summary report

Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=ucis20

Download by: [University of Waikato Library] Date: 03 July 2017, At: 18:45

Journal of Computer Information Systems

ISSN: 0887-4417 (Print) 2380-2057 (Online) Journal homepage: http://www.tandfonline.com/loi/ucis20

Actionable Social Media Competitive Analytics For Understanding Customer Experiences

Wu He, Xin Tian, Yong Chen & Dazhi Chong

To cite this article: Wu He, Xin Tian, Yong Chen & Dazhi Chong (2016) Actionable Social Media Competitive Analytics For Understanding Customer Experiences, Journal of Computer Information Systems, 56:2, 145-155, DOI: 10.1080/08874417.2016.1117377

To link to this article: http://dx.doi.org/10.1080/08874417.2016.1117377

Published online: 15 Jan 2016.

Submit your article to this journal

Article views: 165

View related articles

View Crossmark data

Citing articles: 3 View citing articles

ACTIONABLE SOCIAL MEDIA COMPETITIVE ANALYTICS FOR UNDERSTANDING CUSTOMER

EXPERIENCES

WU HE Old Dominion University, Norfolk, VA 23529,

USA

XIN TIAN Old Dominion University, Norfolk, VA 23529,

USA

YONG CHEN Old Dominion University, Norfolk, VA 23529,

USA

DAZHI CHONG Old Dominion University, Norfolk, VA 23529,

USA

ABSTRACT

A large amount of user-generated content is now freely available on social media sites. To increase their competitive advantage, companies need to monitor and analyze not only the customer- generated content on their own social media sites, but also the content on their competitors’ social media sites. In this article, we describe a framework to integrate several techniques including quantitative analysis, text mining, and sentiment analysis for analyzing and comparing social media content from business competitors. Specifically, we conducted an in- depth case study which applies our developed framework to the analysis and comparison of social media content on the Facebook sites of the three largest drugstore chains in the United States: Walgreens, CVS, and Rite Aid. We found similarities and differences in the social media use among the three drugstore chains. We discuss the implications of our findings and provide recommendations to help companies develop their social media competitive analysis strategies.

Keywords: Social media, text mining, sentiment analysis, business intelligence, competitive intelligence, competitive analytics

INTRODUCTION

More and more customers use social media sites to express opinions, feelings, and concerns about the services and products they have purchased. Customers’ conversations on social media sites can help us learn about their purchasing behaviors and shop- ping experiences and can provide valuable knowledge to help businesses improve their marketing and customer service. He, Zha, and Li [23] suggest that businesses leverage and analyze a wealth of textual data on social media to reveal hidden knowledge and insights to gain a competitive edge. However, analyzing social media content can be challenging and very time-consuming [47]. The rapid growth of social media content demands the use of automatic social media analytic techniques. Furthermore, since competitive intelligence is an important factor for businesses to use in managing risks and making decisions [52], there is a need for businesses to monitor not only their own social media sites but also their competitors’ social media sites. Successful businesses need to develop the capability to process all available information (e.g., customers’ opinions, product prices from competitors, reviews of services and products), identify what has happened, and predict what may happen later. Studies indicate that businesses that can harness data analytics significantly outperform their peers on the key business metrics of growth, earnings, and performance [65]. With the wide adoption of social media by businesses, the large

amount of customer-generated content on social media sites has become a new source of mining competitive intelligence [59]. In addition to extracting data from social media sites, businesses need to be able to quickly make sense of the data to provide information that is relevant, and actionable. By understanding what people think about their company, products, and services, companies can act quickly and compete more effectively. Furthermore, corporate employees can transform what they have learned from social media data analytics into action to deliver real business results or to guide the development of better products and services [26]. Thus, busi- nesses are increasingly expected to harness this user-generated social media content to create a competitive advantage in order to excel in the business environment [14]. In this article, we devel- oped a framework to integrate several techniques including quan- titative analysis, text mining, and sentiment analysis to analyze and compare social media content from business competitors. In an effort to help businesses understand how to perform actionable social media analytics and how to transform social media content into strategically actionable knowledge, we conducted an in-depth case study which applies our developed framework to analyze and compare unstructured text content on the Facebook sites of the three largest drugstore chains in the United States: Walgreens, CVS, and Rite Aid. We focused on comparing their Facebook posts to understand the issues and problems with customers’ shop- ping experiences.

The remainder of the article is organized as follows: Section 2 provides a review of social media and social media competi- tive analytics. Section 3 proposes a framework for using social media competitive analytics to analyze and compare social media content from competing organizations. Section 4 presents an in-depth case study of the three largest drugstore chains in the United States. Section 5 discusses the findings. Section 6 pro- vides managerial implications and insights. Section 7 notes conclusions and suggestions for future research.

LITERATURE REVIEW

Social Media

Social media generally refers to online communication plat- forms including websites and web applications used for social networking, photo and video sharing, blogging, etc. [17, 51]. Compared to traditional media, social media allows businesses to easily create online communities and utilize a two-way parti- cipatory media model, rather than the traditional one-way broad- cast media model, to communicate with customers [62].

Volume 56 Issue 2, Spring 2016 Journal of Computer Information Systems 145

Common types of social media include collaborative projects (e.g., Wikipedia), blogs and microblogs (e.g., Twitter), content communities (e.g., YouTube), social networking (e.g., Facebook), virtual game worlds, and virtual social worlds [29]. Among various social media networks, Facebook is the most popular in terms of its network traffic and has been widely used by businesses to engage customers and to influence and track consumer beliefs and attitudes [23, 58].

Social media tools present unparalleled opportunities for businesses. Radick [45] found that social media has been “instrumental in promoting consumer awareness and providing access to vast amounts of information, which impact decision- making processes.” As social media can reach a large audience at a low cost [24, 29], they are excellent vehicles for businesses to access markets, to communicate with customers [25], to foster stronger relationships with customers [13, 55], and to facilitate brand communities [29, 31]. On the social media platform, businesses are able to develop new approaches to rapidly dis- seminate information[4, 27], learn customer perceptions of new product offerings and competitive actions, and maximize oppor- tunities to attract and collaborate with loyal customers [20, 38], influence customers’ evaluations and actions [40], and cultivate loyal customers [38].

Social Media Competitive Analytics

In order to achieve competitive advantage, it is important for businesses to constantly collect and analyze information about their competitors’ products, services, and plans [23, 59]. Traditionally, companies mainly collect information about com- petitors from marketing reports, trade journals, newspaper arti- cles, and competitors’ websites. However, most of the information is secondary information whose objectivity can be questionable and the amount of such information is typically limited [59]. Since social media has been used by many busi- nesses to interact with customers, it is necessary for companies to monitor their own and their competitors’ social media sites. As the market competition among big brands is becoming increasingly fierce, it is of great importance for large companies to identify critical situations or moves by their competitors at an early stage so that they are in a position to initiate counteractive marketing measures [28]. This is especially critical for a com- pany when its competitors use a new promotional technique and customers express negative opinions about their products or services through social media. For example, Coombs [9] found that the diffusion of negative opinions poses a reputational and financial threat and can harm the company’s image and future sales volume. On the other hand, Dey, Haque, Khurdiya, and Shroff [14] found that social media not only provide competitors information but also provide direct comparison of customer behaviors with respect to different verticals among competing organizations. Customers often compare several competitive products with similar functions and share their opinions and sentiments on social media. These comparison opinions have large impact on other customers’ purchasing behaviors and thus need to be constantly collected and analyzed by businesses to identify the relative strengths and weaknesses of their products or services, to analyze the potential business risk and threats from competitors, and to further develop corresponding business strategies or tactics [59]. In summary, it is crucial for companies to develop social media competitive analytics skills so that they can get prompt feedback and can generate daily/weekly/monthly analytic reports for quicker strategy changes that allow them to work more effectively to attract and retain customers [19, 22]. However, a recent literature review reveals that there are only a few studies that use social media competitive analytics in busi- ness although a lot of research has been devoted to analyzing the

data presented in social media [16, 44]. He, Zha, and Li [23] applied a social media competitive analytics approach to analyze unstructured text content on the Facebook and Twitter sites of the three largest US pizza chains: Pizza Hut, Domino’s Pizza, and Papa John’s Pizza. Their results reveal the value of social media competitive analysis and the power of text mining as an effective technique to extract business value from social media content. Dey, Haque, Khurdiya, and Shroff [14] used text mining to gather competitive intelligence about competing products and companies. They found that social media data can be used to derive competitive intelligence and to study the correlation of rival brand promotion events on sales data and consumer sentiment.

Many techniques can be used to conduct social media com- petitive analytics. For instance, a combination of traditional statistical analysis, content analysis, text mining, and sentiment analysis techniques can be used to examine the social media content collected from competing organizations’ social media sites in order to gain insights and compare customers’ experi- ences and sentiments.

As an emerging technology, text mining aims to extract meaningful information from unstructured textual data [23, 36]. The main purpose of text mining is to automatically extract knowledge, insights, and useful patterns or trends from a given set of text documents [64]. Text mining techniques have been used to analyze large amounts of social media data. Chen, Vorvoreanu, and Madhavan [8] mined students’ informal con- versations on social media (e.g., Twitter, Facebook) to under- stand students’ educational experiences, including their opinions, feelings, and concerns about the learning process. Corley et al. [10] used text mining to identify trends in posts about the flu that correlate to real-world influenza-like illness patient report data. Key text mining techniques include cluster analysis, categorization, information extraction (text summariza- tion), and link analysis [64]. In particular, cluster analysis is a key application of text mining and includes four main building blocks: feature selection, the clustering algorithm, validation of the results, and interpretation of the results [18]. By dividing a population into clusters which are different from one another (maximal distance between clusters) but whose members are similar (minimal distance within each cluster), cluster analysis can enhance the understandability of datasets and can support effective decision-making [30]. Currently, there are a wide range of tools that can be used for text mining and analytics, such as the IBM SPSS Modeler (formerly Clementine), Leximancer, Clarabridge, and the SAS Enterprise Miner.

Sentiment analysis is the computational detection and study of opinions, sentiments, emotions, and subjectivities in text [34, 35, 42]. As a special application of text mining, sentiment analysis is concerned with the automatic extraction of positive or negative opinions from text [42]. Since texts often contain a mix of positive and negative sentiment, sentiment analysis is often useful to identify the polarity of sentiment in text (positive, negative, or neutral) and even the strength of sentiment expressed [42, 54]. Sentiment analysis relies mainly on machine learning techniques such as support vector machine (SVM) and naive Bayes to clas- sify texts into positive or negative categories [34, 43]. In the business field, sentiment analysis makes it easier for companies to mine customer opinions on products or companies through their reviews or online posts, so it has been widely used in marketing and customer relationship management. Bollen et al. [3] used sentiment analysis to mine a large corpus of Twitter messages to determine the mood of the Twitter population on a given day. They found that the mood of the Twitter population was able to predict the movement of the Dow Jones Industrial Average (DJIA) on the following day with a claimed 87.6% accuracy. Duan, Cao, Yu, and Levy [15] used sentiment analysis to mine 70,103 online

146 Journal of Computer Information Systems Volume 56 Issue 2, Spring 2016

user reviews posted in various online venues from 1999 to 2011 for 86 hotels in the Washington D.C. area. Stieglitz and Dang- Xuan [50] examined whether sentiment occurring in Wall posts on public political Facebook pages had an effect on feedback in terms of the quantity of triggered comments in the political domain. Stieglitz and Dang-Xuan [49] also used a sentiment analysis tool called SentiStrength [54] to analyze two datasets of more than 165,000 tweets and found that emotionally charged Twitter mes- sages tend to be retweeted more often and more quickly, compared to neutral ones.

A FRAMEWORK FOR CONDUCTING SOCIAL MEDIA COMPETITIVE ANALYTICS

Figure 1 illustrates a framework that includes some possible approaches (traditional statistical analysis, content analysis, text mining, sentiment analysis, social network analysis) for con- ducting data-intensive research on social media competitive analytics. The framework considers methodological approaches from a variety of disciplines such as computer science, statistics, and computational linguistics, as well as from social science. A variety of algorithms and methods such as text classification, n-gram, topic modeling, and sentiment analysis [7, 64] can be used to support and implement these approaches. Additional analytics methods and mining algorithms can be added to this framework as technology evolves.

Data from social media sites can be gathered in a number of ways. The most straightforward way is to use web-crawl- ing software to access the web interface of the site. Currently, many social media sites such as Twitter, Facebook, and YouTube offer application programming interfaces (APIs) for data tracking. These APIs allow organizations to create custom applications for more convenient data collection. In contrast, blogs and online forums typically do not provide APIs for data tracking. However, most blogs and online for- ums do offer RSS feeds which can be easily tracked. For

those without an RSS function, manual copying or web- crawling techniques such as HTML parsing can be used to collect data as well, although they may be more time- consuming [50]. As more social media platforms are provid- ing APIs for accessing their content, gathering large amounts of data from social media sources will become easier over time. A few social media data collection and gathering tools or applications such as NVivo’s NCapture, X1 Social Discovery, and IBM Social Media Capture 4 have also been developed in recent years to facilitate data gathering from various social media platforms.

The data gathered from social media sites will then be stored in a back-end repository for subsequent analysis. Since companies can remove information on their social media sites, there is a need to store longitudinal social media data in the repository. The framework can be used to facilitate the con- struction of a social media data repository for managing het- erogeneous longitudinal social media data from competitors. Based on this framework, social media monitoring and analy- tics software systems can be further developed [48] to collect, store, and analyze data by conducting continuous longitudinal monitoring and analysis of all known social media sites from target business competitors. Competitive intelligence reports for each competitor can be generated on a daily, weekly, or monthly basis to provide actionable insights into critical busi- ness functions such as supply chain, customer service, and marketing. For example, by comparing customer comments and sentiments about a particular competing product, a com- pany could make its marketing strategies more precise, agile, and responsive to consumer demands and thereby increase its sales potential. Based on an understanding of the patterns, trends, issues, and problems reflected in social media content, companies can make more informed recommendations and decisions on proper interventions and services that can improve the customer experience and can help achieve better business outcomes for the immediate future [52].

FIGURE 1. A Framework for Conducting Social Media Competitive Analytics

Volume 56 Issue 2, Spring 2016 Journal of Computer Information Systems 147

A CASE STUDY

Research Questions

This case study examines the Facebook sites of three largest drugstore chains in the United States and applies a social media competitive analytics approach to analyze unstructured text con- tent on their Facebook sites. We analyze the content shared on their Facebook sites in terms of topics, categories, and shared sentiment. Specifically, the study attempts to answer the follow- ing two questions:

• What patterns can be found from their Facebook sites? • What are the main differences in their Facebook patterns?

Methodology

Context of the Study

The retail pharmacy industry is highly competitive. Numerous pharmacies and drug stores use social media both to connect with their consumer bases and to attract new customers. In addition to selling prescription and over-the-counter (OTC) medications, many pharmacies and drug stores in the United States also provide other services and sell products including cosmetics, tobacco, consumables, stationery, household items, and the like. Currently, there are approximately 67,000 pharma- cies in the United States. Many of these pharmacies and drug stores aggressively market on the Internet to capture market share. Popular social media applications used by drugstores include Facebook, Twitter, and YouTube. Among them, Facebook is the most popular social media application used by the drugstore chains. Since pharmacies and drugstores have a large social media user base, and so far there is little research that investigates how large drugstore chains are using social media to support their business, we conducted our social media competitive analysis with the three largest drugstore chains in the United States: Walgreens, CVS, and Rite Aid. Currently, by pharmacy count, Walgreens is the largest, CVS is the second largest, and Rite Aid is the third largest drug retailing chain in the United States.

Procedures

To answer the research questions, we followed the proposed framework listed in Figure 1 to conduct our social media com- petitive analysis. First, we collected quantitative data manually from the chains’ individual Facebook sites, such as the number of fans/followers, the number of postings, comments, shares, and likes, and the frequency of posting. Our data analysis mainly focuses on the information posted on their Facebook sites during June 2013. Next, we applied several analytics methods including statistical analysis, text mining, and sentiment analysis to ana- lyze the gathered text messages in order to discover business insights, new knowledge, and patterns and to acquire a deeper understanding of how the three drug chains are using social media in practice. Our study used the posts collected between June 1 and June 30, 2013, as the sample for analytics. All of the posts and comments were saved into Excel Spreadsheets for competitive analysis.

Case study relies on multiple sources of evidence with data needing to converge in a triangulating fashion [63]. The data analysis in this case study was done by using both qualitative and quantitative data. We first used statistical analysis including descriptive analysis and correlation analysis to examine the relationships among the numerical data we collected from the drugstore chains’ social media sites. We then used a text mining tool and a sentiment analysis tool to mine and analyze the

textual content in order to discover categories and insights and to understand any potential patterns or issues. In this case study, we chose to use Leximancer (www.leximancer.com) as the text mining tool. As a popular text mining tool, Leximancer has a very easy-to-use interface; it has been used by many researchers to analyze the content of collections of textual documents [6, 12] (Campbell, Pitt, Parent, & Berthon, 2011; Dann, 2010). For example, Martin and Rice [37] used it to analyze a collection of textual statements gathered from government web pages in order to identify the major issues and areas of concern of computer users and organizations that had borne an exposure to cybercrime.

Sentiment analysis is often used to monitor brand reputation [39] and to help companies understand the perception that cus- tomers have about their products or services. To conduct senti- ment analysis, we used a highly popular program named SentiStrength (http://sentistrength.wlv.ac.uk/) to measure the intensity of positive, negative, and neutral sentiment in each comment from social media users [49, 50]. A group of four researchers reviewed the text mining and sentiment analysis results carefully to identify patterns, to discuss and interpret the findings, and to come up with recommendations and insights. We had several rounds of evaluation and discussion of the generated results to develop and refine our findings.

Findings

Results of Quantitative Analysis on Facebook Content. We collected quantitative data, such as number of fans/followers, number of postings, comments and likes, frequency of posting, posting, and response time, manually from each chain’s indi- vidual Facebook site. In our analysis of the Facebook data gathered for the month of June 2013, we focused on compar- ing the level of engagement on their Facebook sites. Figures 2–4 show the customer engagement levels for Walgreens, CVS, and Rite Aid in June 2013. As Dey, Haque, Khurdiya, and Shroff [14] indicate, social media data can be an indicator of brand popularity. In the three Figures and in Table 1, we note that Walgreens, the largest drugstore chain, typically had a higher level of engagement than its two competitors, in terms of likes, comments, and shares from customers. This result correlates well with the number of likes it had, since Walgreens had many more likes than its competitors. We also identified spikes from Figures 2–4 and examined the posts from the dates of the spikes to see what might have caused them. We found that these spikes were mainly caused by holidays or by promotional sales as well as by special events. For example, since June 16 was Father’s Day in 2013, we found a relatively higher level of engagement during that weekend, which led to a spike in the number of likes and comments.

Table 1 lists the total number of likes, shares, comments, and posts for each drugstore chain received in June 2013.

Figure 5 compares the number of daily posts among the three drugstore chains in June 2013.

We found that Walgreens posted approximately two mes- sages on average per day, Rite Aid posted 1.5 messages on average, and CVS posted about one message per day. Compared to CVS and Rite Aid, Walgreens seemed more active in engaging and interacting with customers. Based on the data shown in Table 1, we conducted a correlation analysis. A finding observed is that the frequency of post activity was positively correlated to number of likes (r = 0.755), shares (r = 0.838), and comments (r = 0.576). This indicates that active posting on Facebook is likely to engage more customers or to attract the attention of more customers.

148 Journal of Computer Information Systems Volume 56 Issue 2, Spring 2016

Table 2 lists the latest number of likes, the number of people talking about it (in the past seven days), number of people who were there (place feature), and the Facebook starting date for each of these drugstore chains as of April 13, 2014.

The results show that the number of people who liked the pages showed a strong correlation with the number of people commenting on Facebook pages in the past seven days (rs = 0.916); the number of people who liked the pages also had strong correlation with the number of people who visited the place(store) (rs = 1); and the number of people commenting on the Facebook pages in the past seven days had a strong correlation with the number of people who visited the place(store) (rs = 1). This indicates that a drugstore that has more fans, and a higher level of engagement on Facebook is likely to attract more potential customers to visit its physical stores.

The results also reveal that the year when a drugstore joined Facebook had a strong relationship with both the number of

0

1000

2000

3000

4000

5000

6000

7000

8000

2 0

1 3

/ 6

/ 1

2 0

1 3

/ 6

/ 3

2 0

1 3

/ 6

/ 5

2 0

1 3

/ 6

/ 7

2 0

1 3

/ 6

/ 9

2 0

1 3

/ 6

/ 1

1

2 0

1 3

/ 6

/ 1

3

2 0

1 3

/ 6

/ 1

5

2 0

1 3

/ 6

/ 1

7

2 0

1 3

/ 6

/ 1

9

2 0

1 3

/ 6

/ 2

1

2 0

1 3

/ 6

/ 2

3

2 0

1 3

/ 6

/ 2

5

2 0

1 3

/ 6

/ 2

7

2 0

1 3

/ 6

/ 2

9

Like

Comments

Share

FIGURE 2. Walgreens’ Customer Engagement Trend in June 2013

0

200

400

600

800

1000

1200

1400

1600

1800

2 0

1 3

/ 6

/ 1

2 0

1 3

/ 6

/ 3

2 0

1 3

/ 6

/ 5

2 0

1 3

/ 6

/ 7

2 0

1 3

/ 6

/ 9

2 0

1 3

/ 6

/ 1

1

2 0

1 3

/ 6

/ 1

3

2 0

1 3

/ 6

/ 1

5

2 0

1 3

/ 6

/ 1

7

2 0

1 3

/ 6

/ 1

9

2 0

1 3

/ 6

/ 2

1

2 0

1 3

/ 6

/ 2

3

2 0

1 3

/ 6

/ 2

5

2 0

1 3

/ 6

/ 2

7

2 0

1 3

/ 6

/ 2

9

Like

Comments

Share

FIGURE 3. CVS’ Customer Engagement Trend in June 2013

0

200

400

600

800

1000

1200

2 0

1 3

/ 6

/ 2

2 0

1 3

/ 6

/ 4

2 0

1 3

/ 6

/ 6

2 0

1 3

/ 6

/ 8

2 0

1 3

/ 6

/ 1

0

2 0

1 3

/ 6

/ 1

2

2 0

1 3

/ 6

/ 1

4

2 0

1 3

/ 6

/ 1

6

2 0

1 3

/ 6

/ 1

8

2 0

1 3

/ 6

/ 2

0

2 0

1 3

/ 6

/ 2

2

2 0

1 3

/ 6

/ 2

4

2 0

1 3

/ 6

/ 2

6

2 0

1 3

/ 6

/ 2

8

2 0

1 3

/ 6

/ 3

0

Like

Comments

Share

FIGURE 4. Rite Aid’s Customer Engagement Trend in June 2013

TABLE 1. Number of Likes, Shares, Comments, and Posts for Each Drugstore Chain Received in June 2013

Drugstore Like Share Comments Post

Walgreens 60,976 10,895 3642 64 CVS 12,039 994 1732 28 Rite Aid 9270 1807 879 50

0

1

2

3

4

5

6

7

2 0

1 3

/ 6

/ 1

2 0

1 3

/ 6

/ 3

2 0

1 3

/ 6

/ 5

2 0

1 3

/ 6

/ 7

2 0

1 3

/ 6

/ 9

2 0

1 3

/ 6

/ 1

1

2 0

1 3

/ 6

/ 1

3

2 0

1 3

/ 6

/ 1

5

2 0

1 3

/ 6

/ 1

7

2 0

1 3

/ 6

/ 1

9

2 0

1 3

/ 6

/ 2

1

2 0

1 3

/ 6

/ 2

3

2 0

1 3

/ 6

/ 2

5

2 0

1 3

/ 6

/ 2

7

2 0

1 3

/ 6

/ 2

9

CVS

Walgreens

Rite Aid

FIGURE 5. A Comparison of Daily Posts Among the Three Drugstore Chains in June 2013

TABLE 2. Total Number of Likes, Number of People Talking about it, Number of People Who Were there, and the Facebook Starting Date for These Drugstore Chains as of April 13, 2014

Drugstore Likes

Number of people

talking about it (in the past seven days)

Number of people

who were there (place feature)

Facebook starting Date

Walgreens 4,265,248 37,981 991,654 August 19, 2009

CVS 1,323,578 15,065 115,732 June 12, 2009

Rite Aid 677,362 20,818 N/A April 4, 2010

Volume 56 Issue 2, Spring 2016 Journal of Computer Information Systems 149

likes (rs = 0.639) and the number of people commenting on Facebook pages in the past seven days (rs = 0.276). In other words, a drugstore with an earlier Facebook setup tends to attract more Facebook users who “like” the drugstore and who comment on its Facebook pages than those with a later Facebook setup. This indicates that those drugstores that joined Facebook earlier tend to have more Facebook fans, which typi- cally leads to more comments on their Facebook pages.

We manually analyzed the original Facebook posts made by each drugstore chain. Table 3 lists the distribution of Facebook post categories among Rite Aid, CVS, and Walgreens. We found that the three drugstore chains mainly use Facebook to survey customers for learning their interests or opinions on certain events, to announce new services, sales or special offers, or to providing tips, etc. The focuses of the Facebook posts are different among Rite Aid, CVS, and Walgreens. Customer survey contributes to 57.1% of CVS posts and 54% of Rite Aid posts in Rite Aid. Compared with the other two drugstores, Walgreens posted more special offers, which accounted for 40.6% of all of its posts. Unlike CVS, Walgreens and Rite Aid use a variety of approaches to promote their products and services. For instance, advertisements and tips account for 34.3% of all of Walgreens’ Facebook posts.

Content-wise, we found that the funny pictures with animals were the most popular posts among all posts on the drugstore chain’s Facebook sites and typically led to a high level of engage- ment. Many people left comments and liked/shared such posts. In addition, the open-ended survey questions seemed an effective way to engage customers and to attract more people to leave comments/to answer the questions. Some of the survey questions had over 150 comments. Other posts, such as discounts, new

services, or information about free items, are relatively less pop- ular and do not attract many comments or likes (in most cases).

Results of Text Mining. After applying the Leximancer soft- ware to conduct an automatic analysis of the Facebook comments (6253 comments in total), we examined the generated maps care- fully. Leximancer produced several concept maps that contain the extracted concepts and their interrelationships. Each concept map contains the names of the main concepts that occur within the text. Figure 6 shows an example of the concept maps generated for CVS, Rite Aid, and Walgreens, respectively. The map uses large circles to represent key themes, uses dots to represent concepts, and uses brighter color and larger theme circles and concept dots to indicate greater importance within the text [6, 33]. The themes in Leximancer are heat-mapped to indicate importance. According to a predefined color scheme, Leximancer displays the most important theme in red, and the second important in orange, and so on. In general, concepts that are strongly semantically linked will be close to each other and will form clusters [37, 46]. Figure 7 shows some of the key words identified by Leximancer include pharmacy, experience, service, member, love, care, discount, coupon, etc.

Since we are mainly interested in the overall patterns of the customer comments, we further combined the comments of the three drugstores and did a text mining operation. We reviewed the generated concept terms and themes carefully and, since the soft- ware also allowed us to go deeper and look at the original posts related to those concept terms for more details, we held an in-depth discussion on clusters. We manually examined the original posts, related the concept terms in the dataset, and then discussed what clusters they should be placed into based on the meaning of each

TABLE 3. Facebook Post Categories Among Rite Aid, CVS, and Walgreens

CVS Walgreens Rite Aid

Post categories Number of posts Post categories Number of posts Post categories Number of posts

Customer Survey 16 Customer Survey 7 Customer Survey 27 New Service 2 New Service 2 New Service 2 Special Offer 9 Special Offer 26 Special Offer 9 Others 1 Advertisement 13 Advertisement 2

Tips 9 Tips 3 Others 7 Store-related news 3

Others 4 Total 28 Total 64 Total 50

Concept map generated from

CVS data

Concept map generated from

Rite Aid data

Concept map generated from

Walgreens data

FIGURE 6. Generated Concept Maps

150 Journal of Computer Information Systems Volume 56 Issue 2, Spring 2016

FIGURE 7. Concept Map Generated from All Customer Comments

TABLE 4. Categories Related to Customer Comments

Categories Examples (Quotes from Facebook Users)

Suggestions for products or services ● We were at the pharmacy today & don’t u think a hour wait with kids is a bit too much? . . ... U guys gotta hire more people.

● From the last time you guys updated the android app in March, the coupon feature stopped working. Can you guys please fix that?

Customer’s questions for other Facebook customers

● How much did u pay for tax? ● Has anyone tried these? Do they work? ● Do you take both, or just pick which applies to you more?

Customers’ compliments ● I just wanted to say my CVS is great! ● CVS is the best for coupons.

Customers’ complaints ● I have had the worst experiences with XX pharmacy. Will never use them again. ● I received an email with coupon links that don’t work, will you please look into this ● I was told it would take 45 minutes to fill the prescription its a long time.

Customers’ responses or suggestions to other customers

● You CAN add coupon to your card when you log on to CVS.com. When you place your card info your allowed to put all of the coupons on your card which will come off when you make your purchase.

● I put my cvs coupons in my purse so I don’t forget to use them. ● To best protect your skin from the sun, you should choose a sunscreen that has an SPF of 15

or higher. Drugstores’ Facebook response to customer comments

● Rite Aid: We apologize if you’ve experienced any issues when attempting to visit the link provided in our post. This technical issue has since been resolved.

● Rite Aid: We do not tolerate the harassment of individuals, so any comments that disparage an individual by name or a specific entry will be removed from the page.

● Walgreens: I am sorry to hear this. I will certainly follow up on your issue to get you a resolution as soon as possible.

● Walgreens: I’m so sorry that happened and I completely understand your concerns. Please Private Message me so that I can get more details to address your concerns.

● Walgreens: Please email us directly at [email protected].

Volume 56 Issue 2, Spring 2016 Journal of Computer Information Systems 151

post. As a result of the repetitive comparison, contrast, and discus- sion, we finalized a few major clusters and collectively decided a descriptive name for each of the clusters. We went through several iterations to finalize the clusters/concept terms and their labels.

In summary, based on the text mining results, we classified the 6253 comments posted by the drugstore chains’ customers on Facebook during June 2013, as well as the responses to customer comments from the drugstores, into five major cate- gories (see Table 4).

Results of Sentiment Analysis. Table 5–7 list the sentiment analysis results we derived by using SentiStrength for the Facebook comments of Walgreens, CVS, and Rite Aid, respec- tively. By comparing consumer sentiments based on the custo- mer comments we collected from the three drugstore chains’ Facebook sites, we found that Rite Aid received a relatively higher percentage of positive comments and a lower percentage of negative comments from their customers than did CVS and Walgreens.

The comments of customers reflect the customers’ personal opinions, perceptions, and preferences. Different customers often have different experiences regarding the same type of services due to their different experiences in different stores. For example, some customers complained that they waited too long at the pharmacy. Other customers reported that they love their pharmacy because the service that they received was fast, they found good deals and had great savings, the pharmacy staff was friendly to them, and so on.

Below are some quotes of positive comments:

● They call we when my pills are ready and when I ask them a question they answer it and they are friendly.

● I am so ecstatic about the rewards program at rite aid, i save so much money! I should have done it a long time ago. And I must say the service I receive at my Rite Aid is beyond exceptional.

● Dr XX is very good and reasonable. ● Just want to say how much I love my rite aid. My

pharmacist is awesome!!! Below are some quotes of negative comments: ● The pharmacy department is inept and rude. ● One of your pharmacies gave me the wrong medication ● The pharmacy in XX is horrible with the wait there

DISCUSSION

Social media allow firms to engage consumers at a rela- tively low cost and with higher levels of efficiency [29]. Businesses can develop relationships with customers through repeated interactions with them on Facebook [56]. However, an effective use of social media for business purposes is not always easy. Businesses may wonder how often they should post on Facebook. Too many Facebook posts may annoy customers and may drive customers away; too few postings may lead to a low level of engagement on Facebook. A commonly recommended practice for businesses is to have one-to-two posts on Facebook per day, which is suggested to be more effective than fewer or more posts per day [1]. Our case study shows that the three drugstores are following this recommended practice and usually post one to two messages per day. Based on the number of likes, shares, and comments they received during June 2013, we can see that their social media practices were fairly effective in terms of engaging their customer bases.

Many businesses also wonder what kind of a message will attract and engage customers on social media. In our case study, we found that the content of the post plays a critical role in terms of engaging customers. We observed that those drugstore chains that offered visually appealing posts with pictures and survey questions got more engagement and attracted more com- ments, likes, and shares. For example, the open-ended survey questions received more comments than posts related to sales or to promotional announcements. In terms of designing the open- ended survey questions, we found that the three drugstores tended to ask questions that reflected current popular events like Father’s Day, summer vacation, etc. These posts could allow the customer to feel that the drugstore cared about its customers’ interests on a personal level and was willing to take the time to learn about and communicate with its customers. As a result, these posts could bring those customers closer to the business and increase their loyalty.

Instead of passively receiving information sent by busi- nesses [21], customers can also use social media for their own benefit [53]. The comments posted by customers reflect their opinions, perceptions, and preferences. These comments, at first, were responses to the original posts from the drug- stores. Subsequently, some customers had further discussions and exchanged opinions and experiences among them. They also often compared the drugstore with other drugstores in their comments and offered tips to other customers. We also noticed that a few customers (ranging from 4.9% to 12.9%) made negative comments and aired complaints about their experience with the services they had been given by the drugstores. O’Brien and Marakas [41] indicate that 70% of complaining customers will do business with the company again if their problems and concerns are quickly addressed. We noted in our study that when negative comments were posted, each of the three drugstore chains made a good effort to minimize their impact and posted their responses or solutions (e.g., giving

TABLE 5. Sentiment Analysis Results for Walgreens

Sentiment

Occurrence Percentage Percentage

s (all) (all) (sentiment only)

Positive 333 32.80 75.70 Negative 107 10.5 24.30 Neutral 574 56.70 TOTAL 1014 100 100

TABLE 6. Sentiment Analysis Results for CVS

Sentiment

Occurrence Percentage Percentage

s (all) (all) (sentiment only)

Positive 167 30.30 70.20 Negative 71 12.90 29.80 Neutral 313 56.80 TOTAL 551 100 100

TABLE 7. Sentiment Analysis Results for Rite Aid

Sentiment

Occurrence Percentage Percentage

s (all) (all) (sentiment only)

Positive 229 34.80 87.70 Negative 32 4.90 12.30 Neutral 396 60.30 TOTAL 657 100 100

152 Journal of Computer Information Systems Volume 56 Issue 2, Spring 2016

email or phone contact methods). Rite Aid and Walgreens were very good at this and typically responded fairly quickly.

On the other hand, many customers use their comments to provide suggestions to the drugstores or to express their opi- nions. Instead of discussing the posts, some customers expand topics to other fields. For instance, due to a controversy about the use of a racial epithet, Walgreens announced that it was going to cut ties with Paula Deen on June 28, 2013. This announcement quickly resulted in a negative reaction from customers on Facebook. We found quite a few negative com- ments made on June 29, 2013. In particular, 16 out of 18 negative comments were related to this announcement. Since negative word-of-mouth initiated by users is generally more persuasive than most positive comments and can trigger bad consequences [2], we recommend that businesses use proactive methods to remediate the impact of negative comments on their businesses. In particular, businesses need to monitor their social media sites very closely, as social media are being increasingly used by their customers. Furthermore, since cus- tomers often compare services or products in the competitors’ social media sites, monitoring competitors’ social media sites is also crucial. By monitoring and analyzing the words posted by their competitors’ customers, businesses can develop better business intelligence and can use the knowledge they gain to identify relative strengths and problems when comparing their products and services with those of their competitors. However, the social media monitoring and analyzing process can be very challenging because of the huge amount of data created daily on social media.

In summary, the case study shows how the social media analytics results from three drug store chains can be used to benefit their business relationship with their customers. In parti- cular, social media analytics help a business to know more about both its customers and its competitors’ customers. Therefore, use of social media analytics can help a business develop a more effective competitive benchmarking strategy in terms of compar- ing customers’ sentiments about the services and products of a particular company with sentiments about the same or similar services and products offered by competitors. For example, the comparison may identify some additional benefits or values that were previously overlooked. The comparison may also identify some consumer needs that have not been met by competitors [57] or may help to identify potential risks as early as possible. Based on the results of such a comparison, businesses can better under- stand their strengths and weaknesses, can plan appropriate strate- gies, and can create action plans to compete more effectively in the marketplace [61].

IMPLICATIONS

The results of this study indicate that using and mining social media data has value and can help businesses produce useful intelligence. With the growing importance of social media in our society, companies and customers are increasingly using social media as a medium for communication. To help businesses more effectively manage social media data, we recommend that interested businesses adopt or adapt our proposed framework for conducting social media competitive analytics. Below are some suggestions.

First, companies need to consider implementing a corporate database system to capture, store, and manage the messages posted on their social media sites and their main competitors’ social media sites, because implementation of competitive intel- ligence requires the collection of information from both the internal environment and the external environment. In practice, many companies often fail to “scan the external environment to which the companies are exposed, and fail to yield meaningful intelligence” [60]. Since scanning the whole environment is

costly, companies have to decide from what channels they will collect information. Since social media data are public and can be easily accessed online, we believe that social media data, as a valuable information source, should not be overlooked by those companies who want to excel in the marketplace. Since postings and comments on social media can be deleted by companies, using a database to store longitudinal social media data is also important. The longitudinal social media data stored in the database can be used as supporting documentation or as evi- dence for future decision-making, marketing campaign develop- ment, or policy making.

Second, companies need to have a professional staff in place to manage the social media data they collect, to integrate the scattered social media records from different social media sites, and to analyze the data in an efficient way. Existing business intelligence tools, including data mining, text mining, sentiment analysis, and statistical tools, should be used to expedite the data analysis. There are a variety of ways to analyze the data, as this case study shows. For example, the social media content they collect can be classified by type of request, complaints, and suggestion [62]. Social media content can also be classified by sentiment including positive sentiment and negative sentiment. Fisher and Miller [19] recommend that companies “obtain real time snapshots of sentiment related to service, reaction to mes- saging at different locations, for different demographics, from different media outlets, and in a meaningful format.”

Third, companies need to provide managers, executives, and other decision-makers with easy access to the data and the results of the analytics. To gain a competitive advantage, it is imperative for companies to use the data and analytics results to quickly respond to changes. Query-and-reporting tools can be provided to decision-makers who can then explore the data on their own. Online analytical processing (OLAP) tools can be provided to decision-makers to support them in their use of common operations including slice and dice, drill down, roll up, and pivot. Daily, weekly, or monthly reports can be sent to decision-makers who can then turn the information into knowl- edge for decision-making and other actions.

Finally, we need to point out that businesses need to pay specific attention to the quality of the data when they use social media data for decision-making. There are some concerns about the quality and reliability of the data posted on social media. To mediate such concerns, businesses need to integrate social media data with other data such as their internal business data in order to make well-informed decisions. Teo and Choo [52] found that the quality of competitive intelligence information is positively related to organizational impact. By integrating longitudinal social media data with other data sources, companies can more reliably determine trends in a way that provides more accurate predictions that can be incorporated into decision-making and policy development.

There are several limitations to this study. First, the Facebook data were collected for only one month and thus the conclusions should not be overgeneralized. Ideally, a longitudi- nal study is needed to examine the Facebook data for an exten- sive period of time to identify long-term trends of how businesses like drugstores use social media for their business development. Second, we did not interview drugstore managers, employees, and customers directly to understand their social media experience. For future research, it would be interesting to interview them and to get their direct input on their social media experience.

CONCLUSION AND FUTURE RESEARCH

Business cannot afford to ignore social media data as the competition in the marketplace intensifies. Businesses often

Volume 56 Issue 2, Spring 2016 Journal of Computer Information Systems 153

have different perspectives on the use of social media tools. Some businesses tend to use social media to broadcast sales information, and some businesses tend to build relationships with customers through social media. Our case study reveals that the three largest drugstore chains in the United States have used Facebook not only to provide information but also to build relationships with their customers. By monitoring competitors’ social media sites, businesses can learn from each other ways to enhance their social media usage and can detect new commer- cial trends in developing marketing strategies and tactics [5].

Since many companies are not familiar with social media analytics and are especially new to conducting competitive intelligence using social media content [11], the authors propose a novel actionable framework for conducting social media com- petitive analytics, with which they conducted a case study to illustrate how social media data can be transformed into knowl- edge to guide businesses’ decision and action plans. This case study makes a contribution by using several techniques to per- form social media competitive analytics to review the user- generated data on the Facebook sites of the three largest US drugstore chains. The results of the case study show that the proposed framework and the methods applied in the case study are effective ways to perform social media competitive analyses. Businesses should take necessary steps to leverage social media data to complement their competitive intelligence strategies for achieving better competitive advantage.

Our future research plan is to build a social media competitive intelligence monitoring and analytics system, which can be custo- mized to collect different types of data from social media. The purpose of this system is to help companies identify relevant social media content, including issues, complaints, and suggestions that, with minimal effort in collection, may influence business decisions or strategies. The social media data will be stored locally in a large-scale database repository for generating weekly or monthly competitive analysis reports for interested businesses. Businesses can also combine the gathered social media data with other struc- tured data such as sales data and stock prices to determine how social media and competition influence their businesses.

REFERENCES

[1] Amber. 2014. Why You Should Consider Posting Daily; [cited 2014 May 24]. Available from: http://thesocialbusi- ness.com/blog/why-you-should-consider-posting-daily/

[2] Bhattacherjee A. 2001. An empirical analysis of the ante- cedents of electronic commerce service continuance. Decis Support Syst. 32:201–214.

[3] Bollen J, Mao H, Zeng X. 2011. Twitter mood predicts the stock market. J Comput Sci. 2:1–8.

[4] Brown SFX, Kozincts RV, Sherry JF Jr. 2003. Teaching old brand, new tricks: Rctro branding and the revival of brand meaning. J Marketing. 67:19–33.

[5] Cachia R, Compañó R, Da Costa O. 2007. Grasping the potential of online social networks for foresight. Technol Forecasting Soc Change. 74:1179–1203.

[6] Campbell C, Pitt LF, Parent M, Berthon PR. 2011. Understanding consumer conversations around ads in a Web 2.0 world. J Advertising. 40:87–102.

[7] Cavnar WB, Trenkle JM. 1994. N-gram-based text cate- gorization. Ann Arbor MI. 48113:161–175.

[8] Chen X, Vorvoreanu M, Madhavan K. 2014. Mining social media data for understanding students’ learning experiences. IEEE Trans Learn Technol. DOI:10.1109/ TLT.2013.2296520

[9] Coombs WT. 2007. Protecting organization reputations during a crisis: the development and application of situa- tional crisis communication theory. Corporate Reputation Rev. 10:163–176.

[10] Corley CD, Cook DJ, Mikler AR, Singh KP. 2010. Text and structural data mining of influenza mentions in web and social media. Int J Environ Res Public Health. 7:596– 615.

[11] Dai Y, Kakkonen T, Sutinen E. 2011. MinEDec: a deci- sion-support model that combines text-mining technolo- gies with two competitive intelligence analysis methods. Int J Comput Inf Syst Ind Manage Appl. 3:165–173.

[12] Dann S. 2008. A Leximancer analysis of social marketing definitions versus social marketing literature. Proceedings of 2008 Australian and New Zealand Marketing Academy Conference; Sydney, Australia; p. 1–3.

[13] De Vries L, Gensler S, Leeflang PS. 2012. Popularity of brand posts on brand fan pages: an investigation of the effects of social media marketing. J Interact Marketing. 26:83–91.

[14] Dey L, Haque SM, Khurdiya A, Shroff G. 2011. Acquiring competitive intelligence from social media. Proceedings of the 2011 Joint Workshop on Multilingual OCR and Analytics for Noisy Unstructured Text Data; p. 3. Beijing, China.

[15] Duan W, Cao Q, Yu Y, Levy S. 2013. Mining online user- generated content: using sentiment analysis technique to study hotel service quality. Proceedings of the 46th Hawaii International Conference on System Sciences; p. 3119– 3128, mavi, Hawail.

[16] Earle PS, Bowden DC, Guy M. 2011. Twitter earthquake detection: earthquake monitoring in a social world. Ann Geophys. 54:708–715.

[17] Ellison NB. 2007. Social network sites: definition, history, and scholarship. J Comput Mediated Commun. 13:210– 230.

[18] Fayyad MU, Piatesky-Shapiro G, Smuth P, Uthurusamy R. 1996. Advances in knowledge discovery and data mining. AAAI/MIT Press.

[19] Fisher B, Miller H. 2011. Social Media Analytics; [cited 2014 Sep 21]. Available from: http://www.microtech.net/ sites/default/files/socialmediaanalytics.pdf.

[20] Franke N, Shah S. 2003. How communities support inno- vative activities: an exploration of assistance and sharing among end-users. Res Policy. 32:157–178.

[21] Hanna R, Rohm A, Crittenden VL. 2011. We’re all con- nected: the power of the social media ecosystem. Bus Horiz. 54:265–273.

[22] He W, Yan G. 2014. Mining blogs and forums to under- stand the use of social media in customer co-creation. Comput J. DOI:10.1093/comjnl/bxu038.

[23] He W, Zha S, Li L. 2013. Social media competitive ana- lysis and text mining: a case study in the pizza industry. Int J Inf Manage. 33:464–472.

[24] He W, Zha SH. 2014. Insights into the adoption of social media Mashups. Internet Res. 24:160–180.

[25] Holzner S. 2008. Facebook marketing: leverage social media to grow your business. Pearson Education.

[26] IBM. 2013. Business Analytics: Social Business Analytics – Gaining Business Value from Social Media; [cited 2014 Apr 15]. Available from: http://www-01.ibm.com/com- mon/ssi/cgi-bin/ssialias? infotype=SA&subtype=WH&htmlfid=YTW03279USEN

[27] Jin XL, Cheung CM, Lee MK, Chen HP. 2009. How to keep members using the information in a computer-

154 Journal of Computer Information Systems Volume 56 Issue 2, Spring 2016

supported social network. Comput Hum Behav. 25:1172– 1181.

[28] Kaiser C, Schlick S, Bodendorf F. 2011. Warning system for online market research–identifying critical situations in online opinion formation. Knowl-Based Syst. 24:824–836.

[29] Kaplan AM, Haenlein M. 2009. Consumers, companies, and virtual social worlds: a qualitative analysis of second life. Adv Consum Res. 36:873—874.

[30] Kim Y. 2007. Weighted order-dependent clustering and visualization of web navigation patterns. Decis Support Syst. 43:1630–1645.

[31] Laroche M, Habibi MR., Richard MO, Sankaranarayanan R. 2012. The effects of social media based brand commu- nities on brand community markers, value creation prac- tices, brand trust and brand loyalty. Comput Hum Behav. 28:1755–1767.

[32] Leximancer. 2011. Leximancer Manual; [cited 2014 Sep 11]. Available from: https://www.leximancer.com/site- media/lm/science/Leximancer_Manual_Version_4_0.pdf

[33] Li N, Wu DD. 2010. Using text mining and sentiment analysis for online forums hotspot detection and forecast. Decis Support Syst. 48:354–368.

[34] Liu B. 2010. Sentiment analysis and subjectivity. In: Indurkhya N, Damerau FJ, editors. Handbook of natural language processing. Boca: Taylor and Francis Group.

[35] Liu B, Cao SG, He W. 2011. Distributed data mining for E-business. Inf Technol Manage. 12:1–13.

[36] Martin NJ, Rice JL. 2007. Profiling enterprise risks in large computer companies using the Leximancer software tool. Risk Manage. 9:188–206.

[37] McAlexander JH, Schouten JW, Koenig HF. 2002. Building brand community. J Marketing. 66:38–54.

[38] Mostafa MM. 2013. More than words: social networks’ text mining for consumer brand sentiments. Expert Syst Appl. 40:4241–4251.

[39] Muniz AM Jr, Schau HJ. 2005. Religiosity in the aban- doned Apple Newton brand community. J Consum Res. 31:737–747.

[40] O’Brien JA, Marakas GM. 2008. Management information systems. Boston: McGraw-Hill/Irwin.

[41] Pang B, Lee L. 2004. A sentimental education: sentiment analysis using subjectivity summarization based on mini- mum cuts. Proceedings of the 42nd Annual Meeting on Association for Computational Linguistics; Barcelona, Spain; p. 271–278.

[42] Pang B, Lee L, Vaithyanathan S. 2002. Thumbs up?: sentiment classification using machine learning techniques. Proceedings of the ACL-02 Conference on Empirical Methods in Natural Language Processing-Volume 10, Association for Computational Linguistics; philadelphia, PA p. 79–86.

[43] Prochaska JJ, Pechmann C, Kim R, Leonhardt JM. 2012. Twitter= quitter? An analysis of Twitter quit smoking social networks. Tob Control 21:447–449.

[44] Radick S. 2010. The Getting Started with Government 2.0 Guide; [cited 2011 Jul 19]. Available from: http://stever- adick.com/2010/04/02/the-%E2%80%9Cgetting-started- with-government-2-0%E2%80%9D-guide/

[45] Rooney D. 2005. Knowledge, economy, technology and society: the politics of discourse. Telematics Inf. 22:405–422.

[46] Rost M, Barkhuus L, Cramer H, Brown B. 2013. Representation and communication: challenges in inter- preting large social media datasets. Proceedings of the 2013 Conference on Computer Supported Cooperative Work; p. 357–362. San Antonio, TX, USA, Februayr 23– 27, 2013.

[47] Semenov A. 2013. Principles of Social Media Monitoring and Analysis Software; [cited 2013 Sep 11]. Available from: https://jyx.jyu.fi/dspace/bitstream/handle/123456789/41559/ 978-951-39-5225-9.pdf

[48] Stieglitz S, Dang-Xuan L. 2013. Emotions and information diffusion in social media—sentiment of microblogs and sharing behavior. J Manage Inf Syst. 29:217–248.

[49] Stieglitz S, Dang-Xuan L. 2013. Social media and political communication: a social media analytics framework. Soc Network Anal Min. 3:1277–1291.

[50] Television Bureau of Advertising, Inc. 2009. Multiplatform Glossary; [cited 2014 Sep 11]. Available from: http://www.tvb.org/trends/4758#s

[51] Teo TS, Choo WY. 2001. Assessing the impact of using the Internet for competitive intelligence. Inf Manage. 39:67–83.

[52] Thackeray R, Neiger BL., Hanson CL, McKenzie JF. 2008. Enhancing promotional strategies within social mar- keting programs: use of Web 2.0 social media. Health Promotion Pract. 9:338–343.

[53] Thelwall M, Buckley K, Paltoglou G. 2012. Sentiment strength detection for the social web. J Am Soc Inf Sci Technol. 63:163–173.

[54] Trainor KJ, Andzulis JM, Rapp A, Agnihotri R. 2013. Social media technology usage and customer relationship performance: a capabilities-based examination of social CRM. J Bus Res. 67:1201–1208.

[55] Waters RD, Burnett E, Lamm A, Lucas J. 2009. Engaging stakeholders through social networking: how nonprofit organizations are using Facebook. Public Relations Rev. 35:102–106.

[56] Watson GH. 1993. Strategic benchmarking: How to rate your Company’s Performance against the world’s best. NY: John Wiley & Sons.

[57] Weinberg BD, Pehlivan E. 2011. Social spending: mana- ging the social media mix. Bus Horiz. 54:275–282.

[58] Xu K, Liao SS, Li J, Song Y. 2011. Mining comparative opinions from customer reviews for competitive intelli- gence. Decis Support Syst. 50:743–754.

[59] Xu XZ, Kaye GR. 1995. Building market intelligence systems for environment scanning. Logist Inf Manage. 8:22–29.

[60] Yasin MM. 2002. The theory and practice of benchmark- ing: then and now. Benchmarking Int J. 9:217–243.

[61] Yi M, Oh SG, Kim S. 2013. Comparison of social media use for the US and the Korean governments. Government Inf Q. 30:310–317.

[62] Yin RK. 2009. Case study research: design and methods. Vol. 5. Thousand Oaks, CA: Sage.

[63] Zhong N, Li Y, Wu S. 2012. Effective pattern discovery for text mining. IEEE Trans Knowl Data Eng. 24:30–44.

[64] Zikopoulos P, Parasuraman K, Deutsch T, Giles J, Corrigan D. 2012. Harness the power of big data: the IBM big data platform. New York, NY: McGraw Hill Professional.

Volume 56 Issue 2, Spring 2016 Journal of Computer Information Systems 155