as to why many firms are investing heavily in social media management (Wessel 2011) and are interacting with their customers on firm-initiated social media sites. Firms operating in a competitive industry face challenges in the valuation of their intangible assets, and a firm’s social media efforts, viewed as endeavors to deliver informative messages about the firm and to communicate with its customers more efficiently and effectively, may function as a signal to investors that a firm has valuable assets (Marks 2014). Thus, our findings can help firms by providing evidence about the positive effect of investments in social media on firm performance.

Second, our study findings can help managers formulate their social media strategies that will affect the returns to their investments in social media. In general, our results suggest that among the various dimensions of a firm’s social media efforts, richness and responsiveness (related to quality-oriented social media strategies) are directly germane to firm value (especially in the short run). Thus, social media managers need to focus on delivering useful, richer information and actively responding to consumer



comments; merely proving a large amount of uninformative messages is not an effective strategy. Furthermore, our results from an additional analysis indicate that for non-transaction-oriented firms, there is no significant association between consumer engagement on social media and website traffic. Thus, managers should tailor their social media strategies, depending on the characteristics of their business, to increase their website traffic and eventually their firm performance.

Finally, our study provides a useful methodology based on big data analytics that firms can use to collect and analyze data from their own and their competitors’ social media sites, especially with respect to the three key dimensions of social media efforts. By collecting and analyzing data based on the big data analytics used in our study, firms can benchmark against their competitors with respect to their levels of intensity, richness, and responsiveness. This, in turn, can enable them to formulate and revise their social media strategies. By doing so, they can achieve higher returns to their social media investments, compared to when they blindly invest in social media.

Limitations and Future Research

Our study has several limitations and offers directions for future research. First, although we studied the impact of a firm’s social media efforts, we did not examine how the valence of messages affects consumer engagement and attention, and firm performance. This was because most of the message postings in our data are either positive or neutral, which limits us from studying the impact of message valence. However, the results would have generally been weaker if there had been a significant proportion of firms that had inevitably faced negative publicity, such as in the cases of product recalls and industrial disasters. In addition, we did not study the differential effects of different types of information that firms post on their social media sites (e.g., promotion related vs. CSR related). It would be interesting to examine whether and how various contents of postings and comments by firms and customers influence consumer behavior and firm performance differently. For example, Lee et al. (2013) examined the impact of social media content (persuasive vs. informative) on customer engagement by analyzing the advertising messages firms post on their Facebook pages. Nevertheless, the use of frequencies of postings and comments to operationalize the dimensions of a firm’s social media efforts appear to have great promise for empirical research, and offers a relatively easy way to capture these concepts objectively. Second, we focused on large firms that are listed in public stock markets and that have already adopted social media. Our results may not generalize to small firms, which have not invested in social media as actively. Although a recent report suggests that small companies need social media too (Hardy 2012), future research should replicate our analysis for small firms. Further, our research design does not allow us to consider the difference in firm performance before and after adopting social media. To alleviate some of these issues, one fruitful direction for future search is the use of field experiments and difference-in-difference techniques (Rishika et al. 2013). Finally, similar to other empirical studies using secondary data, we cannot claim causality, although we have done our best to address reverse causality and endogeneity. Further studies are needed to demonstrate causality between firms’ social media efforts on consumer behavior and firm performance.

Conclusions

Our study examined the relationship between a firm’s social media efforts and firm performance, while also considering the mediating role of consumer engagement and attention. Our findings suggest that a firm’s social media efforts do matter for firm performance across industries, and the value of a firm’s social media efforts critically depends on the strategies (i.e., dimensions) of social media management. Engagement is at the heart of social media, as it provides a firm an opportunity to interact with its customers from all over the world. To extract value from social media, a firm needs to engage with its current and potential customers, as well as other stakeholders. Our study highlights the importance of focusing on the richness and responsiveness of the communication, in addition to the volume of communication in order to engage a firm’s audience, influence them to act and spread the firm’s messages, and attract the attention of current and potential audiences to the firm’s website. We also show that a firm’s efforts to facilitate such engagement and generate attention allow the firm to take advantage of social media and improve its financial performance. Thus, we underscore the business value of being “connected” in a digital economy in which people are increasingly networked. We hope that our research inspires future studies exploring the business value of social media.



References

Adler, P.S., and Kwon, S.-W. 2002. "Social Capital: Prospects for a New Concept," The Academy of Management Review (27:1), pp. 17-40.

Algesheimer, R., Dholakia, U.M., and Herrmann, A. 2005. "The Social Influence of Brand Community: Evidence from European Car Clubs," Journal of Marketing (69:3), pp. 19-34.

Anderson, E.W., Fornell, C., and Mazvancheryl, S.K. 2004. "Customer Satisfaction and Shareholder Value," Journal of Marketing (68:4), pp. 172-185.

Aral, S., and Walker, D. 2011. "Creating Social Contagion through Viral Product Design: A Randomized Trial of Peer Influence in Networks," Management Science (57:9), pp. 1623-1639.

Bardhan, I., Krishnan, V., and Lin, S. 2013. "Business Value of Information Technology: Testing the Interaction Effect of It and R&D on Tobin's Q," Information Systems Research (24:4), pp. 1147-1161.

Bond, R.M., Fariss, C.J., Jones, J.J., Kramer, A.I., Marlow, C., Settle, J.E., and Fowler, J.H. 2012. "A 61-Million-Person Experiment in Social Influence and Political Mobilization," Nature (489), pp. 295-298.

Chen, Y., Liu, Y., and Zhang, J. 2011. "When Do Third-Party Product Reviews Affect Firm Value and What Can Firms Do? The Case of Media Critics and Professional Movie Reviews," Journal of Marketing (76:2), pp. 116-134.

Coleman, J.S. 1988. "Social Capital in the Creation of Human Capital," American Journal of Sociology (94), pp. S95-S120.

Culnan, M.J., McHugh, P.J., and Zubillaga, J.I. 2010. "How Large U.S. Companies Can Use Twitter and Other Social Media to Gain Business Value," MIS Quarterly Executive (9:4), pp. 243-259.

Da, Z., Engelberg, J., and Gao, P. 2011. "In Search of Attention," Journal of Finance (66:5), pp. 1461-1499. Daft, R.L., and Lengel, R.H. 1986. "Organizational Information Requirements, Media Richness and

Structural Design," Management Science (32:5), pp. 554-571. Dellarocas, C. 2006. "Strategic Manipulation of Internet Opinion Forums: Implications for Consumers

and Firms," Management Science (52:10), pp. 1577-1593. Demers, E., and Lev, B. 2001. "A Rude Awakening: Internet Shakeout in 2000," Review of Accounting

Studies (6:2-3), pp. 331-359. Dewan, R.M., Freimer, M.L., and Zhang, J. 2002. "Management and Valuation of Advertisement-

Supported Web Sites," Journal of Management Information Systems (19:3), pp. 87-98. Dyer, P. 2013. "10 Examples of Social Media Roi," in: Social Media Today. Emerson, M.F. 2012a. "Blogging to Build Your Business," in: New York Times. New York, NY: New York

Times. Emerson, M.F. 2012b. "Social Media Marketing from a to Z," in: The New York Times. New York. Fama, E.F., and French, K.R. 1993. "Common Risk Factors in the Returns on Stocks and Bonds," Journal

of Financial Economics (33:1), pp. 3-56. Fama, E.F., and French, K.R. 1996. "Multifactor Explanations of Asset Pricing Anomalies," Journal of

Finance (51:1), pp. 55-84. Fornell, C., Mithas, S., III, F.V.M., and Krishnan, M.S. 2006. "Customer Satisfaction and Stock Prices:

High Returns, Low Risk," Journal of Marketing (70:1), pp. 3-14. Fouts, R. 2013. "What to Do About Social Media Fatigue," in: Gartner. Gallaugher, J., and Ransbotham, S. 2010. "Social Media and Customer Dialog Management at Starbucks,"

MIS Quarterly Executive (9:4), pp. 197-212. Godes, D., and Mayzlin, D. 2009. "Firm-Created Word-of-Mouth Communication: Evidence from a Filed

Test," Marketing Science (28:4), pp. 721-739. Goh, K.-Y., Heng, C.-S., and Lin, Z. 2013. "Social Media Brand Community and Consumer Behavior:

Quantifying the Relative Impact of User- and Marketer-Generated Content," Information Systems Research (24:1), pp. 88-107.

Gruca, T.S., and Rego, L.L. 2005. "Customer Satisfaction, Cash Flow, and Shareholder Value," Journal of Marketing (69:3), pp. 115-130.

Hall, B.H., Jaffe, A., and Trajtenberg, M. 2005. "Market Value and Patent Citations," RAND Journal of Economics (36:1), pp. 16-38.

Han, K., and Mithas, S. 2013. "Information Technology Outsourcing and Non-It Operating Costs: An Empirical Investigation," MIS Quarterly (37:1), pp. 315-331.



Hardy, Q. 2012. "Salesforce Bets Small Companies Need Social Media Too," in: New York Times. New York, NY.

Healy, P.M., and Palepu, K.G. 2001. "Information Asymmetry, Corporate Disclosure, and the Capital Markets: A Review of the Empirical Disclosure Literature," Journal of Accounting and Economics (31:1), pp. 405-440.

Heckman, J.J. 1976. "The Common Structure of Statistical Models of Truncation, Sample Selection and Limited Dependent Variables and a Simple Estimator for Such Models," Annals of Economic and Social Measurement (5:4), pp. 475-492.

J.D.Power. 2013. "Poor Social Media Practices Can Negatively Impact a Businesses' Bottom Line and Brand Image."

Larkin, J.H., and Simon, H.A. 1987. "Why a Diagram Is (Sometimes) Worth Ten Thousand Words," Cognitive Science (11:1), pp. 65-100.

Lee, D., Hosanagar, K., and Nair, H. 2013. "The Effect of Advertising Content on Consumer Engagement: Evidence from Facebook ".

Ling, B. 2013. "18 Million Obvious Secrets About Facebook Business Pages You Need to Know Today," in: InsideFacebook.

Luo, X. 2009. "Quantifying the Long-Term Impact of Negative Word of Mouth on Cash Flows and Stock Prices," Marketing Science (28:1), pp. 148-165.

Luo, X., and Zhang, J. 2013. "How Do Consumer Buzz and Traffic in Social Media Marketing Predict the Value of the Firm?," Journal of Management Information Systems (30:2), pp. 213-238.

Luo, X., Zhang, J., and Duan, W. 2013. "Social Media and Firm Equity Value," Information Systems Research (24:1), pp. 146-163.

Marks, G. 2014. "Today in Small Business: A Social Media Lesson," in: New York Times. New York, NY. Miller, A.R., and Tucker, C. 2013. "Active Social Media Management: The Case of Health Care,"

Information Systems Research (24:1), pp. 52-70. Miranda, S.M., and Saunders, C.S. 2003. "The Social Construction of Meaning: An Alternative Perspective

on Information Sharing," Information Systems Research (14:1), pp. 87-106. Mueller, M. 2011. "Should Every Business Invest in Social Media?," in: New York Times. New York. Park, O.-C., and Hopkins, R. 1992. "Instructional Conditions for Using Dynamic Visual Displays: A

Review," Instructional Science (21:6), pp. 427-449. Pauwels, K. 2004. "How Dynamic Consumer Response, Competitor Response, Company Support, and

Company Inertia Shape Long-Term Marketing Effectiveness," Marketing Science (23:4), pp. 596-610.

Rajgopal, S., Kotha, S., and Venkatachalam, M. 2000. "The Relevance of Web Traffic for Internet Stock Prices of Internet Firms."

Rishika, R., Kumar, A., Janakiraman, R., and Bezawada, R. 2013. "The Effect of Customers' Social Media Participation on Customer Visit Frequency and Profitability: An Empirical Investigation," Information Systems Research (24:1), pp. 108-127.

Rosenbloom, S. 2011. "For the Plugged-in, Too Many Choices," in: New York Times. New York. Sawhney, M., Verona, G., and Prandelli, E. 2005. "Collaborating to Create: The Internet as a Platform for

Customer Engagement in Product Innovation," Journal of Interactive Marketing (19:4), pp. 4-17. Statman, M., Thorley, S., and Vorkink, K. 2006. "Investor Overconfidence and Trading Volume," The

Review of Financial Studies (19:4), pp. 1531-1565. Stephen, A.T., and Galak, J. 2012. "The Effects of Traditional and Social Earned Media on Sales: A Study

of a Microlending Marketplace," Journal of Marketing Research (49:5), pp. 624-639. Tellis, G.J., and Johnson, J. 2007. "The Value of Quality," Marketing Science (26:6), pp. 758-773. Thompson, S.A., and Sinha, R.K. 2008. "Brand Communities and New Product Adoption: The Influence

and Limits of Oppositional Loyalty," Journal of Marketing (72:6), pp. 65-80. Tirunillai, S., and Tellis, G.J. 2012. "Does Chatter Really Matter? Dynamics of User-Generated Content

and Stock Performance," Marketing Science (31:2), pp. 198-215. Trueman, B., Wong, M.H.F., and Zhang, X.-J. 2000. "The Eyeballs Have It: Searching for the Value in

Internet Stocks," Journal of Accounting Research (38:Supplement), pp. 137-170. Ugander, J., Backstrom, L., Marlow, C., and Kleinberg, J. 2012. "Structural Diversity in Social Contagion,"

Proceedings of the National Academy of Sciences of the United States of America (109:16), pp. 5962-5966.

Vahl, A. 2014. "How to Drive More Facebook Traffic to Your Website in 5 Easy Steps," in: Social Media Examiner.



Wessel, R. 2011. "Activist Investors Turn to Social Media to Enlist Support," in: New York Times. New York, NY.

Yang, S.-B., Lim, J.-H., Oh, W., Animesh, A., and Pinsonneault, A. 2012. "Using Real Options to Investigate the Market Value of Virtual World Businesses," Information Systems Research (23:3), pp. 1011-1029.

York, J. 2012. "Social Business Enlightenment," in: Forbes. Yu, Y., Duan, W., and Cao, Q. 2013. "The Impact of Social and Conventional Media on Firm Equity Value:

A Sentiment Analysis Approach," Decision Support Systems (55:4), pp. 919-926.

Documents

Firmâ s Social Media Efforts, Consumer Behavior, and Firm ... · the impact of social media on customers’ purchase behavior, they are based on data from a single firm and do not