When Social Media Channels Meet: The Effects of Using Social Media
Channels in Interaction on Big Data Analytics
Master Thesis for MSc BA Change Management
Faculty of Economics and Business
University of Groningen
Name: Romée Reef
Student number: S2517221
Date: June 21st 2019
Supervisor/co-assessor: John Dong/Jordi Surroca
2
ABSTRACT
Though the direct effects of social media channels on the potential for the analysis of big data have been researched thoroughly, the phenomenon of interactive social media channel usage has not received much attention. The current study addresses this gap in the literature by researching both the direct and interaction effects of four categories of social media channels – social networks, blogs, multimedia communities, and wikis – on big data analytics. The central argument of this study is that when interactively using these four channels, firms can create super-additive value. By applying a systems perspective, I propose that synergies can be created through the usage of the four
complementary social media channels together, which, in turn, leads to an increased use of big data analytics. Furthermore, this study tests the influence of big data analytics on firm performance. I empirically test the individual and interaction effects of social networks, blogs, multimedia
communities, and wikis on firms’ big data analytics using an Italian database with a final sample of 8,513 firms. The study’s main findings are a positive effect of the four-way interaction of social media channels on big data analytics, and a positive effect of big data analytics on firm performance.
3
INTRODUCTION
In today’s rapidly changing business environment, firms should not only rely on their own Research & Development departments, but they should engage in open innovation - the practice of leveraging the discoveries of others (Chesbrough & Crowther, 2006; Lee & Cole, 2003). It is said that the role customers can play in open innovation is of importance in developing new products (Sandmeier, Morrison, & Gassmann, 2010; Lee & Cole, 2003). Since individual consumers are seen as an
important source for innovative ideas (Nambisan, 2002), many firms try out platforms through which these suggestions can be gathered. The Internet offers such a modern and often used platform to generate ideas from customers; and especially since the emergence of social media technologies, firms have increasingly used equivalent communities the last number of years to leverage ideas about innovation (Dong & Wu, 2015; Bugshan, 2015). The usage of social media channels has grown rapidly in the last decade, as not only young people are using it, but it is also increasingly used by older generations (Kaplan & Haenlein, 2010). Such channels can thus play a big role in capturing relevant information from external parties.
Since social media is used by over a billion people around the world, it generates an overwhelming amount of unstructured data in a relatively short time – big data (Ghani, Hamid, Targio Hashem, & Ahmed, 2018; Sebei, Hadj, & Ben, 2018). Given that there are a lot of social media channels used worldwide, a huge amount of data is generated in just seconds (Sebei et al., 2018). Being able to analyze this big data could create advantages of dramatic cost reductions, improvements in the time to perform a computing task, and the development of new products or services (Fang, Zhang, Wang, Daneshmand, Wang, & Wang, 2015). Big data is considered as technologies useful for managing a huge amount of data that cannot be managed through traditional technologies (Sebei et al., 2018). It seems that social media channels are full of potential for applying corresponding technologies to mine and analyze data (Felt, 2016), since such channels are considered typical examples of big data sources (Bello-Orgaz, Jung, & Camacho, 2016). The data is generated from a wide number of Internet applications and Web sites, among others, Facebook, Twitter, LinkedIn, YouTube, and Instagram (Bello-Orgaz et al., 2016).
4
social media channels. Rutter, Roper, and Lettice (2016) find that multi-channel usage of the two largest social media, Facebook and Twitter, can create synergy between the platforms and increase student recruitment for universities. Also recently, the first attention has been given to diversity of social media usage. Dong and Yang (2019) focus on the interaction of social media diversity and big data analytics on firm performance, and find a positive effect. So although the link between the usage of social media channels and big data analytics by firms is widely acknowledged, to the author’s knowledge, previous studies ignored the potential that the complementarity of different social media channels can have for the analysis of big data. It appears a little odd that the effect that social media interaction can have has not been thoroughly researched to date, since literature does recognize the wide variety of data types and analytic techniques that stem from different social media channels (e.g. Sebei et al., 2018; Stiegliz, Mirbabaie, Ross, & Neuberger, 2018). This study proposes that through the interactive use of four complementary social media channels, firms may be more inclined to apply big data analytics to leverage information from potential customers.
The goal of the present research is to overcome the mentioned gap in the literature by testing all direct and interaction effects of four social media channels, namely social networks, blogs, multimedia communities, and wikis, on big data analytics. The study uses a systems perspective and proposes that interactively using four complementary channels has the biggest potential for the creation of synergies – the phenomenon of super-additive value resulting from the interactions among components of a system (Tanriverdi & Venkatraman, 2005). Super-additive value exists when combinative value of multiple components is greater than the sum of each component’s individual value (Dong & Yang, 2019). The central argument of the study is thus that interactions of the four social media channels are of greater added value than summing up their individual contributions. This study uses survey data from the Italian National Institute of Statistics (Istat) to examine the effect of the different types and interactions of social media channels on firms’ big data analytics.
5
This study is also of important business value, as it will shed light on how different opportunities for social media usage should be addressed by firms. As increasing amounts of big data are available, it creates enormous technological changes that allow firms to analyze more and more data. It is therefore important for managers to discover how to create value from different available technologies, social media channels in this case. As prior research shows that using social media is of increasing
importance to leverage customer ideas, it is imperative to show which social media channels generate most beneficial effects, and what the added value is of using social media channels in interaction. It is of significant managerial interest to discover whether combining complementary social media
channels can lead to super-additive value for firms.
The study’s research questions are as follows: 1) what will be the effect of interactions of social media channels on a firm’s use of big data analytics? And then, 2) how will big data analytics influence firm performance?
This paper proceeds as follows. First, the research background will explain previous research in this area and lay out the paper’s main concepts. Then, the theoretical foundation of this study, the systems theory, will be addressed, followed by the development of the study’s hypotheses. The subsequent method section will describe the data that was used, as well as the analysis strategy. Then, in the results section, the main outcomes of the analyses will be discussed, as well as the results of the hypotheses testing. Finally, in the discussion chapter, the study’s summary, implications, and limitations will be laid out, as well as directions for future research.
RESEARCH BACKGROUND
6 Social media
Social media are 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 & Haenlein, 2010). The definition of Kaplan and Haenlein (2010) is built on eliminating the confusion between the terms social media, Web 2.0 technologies, and user generated content. Web 2.0 is described as a new way in which the Internet was being used from the early 2000s, whereby content and applications on platforms were no longer created and published by individuals, but constantly modified by all Internet users in a collaborative fashion (Kaplan & Haenlein, 2010). Web 2.0 is considered as the platform that caused the evolution of social media. Through Web 2.0 technologies, individuals can constantly modify content on the world wide web (Kaplan & Haenlein, 2010), creating user generated content (UGC). UGC is described as the various forms of media content that are publicly available and created by end users (Kaplan & Haenlein, 2010). It is everything that is produced in the moment of being social (Kaplan & Haenlein, 2010; Smith et al., 2012), and therefore social media allows for the generation of it. Analyzing such content requires social media analytics, which has subsequently risen after the onset of Web 2.0 in the early 2000s (Gandomi & Haider, 2015).
Social networks
The first social media channel to be addressed in this research is the social network, of which Facebook is the most prominent example. On such a website individuals create a personal profile through which they can invite friends, send messages, etc. (Kaplan & Haenlein, 2010). According to Boyd and Ellison (2007), three aspects are considered important for an online platform to be
considered a social network. Firstly, people must be able to create their own profile. Secondly, these individuals are able to connect with other platform users. Lastly, people are able to view and relate to other users and their activities. It has been made clear in literature that social networking websites are of outstanding popularity (Kaplan & Haenlein, 2010), making it an interesting big data source for firms. Because social networks can bring people into contact with others, they are mostly used by firms to enable dialogue and interaction (Konsti-Laakso, 2017). As a result of this interactive part, social networks typically provide network and social data.
Blogs
7
Marques et al., 2013). Blogs are used to update for example customers and employees about relevant developments, on which they can comment (Kaplan & Haenlein, 2010). Blogging can be a necessary tool to discover how people communicate through the web, how they exchange ideas and share knowledge (Marques et al., 2013). Microblogging is a second type of blog, defined as the act of broadcasting short, real-time messages; and it is very popular worldwide (Marques et al., 2013). A microblog can essentially be seen as a smaller version of a blog. The data that stems from company blogs is mostly in the form of text and pictures.
Multimedia communities
Multimedia communities, such as YouTube, are the third category of social media channels that will be dealt with in the present research. According to Smith and colleagues (2012), company blogging is focused on promoting conversation among users, while multimedia communities are more used for self-promoting activities and they seem to have more of a supporting role. They can thus be used by companies to promote and expose new products for instance. However, Kaplan and Haenlein (2010) state that sharing media content between users is the main objective of multimedia communities. Companies can thus use multimedia communities in a self-promoting way by for example informative purposes, but also to encourage customers to share opinions and reviews about the company through adding comments or through uploading their own media. Therefore, the main data that stems from multimedia communities is text, photo, video, and audio data.
Wikis
The final social media channel to be addressed in this research is the wiki. Wikis are defined as websites which allow users to add, remove, and change text-based content (Kaplan & Haenlein, 2010). Because people can also change the content of wikis, the usage of such platforms can create previously unheard-of opportunities for joint content development (Wagner & Majchrzak, 2006). The best-known wiki in the world is Wikipedia, which is of such size and popularity that many people perceive
everything on Wikipedia as true (Kaplan & Haenlein, 2010). Wikis can be used to gather professional knowledge, as it is used a lot by experts, but they can also be used internally to share data and
knowledge within companies. The data generated by wikis is mostly text-based.
Big data analytics
8
and colleagues (2015), there are three V’s that are important in understanding big data analytics, namely volume (large datasets), variety (different types of data from multiple sources), and velocity (data collected in real time). In addition to the three original V’s, there are also other dimensions mentioned in literature (Gandomi & Haider, 2015): veracity (the unreliability of some sources of data), variability and complexity (variation in the data flow rates and the fact that big data is generated through endless sources), and value (data in its original form has little value, but a high value can be obtained by analyzing large volumes of data). According to Gandomi and Haider (2015), there are various techniques that firms can use to analyze big data, such as text analytics/mining (extracting information from textual data), audio analytics (analyzing and extracting information from unstructured audio data), and video analytics/video content analysis (monitoring, analyzing, and extracting meaningful information from video streams).
Chen and colleagues’ (2012) paper is likely the most prominent paper discussing big data analytics, as they divide the phenomenon in three generations: Business Intelligence & Analytics (BI&A) 1.0, BI&A 2.0, and BI&A 3.0. BI&A 1.0, as to Chen and colleagues (2012), is rooted from the database management field. Data is mostly structured, collected by companies through various systems, and often stored in commercial database management systems. It is analyzed using mainly statistical methods. BI&A 1.0 can thus be considered as the collection of big data by the firm itself, and it is stored and analyzed in a structured manner. This type of analytics has been used a lot in the past, but currently it can be seen as outdated, as other sources generating more and better information have now entered the domain.
9
BI&A 3.0, according to Chen and colleagues (2012), is an emerging field, in which firms are using data from mobile and internet-enabled devices from people to gather data about a person’s location, preferences, etc. This field of big data analytics is emerging and has not yet been researched
thoroughly. In the future it is likely that mobile analytics will be used a lot, but as for right now, social media analytics is the largest type. Therefore, the present research will focus on the second type of big data analytics, claiming social media interaction of four channels as a main driver for big data
analytics.
THEORETICAL FOUNDATION
The systems theory is an approach that has been used in scientific research for many years, which focuses on systems as a whole, and not on their individual parts (Ackoff, 1971). According to Ackoff (1971), a system is a set of interrelated elements, with at least two elements and a relation between each element. A system possesses properties that are derived from relationships between the
components of the system (Ackoff, 1971). The most comprising contemporary article about systems theory is the paper of Nevo and Wade (2010), who illustrate that systems are composite things that have interacting components that may be systems in their own right, or they can be basic elements. Systems theory proposes that super-additive value creation emerges from the interactions among several system components (Nevo & Wade, 2010).
The concept of synergies is a central notion to systems theory. The word is often used to label
relationships that result in positive outcomes (Nevo & Wade, 2010). It refers to combined, cooperative effects that arise from the relationships and interactions among various forces, particles, elements, parts, genomes or individuals in a given context – effects that are not otherwise attainable (Corning, 2014). A synergy is therefore a positive emergent capability, stemming from the interaction among system components (Nevo & Wade, 2010). They can emerge, among others, from functional complementarities – a merge of functionally distinct elements (Corning, 1995). The concept of synergies was in the 1980s not widely acknowledged but over time started to get support from more and more scientists (e.g. Wilson & Hölldobler, 2005; Nowak, 2006; Wilson & Wilson, 2008; Clutton-Brock, 2009), making the concept widely acknowledged in the present day.
10
guide, and assist the implementation of the IT asset within the organizational resource (Nevo & Wade, 2010). When these enabling conditions are met, the system should be able to synergistically do something that the parts cannot do separately (Habbershon, Williams, & MacMillan, 2003).
The present research proposes that when interactively using four complementary social media channels, synergies are created that can generate supper-additive value for firms. This study will not look at the synergies generated by IT assets and organizational resources, as was mostly the focus of previous research (e.g. Nevo & Wade, 2010; 2011), but rather to the synergies that exist between technologies, since little seems to be known about how the complementarity of different technologies can lead to value creation for firms (Dong & Yang, 2019). To the author’s knowledge, currently only Dong and Yang (2019) applied a systems perspective to explain synergies that arise from
complementary technologies. The current study will follow this path by exploring the synergies of complementary social media channels.
HYPOTHESES DEVELOPMENT
Because of the user generated content that stems from diverse social media channels, there exists an opportunity for firms to take notice of external parties, such as customers, employees, and investors (e.g. Sebei et al., 2018; Bello-Orgaz et al., 2016; Grover, Chiang, Liang, & Zhang, 2018). This user generated content from social media channels is often highly unstructured (Dong & Yang, 2019), requiring the use of big data analytic techniques to bring to light, in this case the analysis of structured and unstructured data from social media channels (Gandomi & Haider, 2015). Because of this
enormous amount of data, the term social media analytics emerges in literature and is highly influential in business operations (Choi, 2018). To gather big data containing the full picture of customers, preferences, and markets, firms need to use a variety of social media channels. By leveraging the data stemming from for example Facebook messages, tweets, YouTube videos, blogs posts, and wikis, firms can reach external parties who mention something about their companies (Grover et al., 2018).
Direct effects of social media channels
11
better understood once big data analytics are applied, since the data that comes from social networks can indicate trends from large groups of people (Can & Alatas, 2017). It thus seems that social networks are full of potential for analyzing the big data that stems from them. Therefore, I hypothesize:
Hypothesis 1: Social network use has a positive effect on big data analytics.
Other than social networks, it seems that blogs have captured the interest of corporations (Lee, Hwang, & Lee, 2006). Numerous large companies like Microsoft are making use of blogging (Lee et al., 2006). Organizations are using blogs or microblogs for numerous services, like customer involvement, service, and promotion (Lee et al., 2006). Seemingly company blogs are suitable for applying big data analytic techniques like text mining (Liu, Burns, & Hou, 2017; Dong & Yang, 2019). For example, it appears that posts on Twitter could forecast television show ratings more accurately than other online data (Liu, Singh, & Srinivasan, 2016). Especially when companies are accepting comments on their blogs, which around 90% of 2009’s Fortune 500 companies do, firms could engage customers and use their blogs effectively to communicate with external parties, monitor their brands, etc. (Barnes, 2010). It is thus very likely that companies that use blogs apply big data analytics to leverage content from other parties. Therefore, I hypothesize:
Hypothesis 2: Blog use has a positive effect on big data analytics.
Multimedia communities also have potential for the application of big data analytics. As these sites are dominated by images and videos, big data analytic techniques like video analytics can be used to derive insights from others. It is mentioned by Manovich (2011) that computers can be used to quickly explore massive virtual datasets and then select objects for further analysis. That these multimedia communities are full of potential for analyzing data seems no question. For instance, the most prominent multimedia community YouTube has over 1.9 billion logged in users visiting the website each month (Press – YouTube, 2019). Flickr, a picture sharing website, is another example. Flickr received 1.8 million photos per day in 2012, so the billions of pictures on Flicker are packed with information about people and their thoughts (Wu, Zhu, Wu, & Ding, 2013). The main content on these multimedia communities appears to be rather different from the other social media channels we have seen so far, but can be dealt with using big data analytic techniques such as video analytics. Therefore, I hypothesize:
Hypothesis 3: Multimedia community use has a positive effect on big data analytics.
12
a difference in accuracy compared to an established source as the Encyclopedia Britannica. As wikis allow for massive collaboration between parties (Palomo-Duarte, Dodero, Medina-Bulo, Rodríguez-Posada, & Ruiz-Rube, 2014), firms can use such tools to gather insights from other parties and leverage it to create value. It would seem that analytic tools are required to support the assessment of such large amounts of data (Palomo-Duarte et al., 2014). Firms using wikis will thus be more likely to use big data analytics to sort the large amounts of wiki data. This indicates:
Hypothesis 4: Wiki use has a positive effect on big data analytics. Interaction effects of social media channels
The systems theory proposes that once synergies are created between complementary sources, the joint value of using these sources together is greater than the sum of the values of the individual sources (Tanriverdi & Venkatraman, 2005). A system has interacting components that can enhance strategic potential, which in turn can impact firm performance (Nevo & Wade, 2011). As mentioned, Nevo and Wade (2010) applied a systems view to the ‘black box’ leading from IT assets to firm performance. Synergies between IT assets and organizational resources can lead to better achievement of tasks or goals (Nevo & Wade, 2010). The present research will focus not on the synergies between IT assets and organizational resources, but on the synergies that can be created between technological components, social media channels in this case. As said, realizing a synergy depends on two related components: compatibility and integration effort (Nevo & Wade, 2010). Compatibility means that the different system components need to be able to form a relationship (Nevo & Wade, 2011).
Furthermore, the system parts or subsystems must be integrated to become a successful system (Nevo & Wade, 2010; Katz & Kahn, 1978). It thus seems that, in this case, when social media channels are compatible and tightly integrated, synergies can be created that can lead to super-additive value creation.
13
need to use a variety of social media channels. They also shed insight into how the data that social media channels generate is different. Social networks provide network data, blogs and wikis are dominated by texts and images, and multimedia platforms contain audios and videos (Dong & Yang, 2019). Even though blogs and wikis share the type of data that can be analyzed, making them
somewhat similar, they can be seen as different because wiki content can actually be altered by anyone (Wagner & Majchrzak, 2006). That different channels are used for different purposes is also made clear in a meta-analysis of Reuter and Kaufhold (2018) about social media use in crisis situations. In the study’s overview it seems that different social media channels have been used for different
purposes of gathering information. Distinct social media channels also seem to require different forms of big data analytics. As to Dong and Yang (2019), social network analysis can be used to gain insights from social networking websites, text mining and opinion mining can derive insights from blogs and comments, and web analytics can be used to analyze image and video data from multimedia communities. I thus expect the four social media channels to be complementarities. By exploiting such complementarities in a strategic manner, firms will be better able to improve firm performance
(Milgrom & Roberts, 1990). Because I propose that the four social media channels are complements, firms can only create the maximum super-additive value when all channels are used interactively. Only when using all channels in interaction can a company leverage full information from external parties like customer preference, innovative ideas, and upcoming trends. I thus argue that the value creation potential lies in the synergies that stem from combining these four channels, and that it in turn will lead to more value than summing up their individual contributions. Therefore, I hypothesize:
Hypothesis 5: The interaction of all social media channels has a positive effect on big data analytics. Performance impact of big data analytics
14
clear that literature agrees on the added value that big data analytics can give to firms. Therefore, I hypothesize:
Hypothesis 6: Big data analytics has a positive effect on firm performance.
METHODS
Data
For this research, an existing database containing an Italian survey of ICT usage in enterprises was used. According to the Istat (Italian National Institute of Statistics) website (2018), this survey is carried out by Istat on an annual basis since 2001 and it is sent to active enterprises in industrial and services sectors, with at least 10 employees. The survey is part of the European Community of statistics. The aim of the annual survey is to supply users with indicators on information society: internet activities (web site, social media, cloud computing) and connection used (fixes and mobile broadband), e-Business (ERP, CRM systems), e-commerce, etc. Despite some questions about sales proportion and total revenue, survey questions were on a binary scale (1 = yes, 0 = no).
The present research used data from the survey that was conducted in 2016. This survey was sent to 32,834 enterprises, of which 19,089 responded, 58% of the initial sample. Using archival data can be beneficial compared to other methods. The first benefit is the large amount of respondents. Since the survey is carried out by such a large organization, it is a definite added value compared to having to collect own data. The number of responses also increases the likelihood of significant results in this research. Furthermore, the large amount of respondents increases the likelihood of being able to generalize the findings to other enterprises in Italy or perhaps Europe. Also, Istat is a legitimized source that guarantees the quality of the data. Finally, the added value of this survey is that it measures social media usage by four types of channels: social networks, blogs, multimedia communities, and wikis, making it very useful for this research to test the different direct and interaction effects of social media channels.
The respondents are active in either of the following sectors: manufacturing, energy (electricity, gas steam, air conditioning supply, water supply, sewerage, waste management and remediation
15
are usually sold first and then produced and consumed simultaneously (Nie & Kellogg, 2009). I am interested in what more customer participation using social media channels can do for a firm’s ability to analyze and use their ideas, and since the services sector already has easier access to customers because of more face-to-face contact, I chose to focus on the other three sectors, that do not directly meet with their customers, but can interact with them using technological platforms. The present research thus focused on those firms that seem to be more distant from customers. The sample of these three sectors contains 8,585 firms.
Measures
Independent variables
The independent variables that are dealt with in this study comprise social networks, blogs, multimedia communities and wikis. Each independent variable is measured by one question in the survey. This study’s independent variables will thus be measured using one question for each variable, whether a firm uses a certain social media channel or not. These questions are on a binary scale (1 = yes, 0 = no). In order to create the interaction variables, the independent social media variables in question were multiplied.
Dependent variable
Big data analytics will be measured using four binary questions from the survey about whether or not the company engages in a certain category of big data analytics. Initially, there were six survey
questions concerning big data analytics, but because of missing data of over 80% of the respondents in two questions, these were dropped. The dependent variable big data analytics is thus based on four binary questions about whether a firm analyzes big data from the following data sources: smart devices or sensors, geolocation resulting from the use of portable devices, social media, and other sources. A new variable was computed using the four questions about big data analytics, meaning the new variable will range from 0 = not using big data analytics at all, to 4 = using a large variety of big data analytic techniques.
Control variables
16
million, 50 million, and 200 million euros a year, of which classes were made of 1 (lowest class of revenue) to 13 (highest class of revenue).
Analysis strategy
Before starting the analyses, some missing data was dealt with to be able to carry out the analyses. Using the SPSS function ‘Select Cases’, companies of whom there was relevant missing data were removed from the datafile, leaving a final sample of 8,513 companies. After organizing the dataset, a Pearson correlation was performed using SPSS.
Following the correlation analysis, several regression analyses were performed in SPSS. The
regression type that was used was the Ordinary Least Squares (OLS) regression. The main regression analyses are executed in four steps. Step one involves a regression with solely the control variables and the main effects. Secondly, a regression was performed using the controls, the main effects and the two-way interactions. Then, in the third regression, the controls, the main effects, the two-way
interactions, and the three-way interactions are included. In the fourth regression, the four-way interaction variable is added.
Lastly, an OLS regression was carried out to estimate the effect of big data analytics on firm performance. Firm performance was assessed using the survey variable class of revenue. As
mentioned, the revenues respondents reported are 0, 20,000, 100,000, 200,000, 500,000, 1 million, 2 million, 4 million, 5 million, 10 million, 20 million, 50 million, and 200 million euros a year, of which a log transformed variable was created to simplify to coefficients (Ln(revenue + 1)). This log
transformed variable is used as the dependent variable in this model. The big data analytics variable that is used as a dependent variable in the first stage is now used as an independent variable. Furthermore, the second stage model includes, in addition to the beforementioned controls, the individual social media variables as controls.
RESULTS
In the following section, the results of this study will be discussed. Firstly, I will discuss the
descriptive statistics and intercorrelations of the study and mention the noticeable results. Then, I will present and interpret the results of the several regression analyses and check whether they are
consistent with the previously mentioned hypotheses. Finally, the results of the effect of big data analytics on firm performance will be addressed.
Descriptive statistics and intercorrelations
17
social networks. Wikis are the least used social media channel. Only about 4% of the researched enterprises uses such channels. However, of all channels, the wiki is most strongly correlated with big data analytics (r = .197, p < .001).
Table 1. Descriptive statistics and intercorrelations
Variables M SD 1 2 3 4 5 6 1. Use of IT specialists .330 .472 1 2. Class of revenue 8.68 2.61 .569*** 1 3. Social networks .350 .478 .249*** .219*** 1 4. Blogs .100 .304 .249*** .267*** .429*** 1 5. Multimedia communities .190 .391 .338*** .332*** .517*** .515*** 1 6. Wikis .040 .194 .176*** .176*** .188*** .237*** .238*** 1 7. Big data analytics .190 .551 .223*** .235*** .150*** .194*** .177*** .197*** Note: N = 8,513; *** p < .001
It can furthermore be observed that all independent variables about social media usage are positively correlated with big data analytics, which indicates that the chance of using big data analytics increases with an increasing level of using either social networks (r = .150), blogs (r = .194), multimedia communities (r = .177), or wikis (r = .197). The findings are all significant at the .001 level.
Furthermore, the social media channels are all positively and significantly correlated with each other, meaning that if a firm already uses one social media channel, it is more likely to employ other kinds of social media channels as well. The control variables are also positively and significantly correlated with big data analytics, indicating that they can be included as controls in the regression analyses.
Hypotheses Testing
Direct and interaction effects of social media channels
As mentioned, multiple Ordinary Least Squares (OLS) regressions were performed to test this study’s hypotheses, of which the outcomes are granted in table 2. The findings are presented in four models. Model one contains solely the controls and the independent variables. Model two comprises the control variables, the independent variables, and the two-way interactions. The three-way interactions are added to this in the third model. The fourth and final model contains the control variables, the independent variables, the two-way interactions, the three-way interactions and the four-way interaction.
18
of social network usage on big data analytics. Furthermore, in all four regression models, social network usage has a positive and significant (p < .01) effect on big data analytics, confirming hypothesis 1. Hypothesis 2, predicting that blog usage has a positive effect on big data analytics, is moderately supported, since its effect is only fully significant (p < .001) in the first regression model. Though multimedia communities show a positive and significant correlation with big data analytics in the correlation table, none of the coefficients in the regression models is significant, which means there is no sufficing evidence to support hypothesis 3. I do find support for hypothesis 4, since the effect of wikis on big data analytics is positive and fully significant (p < .001) in all four regression models. The regressions also indicate that the direct effect of wikis on big data analytics is the largest.
Table 2. OLS regression results of direct and interaction effects on big data analytics
Variables Model 1 Model 2 Model 3 Model 4
β (SE) β (SE) β (SE) β (SE) Control variables Use of IT specialists .124*** (.015) .125*** (.015) .126*** (.015) .126*** (.015) Sector 2 .149*** (.017) .151*** (.017) .151*** (.017) .151*** (.017) Sector 3 .012 (.014) .012 (.014) .012 (.014) .012 (.014) Class of revenue Main effects .024*** (.003) .025*** (.003) .025*** (.003) .025*** (.003) Social networks .040** (.014) .044** (.016) .046** (.016) .045** (.016) Blogs .159*** (.023) .191 (.098) .156 (.136) .083 (.139) Multimedia communities .024 (.019) -.011 (.038) .001 (.039) -.005 (.039) Wikis .371*** (.030) .315*** (.066) .354*** (.073) .334*** (.073) Two-way interactions
Social networks x blogs -.145 (.099) -.114 (.143) -.034 (.145)
Social networks x multimedia communities
.056 (.043) .038 (.046) .046 (.046)
Social networks x wikis -.036 (.084) -.128 (.101) -.089 (.102)
Blogs x multimedia communities .052 (.050) .094 (.192) .247 (.200) Blogs x wikis .588*** (.074) .780* (.332) 1.386** (.399) Multimedia communities x wikis -.268** (.077) -.489** (.161) -.384* (.166) Three-way interactions Social networks x blogs x multimedia communities
19 Table 2. Continued
Variables Model 1 Model 2 Model 3 Model 4
β (SE) β (SE) β (SE) β (SE) Social networks x blogs x
wikis -.072 (.350) -.832 (.447) Social networks x multimedia communities x wikis .329 (.186) .187 (.193) Blogs x multimedia communities x wikis -.181 (.208) -1.970** (.685) Four-way interaction Social networks x blogs x multimedia communities x wikis 1.970** (.719) R .329 .340 .341 .342 R² .108 .116 .116 .117 Note: N = 8,513; * p < .05 ** p < .01 *** p < .001 Dependent variable: big data analytics
The positive and significant four-way interaction effect (β = 1.970, p < .01) in model 4 confirms hypothesis 5. Interactively using all four social media channels has a positive effect on big data analytics, and this effect is greater than summing up each individual effect. It indicates that making use of all social media channels together generates a greater impact than the sum of the channels in isolation. This finding is consistent with the systems perspective, that indicates that super-additive value can be generated by combining complementary sources.
Besides the four-way interaction, results show that one two-way interaction – the one of blogs and wikis – also has a synergistic effect on big data analytics. Though its effect is not as large as the four-way interaction effect, it is still much larger than summing up the isolated effects of blogs and wikis. An explanation for this effect may be that both blogs and wikis are text-based, making them somewhat related. Firms can apply for instance text mining interactively for both channels. They are however complementary in the sense that wiki content can be altered. But because they are, of all channels, most related, they can likely easily be combined and companies might feel inclined to apply big data analytics when interactively using these two text-based channels.
20
blogs and wikis in interaction produce a positive synergistic effect, it would seem that when multimedia communities are added, this effect twists into a negative one. Since blogs and wikis are somewhat related concerning the data that stems from these channels, it seems that adding a third channel containing different data results in negative outcomes. Only when adding a fourth channel will firms be more inclined again to apply big data analytics. It would seem that super-additive value is only reached when two related channels are being used in interaction or when all channels,
providing a full picture, are being used in interaction. Synergies that create the greatest super-additive value are only generated when all four social media channels are being used interactively.
Performance impact of big data analytics
Concerning the sixth hypothesis of this research, an OLS regression was performed to test the effect of big data analytics on firm performance. As mentioned, the main social media effects are now taken in as controls. The results of the regression can be found in table 3.
Table 3. OLS regression results performance impact of big data analytics
Variables β (SE) Control variables Use of IT specialists .040*** (.009) Sector 2 -.014 (.010) Sector 3 .042*** (.008) Class of revenue .662*** (.002) Social networks .011 (.008) Blogs .061*** (.013) Multimedia communities .014 (.011) Wikis .131*** (.018)
Big data analytics
Big data analytics .013* (.006)
R .985
R² .970
Note: N = 8,513; * p < .05 ** p < .01 *** p < .001 Dependent variable: firm performance
21
DISCUSSION
Theoretical implicationsAs the paradigm of business shifts from intuitive decision making to making data driven decisions (LaValle et al., 2011), enterprises must manage this change through discovering which sources of data can add most value. The present research focused on social media channels as a potential value creation data source. Social media analytics has been a booming field in IS research the last number of years. Analyzing the big data that stems from social media channels has enormous potential for firms trying to make use of information from external parties. The present research investigated the effects of different types and interactions of social media channels on big data analytics. I find that social networks, blogs, and wikis have a positive effect on big data analytics, and that the interactive use of blogs and wikis has a synergistic effect on big data analytics. The main contribution of the present research is that the interactive use of four categories of social media channels generates the largest synergistic effect on big data analytics. I furthermore find that big data analytics has a positive effect on firm performance.
To the author’s knowledge, this study is among the first to draw on systems theory to explain social media usage by firms. By applying this systems perspective to the concept of social media, the present research proposes that when interactively using four social media channels – social networks, blogs, multimedia communities, and wikis – firms can generate the biggest super-additive value. I thus find that synergies emerge from the use of four complementary social media channels, and that big data analytics has potential in creating value through its positive effect on firm performance.
To answer my research questions about the effects of interactions of social media channels on a firm’s analysis of big data and about the effect of big data analytics on firm performance, I find that when interactively using four social media channels, firms can create synergies that lead to the biggest super-additive value. When interactively using these four channels, firms are more likely to apply big data analytics, that can in turn increase the performance of the firm. I also find that firms can create super-additive value from using two channels – blogs and wikis – in interaction, but this effect is not as large as the effect of using all channels in interaction. It thus seems that the greatest value creation potential of social media analytics lies in the interactive use of four categories of social media channels. Only when all four channels are being used interactively, firms are able to uncover the full picture about for example customer ideas and preferences.
22
has been ignored in past research. I find that the interactive use of four social media channels
positively influences a firm’s big data analytics, and that big data analytics positively influences firm performance. By focusing on the synergies between complementary social media channels, the biggest super-additive value emerges when all four complementary social media channels are being used in interaction. This theoretical explanation allows an understanding to be reached about how social media can add value.
Finally, this study makes contributions to systems theory. As currently only Dong and Yang (2019) have focused on synergies that can arise from complementary technologies, I extend their first move by discovering synergies between complementary social media channels. As mentioned, previous research has mostly focused on the synergies that emerge between IT assets and organizational resources (e.g. Nevo & Wade, 2010; 2011; Tanriverdi & Venkatraman, 2005). The current research addresses this phenomenon and contributes to the understanding of the often-ignored explanation of possible synergies between technological components. It therefore extends system theory by indicating that the systems theory also holds for synergies that arise from complementary IT assets.
Practical implications
The present study has important business implications for managers attempting to create value using social media. In this era of huge changes concerning new technologies as sources of big data, advanced ways to analyze this big data and new approaches regarding communication with external parties, it is of important concern for managers to address these changes the right way and invest in those sources that have the highest potential for data analytics. As this study addressed different social media channels as big data sources, it has important implications for managers concerning how social media should be used in order to leverage most potential.
Though the present study finds support for the potential of social networks, blogs, and wikis for analyzing big data, it seems clear that social media channels should not be used in isolation. The current research reveals that using four categories of social media channels interactively has the greatest potential in realizing synergies that can create value. The current findings provide guidance to managers that value can be created through the interactive use of social networks, blogs, multimedia communities, and wikis. When tightly combining and integrating the four compatible channels (Nevo & Wade, 2010), managers can realize super-additive value through an increased utilization of big data analytics. In addition, I find that big data analytics positively impacts firm performance, indicating that applying analytics indeed has value-creation potential for firms that aim to increase their performance.
23
(Waldrop, 2008), it seems from the current results that these can be an excellent big data source. The present research also shows that super-additive value can be reached through interactively using blogs and wikis. Managers could thus also reach super-additive value through the interactive use of two text-based channels.
Limitations and future research
The current research is prone to several limitations. First of all, this study was performed using archival data, which was conducted at one time and in one survey. Furthermore, the data that is dealt with is self-reported. Common method bias could thus be an issue, because respondents scored the independent as well as the dependent variable in the questionnaire, and could be prone to respond as to what they would like to see, and not what is actually the case (Chang, Van Witteloostuijn, & Eden, 2010). This bias is however weakened by the fact that the dependent and independent variables were measured on a binary scale (1= yes, 0 = no), to indicate whether firms use a certain social media channel or not, and a certain big data analytics type or not. Since this is likely quite clear, common method bias should not be such an issue here compared to for example a questionnaire using Likert-scale questions. It is however an interesting pathway for future research to check the robustness of this assumption by avoiding possible common method bias when gathering primary data.
A second limitation of the current study is the possibility of endogeneity issues. This could indicate a third variable problem in which a confounding variable is explaining both the independent and the dependent variable, or possible reverse causality in the data (Sande & Ghosh, 2018). The current study can only claim a significant correlation between four-way social media interaction and big data analytics. An attempt was made in the present study to find an instrumental variable explaining social media usage, but not big data analytics. No such variable has been found within the available data. It is thus an important recommendation for future research to explore the path towards causation, by for instance attempting to find an instrument that explains changes in the independent variable, but not the dependent variable.
Another limitation is a possible generalizability issue when applying the present findings to other sectors. This study chose to focus on three sectors, namely manufacturing, energy, and construction. One must thus exercise caution when generalizing the current findings to for example services settings, since it is acknowledged in academic literature that this sector is very distinct from other sectors (e.g. Chesbrough & Spohrer, 2006; McColgan, 1997; Nie & Kellogg, 2009; Shostak, 1982). A study in a different field could be executed in the future to examine whether the current study’s findings also hold in other industries.
Furthermore, the country the survey was conducted in could create generalizability issues. The survey was only sent to Italian companies. The culture of Italy could have a significant impact on the
24
countries in Asia (Hofstede, 1983; 2001). For example, in Italy there is great individualism (Hofstede, 1983; 2001), and this could make that social media and big data analytics have a greater impact in Italy than in countries with lower individualism. Future research should find out whether the findings of the present research also hold in other cultures or nations.
Another limitation of the study could be that the survey used in this research was only sent to
companies with at least 10 employees. This could cause issues when generalizing the current results to smaller companies. Since these smaller businesses or sole traders are likely to have fewer resources available to invest in technology, it could be questioned whether the current findings also hold for those companies. A study is this field is a necessary step to discover whether the same value-creation potential of social media usage is applicable to smaller firms.
Lastly, the current research was not able to uncover a mediation pathway from four-way social media interaction to firm performance through big data analytics within the present data. A Sobel test was run to discover possible mediation, which was nearly significant, but just not close enough to claim mediation. It is an interesting avenue for future research to further discover this pathway to examine if four-way social media interaction impacts firm performance through big data analytics.
Conclusion
This study is among the first to attempt to apply a systems view to the relationship between social media channels and big data analytics. I find that the interactive use of four complementary social media channels can create synergies that lead to the greatest super-additive value creation through its positive effect on big data analytics. I empirically test the direct and interaction effects of four types of social media channels – social networks, blogs, multimedia communities, and wikis – on big data analytics using a large Italian database with a final sample of 8,513 firms. I find that social networks, blogs and wikis have a positive effect on big data analytics. I also find that the interactive use of blogs and wikis has a synergistic effect on big data analytics, but the interactive effect of combining all channels has the biggest synergistic effect on big data analytics. Finally, I find a positive effect of big data analytics on firm performance.
REFERENCES
Ackoff, R. L. (1971). Toward a system of systems concepts. Management Science, 17(11), 661-671.
Barnes, N. G. (2010). Tweeting and blogging to the top. Marketing Research, 22(1), 8-13.
Bello-Orgaz, G., Jung, J., & Camacho, D. (2016). Social big data: Recent achievements and new
challenges. Information Fusion, 28, 45-59.
25
platforms: The role of meaning-making in social media repertoires. Journal of Computer
Mediated Communication, 23(5), 245-259.
Boyd, D. M., & Crawford, K. (2012). Critical questions for big data: Provocations for a cultural,
technological, and scholarly phenomenon. Information, Communication & Society, 15(5),
662-679.
Boyd, D. M., & Ellison, N. B. (2007). Social network sites: Definition, history, and scholarship.
Journal of Computer‐mediated Communication, 13(1), 210-230.
Bugshan, H. (2015). Open innovation using Web 2.0 technologies. Journal of Enterprise Information
Management, 28(4), 595-607.
Can, U., & Alatas, B. (2017). Big social network data and sustainable economic development.
Sustainability, 9(12), 2027.
Chang, S. J., Van Witteloostuijn, A., & Eden, L. (2010). From the editors: Common method variance
in international business research. Journal of International Business Studies, 41(2), 178-184.
Chen, H., Chiang, R. H., & Storey, V. C. (2012). Business intelligence and analytics: From big data to
big impact. MIS quarterly, 36(4), 1165-1188.
Chesbrough, H., & Crowther, A. (2006). Beyond high tech: Early adopters of open innovation in other
industries. R&D Management, 36(3), 229-236.
Chesbrough, H., & Spohrer, J. (2006). A research manifesto for services science. Communications of
the ACM, 49(7), 35-40.
Choi, T. (2018). Incorporating social media observations and bounded rationality into fashion quick
response supply chains in the big data era. Transportation Research Part E, 114, 386-397.
Clutton-Brock, T. (2009). Cooperation between non-kin in animal societies. Nature, 462(7269), 51-57.
Corning, P. A. (1995). Synergy and self-organization in the evolution of complex systems. Systems
Research, 12(2), 89-121.
26
Systems Research and Behavioral Science, 31(2), 181-196.
Dalla Valle, L., & Kenett, R. (2018). Social media big data integration: A new approach based on
calibration. Expert Systems with Applications, 111, 76-90.
Dong, J. Q., & Wu, W. (2015). Business value of social media technologies: Evidence from online
user innovation communities. Journal of Strategic Information Systems, 24(2), 113-127.
Dong, J. Q., & Yang, C. H. (2019). Business value of big data analytics: A systems-theoretic approach
and empirical test. Information & Management. In press.
Fang, H., Zhang, Z., Wang, C., Daneshmand, M., Wang, C., & Wang, H. (2015). A survey of big data
research. IEEE Network, 29(5), 6-9.
Felt, M. (2016). Social media and the social sciences: How researchers employ big data analytics. Big
Data & Society, 3(1), 1-15.
Gandomi, A., & Haider, M. (2015). Beyond the hype: Big data concepts, methods, and analytics.
International Journal of Information Management, 35(2), 137-144.
Germann, F., Lilien, G. L., Fiedler, L., & Kraus, M. (2014). Do retailers benefit from deploying
customer analytics? Journal of Retailing, 90(4), 587-593.
Ghani, N., Hamid, S., Targio Hashem, I., & Ahmed, E. (2018). Social media big data analytics: A
survey. Computers in Human Behavior.
Giles, J. (2005). Internet encyclopaedias go head to head. Nature, 438(7070), 900-901.
Grover, V., Chiang, R., Liang, T., & Zhang, D. (2018). Creating strategic business value from big data
analytics: A research framework. Journal of Management Information Systems, 35(2),
388-423.
Habbershon, T., Williams, M., & MacMillan, I. (2003). A unified systems perspective of family firm
performance. Journal of Business Venturing, 18(4), 451-465.
27 Management, 1(1), 41-48.
Hofstede, G. (2001). Culture’s consequences: Comparing values, behaviors, institutions, and
organizations across nations. Thousand Oaks, CA: Sage Publications.
Hughes, A. L., St Denis, L. A., Palen, L., & Anderson, K. M. (2014). Online public communications
by police & fire services during the 2012 Hurricane Sandy. In Proceedings of the 32nd Annual
ACM Conference on Human Factors in Computing Systems, 1505-1514. ACM.
Jimenez-Marquez, J. L., Gonzalez-Carrasco, I., Lopez-Cuadrado, J. L., & Ruiz-Mezcua, B. (2019).
Towards a big data framework for analyzing social media content. International Journal of
Information Management, 44, 1-12.
Kaplan, A. M. & Haenlein, M. (2010). Users of the World, Unite! The Challenges and Opportunities
of Social Media. Business Horizons, 53(1), 59-68.
Katz, D., & Kahn, R. L. (1978). The social psychology of organizations, 2, 528. New York: Wiley.
Kircaburun, K., Alhabash, S., Tosuntaş, Ş. B., & Griffiths, M. D. (2018). Uses and gratifications of
problematic social media use among university students: A simultaneous examination of the
Big Five of personality traits, social media platforms, and social media use motives.
International Journal of Mental Health and Addiction, 1-23.
Konsti-Laakso, S. (2017). Stolen snow shovels and good ideas: The search for and generation of local
knowledge in the social media community. Government Information Quarterly, 34(1),
134-139.
LaValle, S., Lesser, E., Shockley, R., Hopkins, M. S., & Kruschwitz, N. (2011). Big data, analytics and the path from insights to value. MIT Sloan Management Review, 52(2), 21-22.
Lee, G., & Cole, R. (2003). From a firm-based to a community-based model of knowledge creation:
The case of the Linux kernel development. Organization Science, 14(6), 633-649.
Lee, S., Hwang, T., & Lee, H. (2006). Corporate blogging strategies of the Fortune 500 companies.
28
Liu, X., Burns, A. C., & Hou, Y. (2017). An investigation of brand-related user-generated content on
Twitter. Journal of Advertising, 46(2), 236-247.
Liu, X., Singh, P. V., & Srinivasan, K. (2016). A structured analysis of unstructured big data by
leveraging cloud computing. Marketing Science, 35(3), 363-388.
Manovich, L. (2011). Trending: The promises and the challenges of big social data. Debates in the
Digital Humanities, 2, 460-475.
Marques, A., Krejci, R., Siqueira, S., Pimentel, M., & Braz, M. (2013). Structuring the discourse on
social networks for learning: Case studies on blogs and microblogs. Computers in Human
Behavior, 29(2), 395-400.
McColgan, E. (1997). How fidelity invests in service professionals. Harvard Business Review, 75(1),
137-144.
McFarland, L., & Ployhart, R. (2015). Social media: A contextual framework to guide research and
practice. The Journal of Applied Psychology, 100(6), 1653-1677.
Milgrom, P., & Roberts, J. (1990). The economics of modern manufacturing: Technology, strategy,
and organization. American Economic Review, 80(3), 511-528.
Mount, M., & Garcia Martinez, M. (2014). Social Media: A tool for open innovation. California
Management Review, 56(4), 124-143.
Nambisan, S. (2002). Designing virtual customer environments for new product development: Toward
a theory. The Academy of Management Review, 27(3), 392-413.
Nevo, S., & Wade, M. R. (2010). The formation and value of IT-enabled resources: Antecedents and
consequences of synergistic relationships. MIS Quarterly, 34(1), 163-183.
Nevo, S., & Wade, M. R. (2011). Firm-level benefits of IT-enabled resources: A conceptual extension
and an empirical assessment. Journal of Strategic Information Systems, 20(4), 403-418.
Nie, W., & Kellogg, D. (2009). How professors of operations management view service operations?
29
Nisar, T. M., Prabhakar, G., & Strakova, L. (2019). Social media information benefits, knowledge
management and smart organizations. Journal of Business Research, 94, 264-272.
Nowak, M. A. (2006). Five rules for the evolution of cooperation. Science, 314(5805), 1560-1563.
Palomo-Duarte, M., Dodero, J. M., Medina-Bulo, I., Rodríguez-Posada, E. J., & Ruiz-Rube, I. (2014).
Assessment of collaborative learning experiences by graphical analysis of wiki contributions.
Interactive Learning Environments, 22(4), 444-466.
Press - YouTube. (2019). Retrieved from https://www.youtube.com/intl/en-GB/yt/about/press.
Reuter, C., & Kaufhold, M. A. (2018). Fifteen years of social media in emergencies: A retrospective
review and future directions for crisis Informatics. Journal of Contingencies and Crisis
Management, 26(1), 41-57.
Rutter, R., Roper, S., & Lettice, F. (2016). Social media interaction, the university brand and
recruitment performance. Journal of Business Research, 69(8), 3096-3104.
Sande, J., & Ghosh, M. (2018). Endogeneity in survey research. International Journal of Research in
Marketing, 35(2), 185-204.
Sandmeier, P., Morrison, P., & Gassmann, O. (2010). Integrating customers in product innovation:
Lessons from industrial development contractors and in-house contractors in rapidly changing
customer markets. Creativity and Innovation Management, 19(2), 89-106.
Sebei, H., Hadj, T., & Ben, A. (2018). Review of social media analytics process and big data pipeline.
Social Network Analysis and Mining, 8(1), 30.
Shoham, Y. (2018). U.S. Patent No. 9,892,416. Washington, DC: U.S. Patent and Trademark Office.
Sigala, M. (2012). Social networks and customer involvement in new service development (NSD)
International Journal of Contemporary Hospitality Management, 24(7), 966-990.
Smith, A., Fischer, E., & Yongjian, C. (2012). How does brand-related user-generated content differ
across Youtube, Facebook, and Twitter? Journal of Interactive Marketing, 26(2), 102-113.
30
topic discovery, data collection, and data preparation. International Journal of Information
Management, 39, 156-168.
Tan, W., Blake, M., Saleh, I., & Dustdar, S. (2013). Social-network-sourced big data analytics. IEEE
Internet Computing, 17(5), 62-69.
Tanriverdi, H., & Venkatraman, N. (2005). Knowledge relatedness and the performance of
multibusiness firms. Strategic Management Journal, 26(2), 97-119.
The corporate blogging revolution (2005). E Learning Age, 9.
Tufekci, Z. (2014). Big questions for social media big data: Representativeness, validity and other methodological pitfalls. In Eighth International AAAI Conference on Weblogs and Social
Media, 505-514.
Wagner, C., & Majchrzak, A. (2006). Enabling customer-centricity using wikis and the wiki way.
Journal of Management Information Systems, 23(3), 17-43. Waldrop, M. (2008). Big data: wikiomics. Nature News, 455(7209), 22-25.
Wamba, S., Gunasekaran, A., Akter, S., Ren, S., Dubey, R., & Childe, S. (2017). Big data analytics
and firm performance: Effects of dynamic capabilities. Journal of Business Research, 70,
356-365.
Warschauer, M., & Grimes, D. (2007). Audience, authorship, and artifact: The emergent semiotics of
Web 2.0. Annual Review of Applied Linguistics, 27(1), 1-23.
Wills, M. J. (2014). Decisions through data: Analytics in healthcare. Journal of Healthcare
Management, 59(4), 254-262.
Wilson, E. O., & Hölldobler, B. (2005). Eusociality: origin and consequences. Proceedings of the
National Academy of Sciences, 102(38), 13367-13371.
Wilson, D. S., & Wilson, E. O. (2008). Evolution "for the good of the group". American
31
Wu, X., Zhu, X., Wu, G. Q., & Ding, W. (2013). Data mining with big data. IEEE Transactions on
