International corporate social media marketing and firm value: Are firms
that are popular on Facebook and Twitter valued higher than their less
popular peers? Do ‘likes’ create value?
This study examines the relationship between social media metrics from Facebook and Twitter and firm value. Results indicate a significant and positive relationship between firm’s social media popularity and market capitalization and market-to-book ratio, while controlling for firm size, leverage, growth, age and country and industry specific factors. The effect of social media popularity on firm value is not significantly influenced by social media activity and the standard of living. The effect of social media metrics on firm value are tested with cross-sectional multivariate OLS
regression models. A unique dataset has been collected consisting of 1968 large, listed firms from 11 different countries (US, UK, Australia, Norway, Sweden, Denmark, Germany, Netherlands, France, Spain, Portugal) and 18 different industries.
Student Number: s1946927 Student Name: Matthias Mainka
Study Programme: MSc Double Degree IFM
Faculty: Faculty of Economics and Business, University of Groningen Supervisor: Dr. M. A. Lamers
Field Key Words: firm value, social media marketing, Twitter, Facebook, market capitalization, market-to-book ratio
1. Introduction
In February 2004, three Harvard students launched a website called ‘thefacebook.com’. In June 2016, the social media service company ‘Facebook Inc.’ counted 1.71 billion monthly active users on its social network site ‘Facebook’ (Facebook, 2016). In other words, 23% of the world population is active on Facebook nowadays. To put this into perspective, around 3.6 billion people worldwide were using the internet in June 2016 (Internet world stats, 2016). Thus, almost half of all internet users are active on the social media site Facebook.
The tremendous increase in the usage and popularity of social media attracts many firms to use social media as well (Agnihotri et al., 2016; Benthaus, Risius, and Beck, 2016; Godey et al., 2015). Although many firms implement social media as a marketing tool, research on whether those social media marketing efforts pay off financially is scare (Kim and Ko, 2012; Karjaluoto et al., 2015; Luo and Zhang, 2013). Therefore, this study examines the following research question:
Are social media marketing efforts on Facebook and Twitter creating firm value? Are firms that perform very well on those social media platforms valued higher than their less popular peers? Is social media marketing a global homogenous phenomenon or does its implementation and effect on firm value differ between countries?
To date, research shows that social media marketing offers firms the opportunity to influence consumers’ purchase intention, increase brand equity (Kim and Ko, 2012), increase brand loyalty (Erdogmus and Cicek, 2012), stimulate word of mouth (Chevalier and Mayzlin, 2006), and create consumer traffic and buzz (Luo and Zhang, 2013), which all imply a positive effect on firm value. On the other hand, Mangold and Faulds (2009) and Hildebrand et al. (2013) point out that companies have less control about their marketing activities in the social media environment, which implies risks that marketing efforts backfire and destroy customer satisfaction, brand image and potentially firm value. This would imply a negative effect of social media marketing on firm value.
There is not only limited empirical evidence about the effect of social media marketing on firm value, but also are those findings mostly based on a single country (Yu et al., 2013; Rapp et al., 2013), single industry (Godey et al., 2016; Luo and Zhang, 2013), or a single social media platform analysis. However, research indicates that the usage of social media differs greatly between countries and industries depending on economical, technological, cultural, legal, and individual factors (Bolton et al., 2013; Van Belleghem et al., 2011; Universal Mccann, 2008).
UK, Norway, Sweden, Denmark, Germany, Netherlands, France, Spain, Portugal, Australia) and 18 different industries.
Results show that social media popularity, measured by the number of follower on Twitter and the number of likes on Facebook, does significantly positively influence firm value. Additional
hypotheses that the relationship between social media popularity and firm value is influenced by social media activity and countries’ GDP per capita are rejected. Twitter is widely used by firms from the US, Spain and Sweden. In contrast, Portuguese and Norwegian firms are rarely using Twitter. The following sections are structured as follows: The first part of the theory section explains why social media as a marketing tool is attractive for firms. Then, most relevant arguments and theories are reviewed to define hypothesis that explain the relationship between social media marketing and firm value. I discuss the research design in the third section of this paper. Results are presented in the fourth section, followed by a conclusion.
2. Theory
This section first explains why social media can be a very attractive marketing tool for firms.
Afterwards theories and empirical findings are summarized to predict the relationship between social media marketing and firm value.
2.1 Social media as marketing tool
Social media is defined as “…a group of internet-based applications that build on the ideological and
technological foundations of web 2.0, and that allow the creation and exchange of user generated content.” (Kaplan and Haenlein, 2010, p. 61). The most popular social media application worldwide is
Many consumers worship the presence and activity of brands in their social media networks and become ‘fans’ by liking and sharing their favorite brands’ content (Van Belleghem, Eenhuizen, and de Vries, 2011). According to Belleghem et al. (2011) more than 50% of social media users are connected to brands on their networks. Moreover, the average time people spend on social media per day
worldwide was steadily increasing in the last couple of years. According to an online survey provided by Statista (2016), worldwide daily social media usage increased from 96 minutes per day in 2012 to 118 minutes in 2016. Currently, Facebook is ranked third on the top 500 sites on the web worldwide, measured by the average amount of daily visitors and page views (Alexa, 2016). Only Google and Youtube are more frequently used websites at the moment. Twitter is currently on rank 14. Social media usage is thus not only growing in its scope, e.g. number of users, but also in the way how much time people spend using it.
That using social media as a marketing tool can be very beneficial for companies show recent studies of the luxury fashion industry. Researchers (Kim and Ko, 2012; Phan et al., 2011; Godey et al., 2016) state that companies from the luxury fashion industry have been early adopters of social media as a marketing tool and find several benefits of using it. Godey et al. (2016) examine social media marketing efforts of 5 luxury brands and its effect on consumer preferences, brand value, and brand loyalty measured by customer interviews conducted in China, France, India and Italy. Their survey results indicate that social media marketing positively influences brand loyalty and brand preferences and even the willingness of consumers to pay premium prices. Kim and Ko (2012) find similar results surveying 362 consumers of luxury brands in Seoul, Korea, who had experience with Louis Vuitton’s social media profiles. They find that social media marketing activities positively affect brand equity. Social media is not only a cheap and efficient tool for companies to relate to customers and society, but it also changes the way people use the internet, by shifting behavior from a passive reading and watching behavior to a more active creating, sharing, modifying, and discussing behavior (Kim and Ko, 2013; Aral, Dellarocas, and Godes, 2013). How this creating, sharing and discussing behavior can influence firm value is explained in the following section in greater detail.
2.2 Social media marketing and firm value
There are several arguments and empirical findings that suggest a positive relationship between social media marketing and firm value:
significant relationship between Twitter sentiment and stock market returns. Information created and shared in social media is constantly updated and spread in high speed so that it might even be ahead compared to other information sources. Investors might use social media information as a sentiment indicator for brand and product performance and even as a tool for predicting future sales and earnings. Traditional firm performance indicators such as financial figures like sales and earnings are only made available in specific intervals such as month, quarters or years. In contrast, social media indicators are available all the time so investors can timely assess product and brand performances even if sales data are not available. Investors can thus use social media as an additional information source to decrease information asymmetry in the stock market (Luo et al., 2013). They also argue that social media is a less biased indicator for customer satisfaction which can easily be absorbed by companies. In other words, firms can use social media as a mirror that reflects word of mouth and customer satisfaction and use those insights for targeted actions to further improve customer satisfaction and increase sales. In several studies, empirical evidence was found that customer satisfaction in turn has a positive relationship with firm performance and firm value (Fornell et al., 2006; Anderson et al., 2004; Sun & Kim, 2013).
The most important studies in the intersecting field of social media and finance examine the
relationship between social media and firm value by linking social media metrics to abnormal stock returns and risk. For example, the number of positive and negative blog posts and the volume of consumer ratings affect abnormal returns of major listed firms from the computer and software
industry (Luo et al., 2013). Moreover, stock risk, measured by the standard deviation of the residual of the extended Fama and French model, is influenced by the number of negative blog posts, the online consumer rating level, and the rating volume (Luo et al., 2013). Their findings are derived by using a vector autoregressive model with exogenous covariates to analyze time series data from 2007 to 2009. Luo and Zhang (2013) partly use the same sample data as Luo et al. (2013), but substitute the
sentiment analysis data with numerical meta data representing consumer buzz and consumer traffic. Consumer buzz refers to ‘the contagious talk about a brand, service, product or idea’ (Carl, 2006, p. 602). Several studies have shown that buzz can positively influence sales (Duan and Whinston, 2008; Senecel and Nantel, 2004; Resnick and Zeckhauser, 2002). Consumer traffic, mostly named web-traffic, is the amount of data that is exchanged between a web site and its visitors (Choi and Limb, 1999). Before the dot-com bubble collapsed in the early 2000s, many investors and researchers believed and proved that web traffic is an important and accurate indicator for internet firms’ value (Seiders and Riley, 1999; Demers and Lev, 2001; Keating et al., 2003). Metaphorically, a visitor of an online shops’ website is like a potential customer walking in a physical store. Thus, Internet
relating to each firm, collected from CNET.com. Their results suggest that both consumer buzz and consumer traffic significantly explain variation in abnormal stock returns and its idiosyncratic risk of major computer and software firms. Consumer buzz does so to a greater extent.
Similar findings present Yu et al. (2013), who show that social media content has even a stronger effect on abnormal returns and risk than conventional media. However, they also point out that the effect varies between different social media platforms. For example, the sentiment of forums
negatively impact abnormal returns while blog sentiment has a positive effect. Sentiment analysis is a text mining technique that automatically screens text and assigns values to identified positive and negative opinions. Their findings are derived by focusing on social media sentiment data of randomly selected sample of 824 listed companies from 6 different industries by using a daily sentiment analysis of 52,746 company related posts collected from Twitter, blogs, forums, and conventional media such as newspaper articles in a three month period in 2011. The sentiment data is combined with meta data of those posts such as the number of positive and negative posts. They measure firm value in a similar way as Luo et al. (2013) using an extended Fama and French model to calculate abnormal returns and risk.
On the other hand, one can argue that social media marketing activities can have a negative impact on firm value:
Mangold and Faulds (2009) point out that companies have less control about their marketing activities in the social media environment. The main reason is that social media offers not only communication between companies and customers, but also stimulates communication between customers. The opinions, comments, critics and compliments, e.g. word of mouth, that customers can easily spread worldwide to the whole social media community are out of companies’ control. Companies make themselves vulnerable by using social media as negative and harmful information and criticism can exaggerate and equally spread throughout the whole community creating so called ‘shit storms’ that can hurt brand image. A citation of a BBC business editor used in Kietzman et al. (2011) brings it nicely to the point: “These days, one witty tweet, one clever blog post, one devastating video –
forwarded to hundreds of friends at the click of a mouse – can snowball and kill a product or damage a company’s share price.” (Kietzman et al., p. 242, 2011). Empirical findings by Hildebrand et al. (2013) show that the social media environment used by the automotive industry to let people design their own cars led to dissatisfaction. The dissatisfaction was mainly caused by feedback from other social media community members.
All in all, the inconclusive relationship between social media metrics and firm value can be summarized in the following hypothesis:
2.3 Social media activity, popularity, and firm value
Researchers from the field of social media often give advise about how firms should use social media to be successful with it. For example, Kaplan and Haenlein (2010) suggest that firms should keep their profiles up to date and be very active in order to develop a good relationship to customers. Moreover, the content presented on social media profiles should be interesting so that fans don’t get bored. Kim and Ko (2012) identify entertainment, interaction, trendiness, customization and word of mouth to be important drivers that influence social media fans of luxury fashion brands. In an empirical study of De Vries and Leeflang (2012), similar social media content characteristics and their impact on
popularity are tested. Their results show that highly vivid content such as videos and pictures get more likes than less vivid content. Adapting this concept to companies’ social media profiles, one can argue that social media activity influences popularity. A higher amount of content gives potential fans a better basis to decide whether or not they like a specific profile. Moreover, being more active implies that more information are provided to investors. This might in turn decrease information asymmetries in the stock market which can have an effect on firm value. This leads to a second hypothesis, namely: 𝐻2: Social media activity significantly influences social media popularity and firm value.
2.4. Economic environment and the effect of social media popularity on firm value
One important factor that influences social media usage is the economic environment (Bolton et al., 2013). It can be argued that the standard of living and the availability and affordability of information technology hardware is determining the density of countries’ social media usage and awareness. Andres et al. (2010) examine the diffusion of the internet across countries and find that the adaption of internet did indeed differ greatly between low-income countries and high-income countries. Although low-income countries have a steeper adaption curve, their overall adaption process starts later
compared to high-income countries. One can thus expect that the usage of social media also differs between low-income countries and high-income countries which in turn influences the effect on social media popularity on firm value. Firm’s social media popularity might be less meaningful in low-income countries compared to high-low-income countries as the audience is smaller and has less
disposable income. From a finance perspective, one can argue that the value creation that firms hope to stimulate via social media marketing is also depending on consumers’ economic environment. Promoting luxury goods via social media in a country with a low standard of living might have a smaller or no effect on firm value than doing the same promotion in a country which has a high standard of living. Generalizing this argument leads to the third hypothesis, namely:
3. Research design
3.1 Sample
Previous research focusing on social media has mostly taken a sample from a single industry, for example luxury brands (Godey et al., 2016; Kim and Ko, 2012), books (Chevalier and Mayzlin, 2006; Forman et al., 2008), movies (Liu 2006, Chintagunta et al., 2010), and Computer hardware and software industry (Luo et al., 2013). However, as mentioned in section 2, there is reason to expect country and industry specific difference. Therefore, this study uses a cross industry and cross country sample containing 1981 listed companies from 11 different countries (UK, US, France, Australia, Norway, Germany, Netherlands, Portugal, Sweden, Spain, Denmark) and 18 different industries (table 2). More than half of the sample consists of companies from native English speaking countries. Namely, 542 companies from the UK, 492 from the US and 178 companies from Australia. The remaining 769 companies are from northern and western European countries mainly represented by 291 companies from France, 117 from Germany and 108 from Norway. The main sample containing 1981 companies is used to create two different sub-samples, a Facebook sample and a Twitter sample. From the 1981 companies 1175 use Twitter and 927 use Facebook. The removal of duplicates, outliers, and a lack of data regarding sales numbers decreases the number of observations used in the regression models to 821 for the Twitter sample and 630 for the Facebook sample. A more detailed country specific sample overview from the main sample is given in Table 1 and an industry specific overview in Table 2.
TABLE 1
Country specific sample overview
Country Sample size Number of companies using Facebook
Number of companies using Twitter Australia 178 56 83 Germany 117 67 71 Denmark 25 14 15 Spain 35 18 27 France 291 128 172 United Kingdom 542 165 269 Netherlands 74 38 47 Norway 108 37 35 Portugal 60 16 9 Sweden 46 26 32 United States 492 362 415 Total 1968 927 1175
TABLE 2
Industry specific sample overview
Industry Sample size Facebook profiles Twitter profiles
1. Banks 81 49 55
2. Chemicals, rubber, plastics, non-metallic products
190 78 104
3. Construction 57 29 38
4. Education, Health 35 11 14
5. Food, beverages, tobacco 71 31 47
6. Gas, Water, Electricity 67 37 47
7. Hotels & Restaurants 34 15 21
8. Insurance Companies 23 19 19
9. Machinery, equipment, furniture, recycling 300 176 208
10. Metals & metal products 82 22 28
11. Other services 548 234 314
12. Post & Telecommunication 52 27 34
13. Primary sector 92 32 36
14. Publishing, printing 72 40 54
15. Textiles, wearing apparel, leather 27 16 17
16. Transport 70 27 40
17. Wholesale & Retail trade 143 73 88
18. Wood, Cork, Paper 23 11 11
Total 1968 927 1175
This table shows the detailed industry composition of the 1968 listed sample companies and the Facebook and Twitter sub-samples. The industry classification is based on Orbis BvD major sector classes.
3.2 Variable descriptions
Most of the financial data has been collected from the Orbis database. However, the Orbis data for some variables, such as book value of assets and market capitalization, was incomplete for 355 US companies. The Datastream database has been used to fill the gab for those companies. There were still some missing accounting values that could not be derived from Orbis and Datastream, so they have been added manually from annual reports. All financial accounting variables have been collected in thousands US dollars and represent 2015 year end values.
Dependent variables
advantages of social media lie in the future. Therefore, a forward looking firm value measure is needed that reflects market based views of investor expectations of firms’ future profit potential. According to Srinivasan and Hanssens (2009), market capitalization and market-to-book ratio are both appropriate measures to represent firm value in this scenario. The market-to-book value has the advantage that values greater than 1 signal a contribution of intangible assets. It also allows direct comparison across industries (Srinivasan and Hanssens, 2009).
Market capitalization is computed by multiplying the stock price with the number of outstanding shares and will be represented by the variable MVE. The variable MVE is transformed to the
logarithm with base ‘e’, e.g. the natural logarithm, to control for skewness. The market to book ratio is calculated as market capitalization divided by the book value of equity and will be represented by the variable MTB.
Independent variables
To test hypothesis 1, that there is a significant correlation between social media popularity and firm value, social media popularity will be measured by the number of likes for Facebook and the number of followers for Twitter data. Facebook’s ‘likes’ and Twitter’s ‘follower’ can be considered to be a better indicator than web traffic, as they indicate not only the potential interest, represented by the visit, but also imply, express, and buzz loyalty and satisfaction. Moreover, unlike web traffic metrics, likes and follower are not anonymous and can only be registered once for each target. The number of likes on Facebook and the number of followers on twitter can thus be seen as sentiment score
summarizing the peoples opinion about the content of companies’ profiles. The corresponding variables are named FLikes and TFollower and are transformed to the natural logarithm to control for their skewness.
addition, it is controlled for sales growth represented by the variable Growth and measured by the percentage increase in sales from 2014 to 2015. The variable AGE is measured by the number of years between the date of incorporation and 2016.
Other research shows that there are industry specific differences in performance and level of R&D investments that might influence firm value (Graves and Waddock, 1994). To control for those industry specific effects, industry dummies are defined as Du_’X’, where ‘x’ corresponds to the defined industry number as in Table 2. To deal with country specific characteristics which influence firm value, country dummies are used and denoted as Du_’CC’, where CC is representative for the corresponding country codes.
For hypothesis 2, which suggest that the degree of social media profile activity influences popularity and firm value, variables TTweets and Fphoto are introduced to represent the number of tweets posted on Twitter and the number of photos posted on Facebook. Both variables are transformed to their natural logarithm to control their skewness.
GDP per capita is used as a proxy for the countries’ standard of living that is used in hypothesis 3. Values have been collected from the World Development Indicators database provided by World Bank and correspond to 2015. The natural logarithm is taken to adjust for skewness and to avoid high scaled values in the interaction terms.
3.3 Descriptive statistics
A summary of the statistics of the variables used in this study is provided in table 3. The mean of the natural logarithms of the market capitalization of companies that use Twitter is 22.20. The
corresponding value for the Facebook sample is 29.13. The market-to-book ratio is 3.90 on average for the Twitter sample and 3.68 for the Facebook sample. Companies of both samples have a capital structure that consists of 59% debt on average. Social media variables show that companies registered earlier at Twitter than Facebook. On average companies registered 6.56 years ago on Twitter and 5.51 years ago on Facebook. The variables Tverified and Fverified show that only 40% and 43% of the companies used the verifying process offered by Twitter and Facebook.
market-to-book ratio between Twitter using firms and non Twitter using firms can be find in Australia, UK, Norway, Portugal and the US. In all those countries, firms that use Twitter have a higher mean market-to-book ratio. This is a first indication that social media usage might indeed affect firm value. Using Twitter might be seen by investors as an intangible asset that might justify a higher valuation compared to non Twitter using firms.
This table provides a summary of the descriptive statistics of the Twitter sample and Facebook sample. On average the market capitalization of the sample firms using Facebook is bigger compared to the Twitter sample. As more companies use Twitter its sample consists mainly of 1122 observations for all variables except of growth. The number of observations is 863 except for the growth proxy. Both samples are quite similar regarding size, leverage, growth and age.
TABLE 3
Summary statistics
Twitter sample
Variable Mean Median Maximum Minimum Std. Dev. Observations
lnMVE 22.20 22.65 27.16 15.16 2.13 1122 MTB 3.90 2.54 42.72 0.07 4.67 1122 lnTFollower 8.81 8.61 16.45 0.69 2.25 1120 lnTTweets 7.62 7.69 14.19 0.00 1.84 1118 Treg 6.56 7 11 1 2.02 1122 Tverified 0.40 0.00 1.00 0.00 0.49 1122 lnTA 15.42 15.67 21.60 8.76 2.34 1122 LEV 0.59 0.60 0.99 0.01 0.19 1122 Growth 0.00 -0.016 1.95 -0.88 0.21 830 AGE 56.98 36 362 1 51.40 1122 Facebook sample
Variable Mean Median Maximum Minimum Std. Dev. Observations
This table compares firms that use social media with firms that do not use it based on both firm value measures lnMVE and MTB. Firms using Twitter have a significantly higher market-to-book ratio in Australia, UK, Norway Portugal and the US. Firms from the UK using Facebook are valued higher based on the market-to-book ratio than firms that do not use Facebook. Significant differences are marked with *, **, *** which corresponds to the statistical 10%, 5%, and 1% significance level.
Table 4
Sample comparison of social media users and non-users based on country specific test of equality of means for market capitalization (lnMVE) and market-to-book ratio (MTB).
Twitter Facebook
Country Twitter No Twitter
Difference Facebook No Facebook Difference
3. 4 Measurement
Hypothesis 1 is tested using a cross-sectional multivariate OLS regression model which has been used in a similar way by other researchers before (Fornell et al., 2006). To see whether the social media variables add explanatory power to the whole model, first only the control variables are regressed and then the social media variables are added step by step. This way, a change of the explanatory power of the model can be examined. To avoid a dummy trap the country dummy for the US and the industry dummy 11 (other services) are used as a base and dropped from the regression model. To test hypothesis 2, whether the relationship between social media popularity and firm value is affected by the degree of social media activity, variables lnTTweets and lnFphoto and their interaction effects with lnTFollower and lnFlikes are added to the regression models (Model 2.1-2.3). Hypothesis 3 is tested by using an interaction term consisting of GDP per capita multiplied with the social media popularity variables lnTFollower for Twitter and lnFlikes. The regression models used hypotheses testing are summarized as follows:
Regression models for Twitter sample to test hypothesis 1:
𝑙𝑛𝑀𝑉𝐸𝑖 = 𝛼 + 𝛽1 𝑙𝑛𝑇𝐹𝑜𝑙𝑙𝑜𝑤𝑒𝑟𝑖+ 𝛽2𝑙𝑛𝐵𝑉𝐴𝑖+ 𝛽3𝐿𝐸𝑉𝑖+ 𝛽4𝐺𝑟𝑜𝑤𝑡ℎ𝑖+ 𝛽5𝐴𝑔𝑒𝑖 + ∑ 𝛽𝑖 16 𝑖=6 𝐶𝑜𝑢𝑛𝑡𝑟𝑦𝑖+ ∑ 𝛽𝑖𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦𝑖 34 𝑖=17 + 𝜀𝑖 (1.1) 𝑀𝑇𝐵𝑖= 𝛼 + 𝛽1 𝑙𝑛𝑇𝐹𝑜𝑙𝑙𝑜𝑤𝑒𝑟𝑖+ 𝛽2𝑙𝑛𝐵𝑉𝐴𝑖+ 𝛽3𝐿𝐸𝑉𝑖+ 𝛽4𝐺𝑟𝑜𝑤𝑡ℎ𝑖+ 𝛽5𝐴𝑔𝑒𝑖 + ∑ 𝛽𝑖 16 𝑖=6 𝐶𝑜𝑢𝑛𝑡𝑟𝑦 + ∑ 𝛽𝑖𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 34 𝑖=17 + 𝜀𝑖 (1.2)
Regression models for Facebook sample to test hypothesis 1:
Regression models for Twitter sample to test hypothesis 2: 𝑙𝑛𝑀𝑉𝐸𝑖 = 𝛼 + 𝛽1 𝑙𝑛𝑇𝐹𝑜𝑙𝑙𝑜𝑤𝑒𝑟𝑖+ 𝛽2 𝑙𝑛𝑇𝑇𝑤𝑒𝑒𝑡𝑠𝑖+ 𝛽3(𝑙𝑛𝑇𝐹𝑜𝑙𝑙𝑜𝑤𝑒𝑟𝑖∗ 𝑙𝑛𝑇𝑇𝑤𝑒𝑒𝑡𝑠𝑖) + 𝛽4𝑙𝑛𝐵𝑉𝐴𝑖 + 𝛽5𝐿𝐸𝑉𝑖+ 𝛽6𝐺𝑟𝑜𝑤𝑡ℎ𝑖+ 𝛽7𝐴𝑔𝑒𝑖+ ∑ 𝛽𝑖 18 𝑖=8 𝐶𝑜𝑢𝑛𝑡𝑟𝑦𝑖+ ∑ 𝛽𝑖𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦𝑖 36 𝑖=19 + 𝜀𝑖 (2.1) 𝑀𝑇𝐵𝑖 = 𝛼 + 𝛽1 𝑙𝑛𝑇𝐹𝑜𝑙𝑙𝑜𝑤𝑒𝑟𝑖+ 𝛽2 𝑙𝑛𝑇𝑇𝑤𝑒𝑒𝑡𝑠𝑖+ 𝛽3(𝑙𝑛𝑇𝐹𝑜𝑙𝑙𝑜𝑤𝑒𝑟𝑖∗ 𝑙𝑛𝑇𝑇𝑤𝑒𝑒𝑡𝑠𝑖) + 𝛽4𝑙𝑛𝐵𝑉𝐴𝑖 + 𝛽5𝐿𝐸𝑉𝑖+ 𝛽6𝐺𝑟𝑜𝑤𝑡ℎ𝑖+ 𝛽7𝐴𝑔𝑒𝑖+ ∑ 𝛽𝑖 18 𝑖=8 𝐶𝑜𝑢𝑛𝑡𝑟𝑦𝑖+ ∑ 𝛽𝑖𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦𝑖 36 𝑖=19 + 𝜀𝑖 (2.2)
Regression models for Facebook sample to test hypothesis 2:
Regression models for Twitter sample to test hypothesis 3:
Regression models for Facebook sample to test hypothesis 3:
4. Results
Before, discussing the regression results one can gain valuable insights by reviewing the sample overview and the descriptive statistics. More companies use Twitter than Facebook, despite the fact that Facebook is older and has way more active users (Table 1). All in all, Twitter is used by 60% of the sample companies and Facebook only by 47% . This finding is not in alignment with previous research. Kim and Ko (2013) state that in 2010 90% of the luxury brand industry companies were active on Facebook and 48% on Twitter. This indicates that generalizing findings from one industry to others does not work for social media research. From the 1986 sample companies there are 844 companies that use both Facebook and Twitter. Twitter is used by more companies than Facebook in all countries except for Norway and Portugal. The general usage of social media strongly differs per country. In the United States 84% of the companies use Twitter and 73% use Facebook. In Contrast, only 15% of the sample companies in Portugal use Twitter and 26% use Facebook. That only few Portuguese firms use social media as a marketing tool can be explained with findings from Tiago and Veríssimo (2014). They find that only 37% of managers of large Portuguese firms see a link between digital presence internal marketing. They would only implement social media due to external
pressures. Social media also differs between industries. Social media is widely used in the insurance and banking industry. 82% of the sample companies from the insurance industry uses Facebook and Twitter. In the banking industry 60% and 67% use Facebook and Twitter. An explanation might be that product differentiation is very limited in those industries. One way to distinguish one selves from competition is via different customer experience and customer relationships. The banking and
insurance industry nowadays has less face to face communication with their customers due to digitalization. Therefore, social media might be especially important for those industries to not lose contact to customers and retain good customer relationships. In the metals and metal products industry the least companies are using social media with 27% using Facebook and 34% using Twitter. In the
primary sector, social media also seem to be less popular with 35% using Facebook and 39% using Twitter. This are both industries that mostly produce intermediate products that are not directly sold to end-consumers. The benefit of using social media might therefore be smaller for business to business industries. Implementing social media might be more risky for companies from the primary sector as the whole industry gets criticized a lot for exploiting natural resources and create pollution.
Twitter is not only used more widely, but companies also started to implement it earlier (Table 3).
Adding the social media popularity variables to the OLS regression model increased the goodness of fit of all 4 models. The correlation table shows that there is a relatively high positive correlation between market capitalization and book value of assets, which is in line with earlier research (Fornell et al., 2006). The OLS regression estimation results support hypothesis 1, that social media popularity significantly correlates with firm value. The estimated coefficient of lnTFollower is positive and significant for lnMVE and market-to-book ratio on the 1% and 5% level respectively (Table 5). Facebook likes show similar coefficients which are also both positive and significant on the 1% level for lnMVE and 5% level for MTB. The null-hypothesis that social media popularity is not affecting firm value can thus be rejected. This means that firms that are more popular on Facebook and Twitter have higher market capitalizations and market-to-book ratios than their less popular firms. Firms that have 1% more follower on Twitter have a 0.06% higher market capitalization. The regression models involving market capitalization (lnMVE) as the explained variable have an adjusted R2 of 0.88. The regression models explaining the market-to-book ratio have a relatively small adjusted R2 of 0.15. The estimated coefficients might be less precise. One would interpret them as firms having 1% more follower have a 0.0019 higher market-to-book ratio. For Facebook the market-to-book ratio is 0.0015% higher for firms that have 1% more likes.
Adding the social media activity variables lnTTweets and lnFhotos and their interaction term with lnTFollower and lnFlikes to the regression models to test hypothesis 2 implies a problem of
multicollinearity as can be seen in Table 7 and 8. The social media activity variables lnTTweets and lnFphotos are highly correlated with social media popularity variables lnTFollower and lnFlikes. This implies that the coefficient estimates of regression 2.1-2.4 can be less precise (Table 6). Regression results indicate that the null hypothesis that social media activity is influencing the effect of social media popularity on firm value can be rejected.
Regression results for hypothesis 3 show no significant coefficient estimate for the interaction terms of lnGDP and social media popularity. Therefore, hypothesis 3 can be rejected. Moreover, adding lnGDP and its interaction term with social media popularity to the regression model diminishes the
*, **, *** indicate significance at the 10%, 5%, and 1% level. This table shows regression results of Twitter follower, Facebook likes, and control variables on market capitalization and market-to-book ratio. For each variable the estimated coefficient and its significance level are shown. The null hypothesis that the estimated coefficient is equal to zero can be rejected for most explanatory variables. The corresponding t-statistic are given in parenthesis. The country dummy for the US and the Industry dummy for ‘Other Services’ are not included in this regression to avoid a dummy trap.
Table 5
Regression results for hypothesis 1
Twitter Facebook (1) lnMVE (2) MTB (3) lnMVE (4) MTB lnTFollower 0.06*** (4.30) 0.19** (2.54) lnFlikes 0.05*** (3.91) 0.15** (2.26) lnBVA 0.82*** (50.27) -0.63*** (7.99) 0.86*** (42.36) -0.51*** (-4.89) LEV -1.43*** (-9.43) 6.42*** (7.99) -1.83*** (-9.55) 5.82*** (5.93) Growth 0.45*** (3.59) 1.83*** (1.83) 0.68*** (4.24) 2.75*** (3.34) AGE -0.00 (-1.86) -0.006* (-2.20) -0.001* (-2.25) -0.005 (-1.45)
Country Yes Yes Yes Yes
Industry Yes Yes Yes Yes
Table 6
Regression results hypothesis 2
Twitter Facebook (1) lnMVE (2) MTB (3) lnMVE (4) MTB lnTFollower 0.07 (1.35) 0.16 (0.58) lnTTweets 0.02 (0.49) 0.17 (0.65) (lnTFollower * lnTTweets) -0.002 (-0.36) -0.007 (-0.23) lnFlikes 0.07** (1.97) -0.00 (-0.02) lnFphotos -0.06 (-1.28) -0.53** (-1.99) (lnFlikes * lnFphotos) 0.00 (0.16) 0.03 (1.45) lnBVA 0.82*** (49.30) -0.62*** (-7.08) 0.86*** (42.07) -0.49*** (-4.73) LEV -1.45*** (-9.50) 6.32*** (7.82) -1.81*** (-9.44) 5.97*** (6.07) Growth 0.45*** (3.58) 1.83*** (2.73) 0.68*** (4.26) 2.77*** (3.36) AGE -0.001 (-1.82) -0.01* (-2.21) -0.001** (-2.32) -0.00 (-1.41)
Country Yes Yes Yes Yes
Industry Yes Yes Yes Yes
Constant 10.15*** (25.50) 8.63*** (4.11) 16.91*** (45.85) 9.40*** (4.96) Observations 823 823 630 630 Adjusted R² 0.87 0.16 0.88 0.16
Table 7
Regression results hypothesis 3
Twitter Facebook (1) lnMVE (2) MTB (3) lnMVE (4) MTB lnTFollower 0.63 (1.17) -0.87 (0.74) (lnTFollower * lnGDP) -0.05 (-1.03) 0.10 (0.40) lnFlikes -0.63 (-1.34) 3.21 (1.38) lnGDP 1.41*** (3.29) 0.97 (0.44) 0.43 (0.97) 4.59** (2.09) (lnFlikes * lnGDP) 0.06 (1.44) -0.28 (-1.31) lnBVA 0.86*** (53.53) -0.51*** (-6.30) 0.92*** (47.86) -0.36*** (-3.83) LEV -1.55*** (-9.75) 6.21*** (7.65) -2.05*** (-10.30) 5.40*** (5.57) Growth 0.52*** (3.94) 2.22*** (3.30) 0.84*** (5.05) 3.17*** (3.85) AGE -0.001*** (2.86) -0.01* (0.01) -0.002*** (-3.07) -0.007** (-1.98)
Industry Yes Yes Yes Yes
Constant -6.15*** (-1.33) -4.70 (-0.20) 15.55*** (68.38) 4.30*** (3.83) Observations 821 821 630 630 Adjusted R² 0.87 0.13 0.86 0.13
5. Conclusion, managerial implications, and limitations
This study shows that firm’s social media popularity on Facebook and Twitter significantly influences market capitalization and market-to-book ratio. The amount of tweets and posted photos is positively correlated to popularity scores but does in turn not add explanatory power to the relationship between social media popularity and firm value. The degree of social media activity does thus not seem to influence the relationship between social media popularity and firm value. Managers should be aware of the risks and advantages of implementing social media as a marketing tool. Investors could monitor social media sentiment metrics such as Facebook likes and Twitter followers and use their growth rates to predict trends, future sales, and value creation. One could even create a portfolio of firms that perform great on social media. Furthermore, this study shows that the economic environment
measured by GDP per capita does not influence the relationship between social media popularity and firm value. Results give reason to monitor social media metrics closer and constantly over time. Unfortunately, social media data is mostly kept private by the network providers or by specialized firms that use advanced datamining techniques to create their own databases. Time series datasets are sold by those companies for thousands of dollar. I hope that in the future the access to social media data will be more balanced between private firms, the public, and universities, as social media is becoming the biggest and probably most powerful database about humankind.
TABLE 8
Correlation table Twitter sample
lnMVE MTB lnTFollower lnTTweets Treg Tverified lnTA LEV Growth AGE lnGDP
lnMVE 1 MTB 0.10 1 lnTFollower 0.54 0.04 1 lnTTweets 0.37 0.07 0.79 1 Treg 0.22 0.02 0.53 0.57 1 Tverified 0.45 0.03 0.64 0.55 0.27 1 lnBVA 0.89 -0.14 0.50 0.35 0.47 0.47 1 LEV 0.23 0.12 0.18 0.19 0.08 0.20 0.43 1 Growth 0.17 0.08 0.25 0.18 0.10 0.14 0.12 0.01 1 AGE 0.20 -0.12 0.09 0.08 0.07 0.10 0.27 0.15 -0.03 1 lnGDP 0.19 0.06 0.11 0.01 0.14 -0.01 0.06 -0.11 -0.02 -0.01 1
This table shows the correlations between all major explanatory variables used in this study and the independent variables lnMVE and market-to-book ratio for the Twitter sample.
TABLE 9
Correlation table Facebook sample
lnMVE MTB lnFlikes lnFphoto Freg Fverified lnGDP lnTA LEV Growth AGE
lnMVE 1 MTB 0.16 1 lnFlikes 0.50 0.11 1 lnFphoto 0.32 0.088 0.70 1 Freg 0.17 0.12 0.39 0.36 1 Fverified 0.34 0.10 0.63 0.44 0.21 1 lnGDP 0.20 0.08 0.06 -0.05 0.05 0.05 1 lnBVA 0.87 -0.09 0.47 0.31 0.12 0.32 0.07 1 LEV 0.18 0.08 0.15 0.13 0.02 0.10 -0.10 0.43 1 Growth 0.01 0.19 0.03 0.04 0.09 0.04 0.03 -0.12 -0.14 1 AGE 0.22 -0.13 0.15 0.10 0.01 0.08 -0.01 0.31 0.11 -0.18 1
6. References
Anderson, R.C. and Reeb, D.M., 2003. Founding‐family ownership and firm performance: evidence from the S&P 500. The journal of finance, 58(3), pp.1301-1328.
Agnihotri, R., Dingus, R., Hu, M.Y. and Krush, M.T., 2016. Social media: Influencing customer satisfaction in B2B sales. Industrial Marketing Management, 53, pp.172-180.
Aivazian, V.A., Ge, Y. and Qiu, J., 2005. The impact of leverage on firm investment: Canadian evidence. Journal of corporate finance, 11(1), pp.277-291.
Alexa, 2016. The top 500 sites on the web. [online] Available at: http://www.alexa.com/topsites [Accessed, 12. Oct. 2016]
Anderson, E.W., Fornell, C. and Mazvancheryl, S.K., 2004. Customer satisfaction and shareholder value. Journal of marketing, 68(4), pp.172-185.
Andrés, L., Cuberes, D., Diouf, M. and Serebrisky, T., 2010. The diffusion of the Internet: A cross-country analysis. Telecommunications Policy, 34(5), pp.323-340.
Aral, S., Dellarocas, C. and Godes, D., 2013. Introduction to the special issue-social media and business transformation: A framework for research.Information Systems Research, 24(1), pp.3-13. Azar, P.D. and Lo, A.W., 2016. The wisdom of twitter crowds: Predicting stock market reactions to fomc meetings via twitter feeds. The Journal of Portfolio Management, 42(5), pp.123-134.
Barth, M.E. and McNichols, M.F., 1994. Estimation and market valuation of environmental liabilities relating to superfund sites. Journal of Accounting Research, pp.177-209.
Benthaus, J., Risius, M. and Beck, R., 2016. Social media management strategies for organizational impression management and their effect on public perception. The Journal of Strategic Information
Systems, 25(2), pp.127-139.
Bolton, R.N., Parasuraman, A., Hoefnagels, A., Migchels, N., Kabadayi, S., Gruber, T., Komarova Loureiro, Y. and Solnet, D., 2013. Understanding Generation Y and their use of social media: a review and research agenda. Journal of Service Management, 24(3), pp.245-267.
Carl, W.J., 2006. What's all the buzz about? Everyday communication and the relational basis of word-of-mouth and buzz marketing practices.Management Communication Quarterly, 19(4), pp.601-634.
Chintagunta, P.K., Gopinath, S. and Venkataraman, S., 2010. Online word-of-mouth effects on the offline sales of sequentially released new products: An application to the movie market. Marketing
Science, 29(5), pp.944-957.
Choi, H.K. and Limb, J.O., 1999. A behavioral model of web traffic. In Network Protocols,
1999.(ICNP'99) Proceedings. Seventh International Conference, pp. 327-334.
Demers, E. and Lev, B., 2001. A rude awakening: Internet shakeout in 2000.Review of Accounting
Studies, 6(2-3), pp.331-359.
De Vries, L., Gensler, S. and Leeflang, P.S., 2012. Popularity of brand posts on brand fan pages: An investigation of the effects of social media marketing. Journal of Interactive Marketing, 26(2), pp.83-91.
Duan, W., Gu, B. and Whinston, A.B., 2008. Do online reviews matter?—An empirical investigation of panel data. Decision support systems, 45(4), pp.1007-1016.
Erdoğmuş, İ.E. and Cicek, M., 2012. The impact of social media marketing on brand loyalty. Procedia-Social and Behavioral Sciences, 58, pp.1353-1360.
Facebook, 2016. Company Inf. [online] Available at: http://newsroom.fb.com/company-info/ [Accessed, 11 Oct. 2016].
Fama, E.F., 1970. Efficient capital markets: A review of theory and empirical work. The journal of Finance, 25(2), pp.383-417.
Forman, C., Ghose, A. and Wiesenfeld, B., 2008. Examining the relationship between reviews and sales: The role of reviewer identity disclosure in electronic markets. Information Systems
Research, 19(3), pp.291-313.
Forman, C., Goldfarb, A. and Greenstein, S., 2005. Geographic location and the diffusion of Internet technology. Electronic Commerce Research and Applications, 4(1), pp.1-13.
Fornell, C., Mithas, S., Morgeson III, F.V. and Krishnan, M.S., 2006. Customer satisfaction and stock prices: High returns, low risk. Journal of marketing, 70(1), pp.3-14.
Godey, B., Manthiou, A., Pederzoli, D., Rokka, J., Aiello, G., Donvito, R. and Singh, R., 2015, June. Luxury brands social media marketing efforts: Influence on brand equity and consumer behavior. In 2015 Global Fashion Management Conference at Florence, pp. 68-68.
Internet World Stats, 2016. World Internet Users and 2016 Population Stats. [online] Available at: http://www.internetworldstats.com/stats.htm [Accessed, 11 Oct. 2016].
Ittner, C.D. and Larcker, D.F., 1998. Are nonfinancial measures leading indicators of financial performance? An analysis of customer satisfaction. Journal of accounting research, 36, pp.1-35. Keating, E.; Lys, T.; and Magee, R., 2003. The Internet downturn: Finding valuation factors in spring 2000. Journal of Accounting and Economics, 34, pp.189–236.
Liu, Y., 2006. Word of mouth for movies: Its dynamics and impact on box office revenue. Journal of
marketing, 70(3), pp.74-89.
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.
Kaplan, A.M. and Haenlein, M., 2010. Users of the world, unite! The challenges and opportunities of Social Media. Business horizons, 53(1), pp.59-68.
Kim, A.J. and Ko, E., 2012. Do social media marketing activities enhance customer equity? An empirical study of luxury fashion brand. Journal of Business Research, 65(10), pp.1480-1486.
Kietzmann, J.H., Hermkens, K., McCarthy, I.P. and Silvestre, B.S., 2011. Social media? Get serious! Understanding the functional building blocks of social media. Business horizons, 54(3), pp.241-251.
Madden, T.J., Fehle, F. and Fournier, S., 2006. Brands matter: An empirical demonstration of the creation of shareholder value through branding. Journal of the Academy of Marketing Science, 34(2), pp.224-235.
Maury, B. and Pajuste, A., 2005. Multiple large shareholders and firm value. Journal of Banking & Finance, 29(7), pp.1813-1834.
Ngai, E.W., Tao, S.S. and Moon, K.K., 2015. Social media research: Theories, constructs, and conceptual frameworks. International Journal of Information Management, 35(1), pp.33-44.
Resnick, P. and Zeckhauser, R., 2002. Trust among strangers in internet transactions: Empirical analysis of ebay’s reputation system. The Economics of the Internet and E-commerce, 11(2), pp.23-25.
Seiders, K. and Riley, E.G., 1999. Stock price performance of Internet firms: Identification of key drivers. Frontiers of Entrepreneurship Research , 392–404.
Senecal, S. and Nantel, J., 2004. The influence of online product recommendations on consumers’ online choices. Journal of retailing, 80(2), pp.159-169.
Statista, 2016. Average daily time spent on social media worldwide 2012-2016. [online] Available at: https://www.statista.com/statistics/433871/daily-social-media-usage-worldwide/ [Accessed, 11 Oct. 2016].
Stelzner, M.A., 2011. Social media marketing industry report. Social Media Examiner, 41.
Srinivasan, S. and Hanssens, D.M., 2009. Marketing and firm value: Metrics, methods, findings, and future directions. Journal of Marketing research, 46(3), pp.293-312.
Sun, K.A. and Kim, D.Y., 2013. Does customer satisfaction increase firm performance? An application of American Customer Satisfaction Index (ACSI). International Journal of Hospitality
Management, 35, pp.68-77.
Tiago, M.T.P.M.B. and Veríssimo, J.M.C., 2014. Digital marketing and social media: Why bother?. Business Horizons, 57(6), pp.703-708.
Tian, Y., & Chen, J. (2009). Concept of voluntary information disclosure and a review of relevant studies. International Journal of Economics and Finance, 1(2), 55.
Tsai,W. H. S., &Men, L. R. (2013). Motivations and antecedents of consumer engagement with brand pages on social networking sites. Journal of Interactive Advertising, 13(2), pp.76–87.
Twitter Inc. 2016, Twitter homepage, 28 August 2016. Available from: https://about.twitter.com/company [28 August2016].
Van Belleghem, Steven, Marloes Eenhuizen, and Elias Veris (2011), Social Media Around the World 2011. InSites Consulting. Retrieved 20-12-2016 from
Waddock, S.A. and Graves, S.B., 1997. The corporate social performance-financial performance link. Strategic management journal, pp.303-319.
