Differentiation between B2B and B2C
companies on LinkedIn: A content
analysis for high customer engagement
Master Thesis
MSc Business Administration – Digital Business track
University of Amsterdam
Amsterdam Business School
Author: Sophie Louise Koopmans Student number: 10556419
Date of submission: June 22nd, 2018 Supervisor: Dr. Shameek Sinha
2
Statement of Originality
This document is written by Student Sophie Louise Koopmans who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
3
Abstract
This study researches to what extent there is a difference in customer engagement on LinkedIn between B2B and B2C companies, and what role social media content has in generating engagement. Posts from six company pages on LinkedIn were collected and a content analysis was conducted over the resulting 572 posts. Each post was assigned one of the twelve content types and one of the three relationship types. The data was quantitatively analyzed. The results indicate that there is indeed a difference between business models: B2C company posts generate higher customer engagement than B2B posts. There are also significant differences between content types: counts of five types differ significantly between B2B and B2C. Concerning the relationship phases (awareness, acquisition and retention), B2B companies post more awareness-related posts, and B2C companies post more acquisition-related posts. There are no significant differences in engagement between the relationship phases. For both content types and relationship phases, there is no significant moderating role for business models on customer engagement. Moreover, no three-way effect between business models, content types and relationship phases has been found. This study narrows the knowledge gap in B2B social media marketing research and the methodology for content analyses on social media is further developed. Furthermore, it helps practitioners to make deliberate decisions on social media strategy and allows them to select the right content types on social media for their respective business model to generate engagement. Concluding, some directions for future research are proposed.
Keywords: social media strategy, LinkedIn, B2B, B2C, content analysis, relationship phases, customer engagement
4
Contents
1. Introduction ... 8
1.1 Expected contributions to theory and practice ... 9
2. Theoretical Framework ... 11
2.1 Social media use in business ... 11
2.2 B2B vs. B2C ... 12 2.3 Customer engagement ... 13 2.4 Content type ... 15 2.5 Relationship phases ... 19 2.5.1 Awareness ... 20 2.5.2 Acquisition ... 21 2.5.3 Retention ... 21
2.6 Relationship between Content Type and Relationship Phase ... 23
2.7 Conceptual Framework ... 26 3. Methodology ... 27 3.1 LinkedIn ... 27 3.2 Sample ... 28 3.3 Measurements ... 29 3.3.1 Business Model ... 29 3.3.2 Content Type ... 29 3.3.3 Relationship Phase ... 30 3.3.4 Customer Engagement ... 31 3.4 Procedure ... 32 3.4.1 Data Collection ... 32 3.4.2 Content Analysis ... 33 3.5 Reliability ... 34 3.6 Data Analysis ... 34
5
4. Results ... 35
4.1 Reliability ... 35
4.2 Result per hypothesis ... 36
5. Discussion ... 47
5.1 Business models ... 47
5.2 Content types ... 48
5.3 Relationship Phases ... 51
5.4 Missing three-way effect ... 52
5.5 Limitations ... 56
5.6 Theoretical and Managerial Implications ... 57
6. Conclusion ... 58
6.1 Future research ... 60
7. References ... 61
8. Appendix ... 67
8.1 Coding guidelines second coder ... 67
8.2 Descriptives tables ... 69
8.3 Content type preferences ... 71
6
Table of Figures
1. Figure 1 Conceptual model of variables business model, content type and customer engagement ... 19 2. Figure 2 Simplified customer journey in relationship phases ... 19 3. Figure 3 Conceptual model of variables business model, relationship phase and
customer engagement ... 23 4. Figure 4 Conceptual model of variables business model, content type and relationship
phase ... 26 5. Figure 5 Conceptual Framework ... 26 6. Figure 6 Interaction plot of the effect of business model on the relationship between
content type and customer engagement ... 42 7. Figure 7 Interaction plot of the effect of business model on the relationship between
7
Table of Tables
1. Table 1 Sample of selected companies ... 29
2. Table 2 Coding sheet Content Types ... 30
3. Table 3 Descriptives of the engagement scores per business model groups ... 36
4. Table 4 Descriptives and chi square results of content types ... 38
5. Table 5 Descriptives and chi square results of broad content categories ... 39
6. Table 6 Games Howell Post Hoc Test: Mean differences ... 41
7. Table 7 Descriptives and chi square results ... 43
8. Table 8 Hochberg's GT2 Post Hoc test results ... 44
9. Table 9 Results of the two-way independent ANOVA ... 45
10. Table 10 Partial Associations: Effect of each variable in the loglinear model ... 47
11. Table 11 The Content Decision Table: Matrix with preferences of content types over other types. ... 49
12. Table 12 Top 10 of the most engaging content types for customers. ... 50
13. Table 13 Descriptives: Counts of posts in each category in the loglinear analysis ... 54
14. Table 14 Overview of hypotheses and corresponding results ... 55
15. Table 15 Coding sheet: Content Categories ... 67
16. Table 16 Coding Sheet: Relationship Phases ... 68
17. Table 17 Descriptives of customer engagement for each content type. Sorted on Means. ... 69
18. Table 18 Descriptives of engagement scores per group ... 70
19. Table 19 Results of engagement comparing content types ... 71
8
1. Introduction
Social media is everywhere and the number of users keeps growing. LinkedIn has passed the 500 million users (Awan, 2017). The advancement of technology in the last decades is making the world more connected. The new domain of digital marketing is becoming ubiquitous and is expanding (Brennan & Croft, 2012; Peters, Chen, Kaplan, Ognibeni, & Pauwels, 2013).
Social media is an important component of digital marketing. Social media platforms are defined as ”forms of electronic communication through which users create online communities to share information, ideas, personal messages, and other content” (Merriam-Webster, 2018). Nowadays, social media have become indispensable in business. In 2015, already 96% of businesses globally used social media (Statista, 2015). Facebook, Twitter, Instagram and LinkedIn are the most popular examples of social media marketing channels.
Some industries are frontrunners in the digital marketing domain. Especially in business to consumer (B2C) marketing, social media grew from an underestimated channel to one that marketers can’t do without (Brennan & Croft, 2012; Peters et al., 2013; Rodriguez, Peterson, & Krishnan, 2012). This caused an increase in research in B2C digital marketing. While research in B2C social media marketing is advancing and becoming more mature, research on business to business (B2B) marketing is far behind and still in its embryonic stage (Salo, 2017). The shortage in research cannot be the result of a lack of size or relevance of the B2B industry, since B2B transactions account for 42% of reported US revenues in 2010 (Lilien, 2016). As more businesses are using social media, marketing scholars are urgently calling for further research on the distinction between B2B and B2C social media (Lilien, 2016; Mora Cortez & Johnston, 2017; Salo, 2017).
The interest of social media usage differs between the business to business and business to consumer contexts. B2B companies are more focused on customer relationship management, whereas B2C companies’ main goal is to create brand awareness (Michaelidou, Siamagka, &
9
Christodoulides, 2011). Hoffman and Fodor (2010) proposed to forget the traditional (monetary) measures of return on investment (ROI) as a measure of performance, but instead assess customer engagement, as customer engagement might better predict the likelihood of payoff in the long term. The common focus on “show me the return” by companies is too much oriented on the short term and dismisses the qualitative opportunities that social media might bring (Hoffman & Fodor, 2010). Thus, performance comes through customer engagement; digital marketers should focus on increasing engagement instead of directly trying to increase sales through social media. The goal of social media is not selling; it’s engaging.
Customers can engage with companies on social media by liking and commenting on firm-generated content or by creating content about companies (user-generated content). Social media content is textual, visual and aural content that is posted on a social media platform by any member, including textual messages, images, audio, videos, website links, and news articles. Content plays a crucial role in the engagement of customers. Good content engages customers, but the question is what content works for B2B and B2C companies.
There has been little research on how the different company contexts influence customer engagement on social media, or if there is any difference in customer engagement between B2B and B2C contexts at all. And what influence does the type of content that firms posts on its social media company pages have on customer engagement? This study addresses this query. It attempts to answer the following research question:
To what extent does customer engagement on social media differ in the B2B and B2C contexts, and what role does content play in this relationship?
1.1 Expected contributions to theory and practice
First, this study adds to the literature about the distinction of B2B and B2C markets. B2B and B2C differ in such a degree that B2C measurements cannot bluntly be copied into B2B
10
applications (Swani, Brown, & Milne, 2014). A distinction in business models must be made for future research to progress. Also, this study contributes to narrowing the B2B knowledge gap proposed by Lilien (2016). Lilien studied the imbalance between the volume of B2B and B2C research and uncovered opportunities and challenges of B2B digital marketing. Also, academics are calling for more explanatory research in the industrial marketing field (La Placa & da Silva, 2016). This study extends the limited quantified knowledge on customer engagement on social media in the B2B and B2C context. The knowledge can be used in practice to develop targeted social media strategies for firms in B2B and B2C markets.
Second, this study might encourage more (non-)social media using companies to seize the opportunities of this type of marketing and adapt their marketing strategies on their business model. In a study on the usage, perceived barriers and the measurement of effectiveness of the social networks of small and medium-sized enterprises (SMEs) in the UK, Michaelidou, Siamagka and Christodoulides (2011) found that only 27% of the B2B SMEs were actively using social media networks. 53% of the social media-using participating companies indicated that they did not measure the effectiveness of social media at all, mostly due to their lack of knowledge of the right metrics. 61% of them did claim to consider evaluating the metrics in the near future (Michaelidou et al., 2011). This thesis attempts to reduce the lack of knowledge and provides tools and guidance for academics and practitioners to increase the effectivity and performance of social media usage for companies.
Third, the results might provide direction for both B2B and B2C companies on which content works best for their specific customer: businesses or consumers. This research looks at what objective the content supports and which type of content increases customer engagement. With this insight, marketers can increase overall customer engagement by posting B2B- or B2C-specific content, which will benefit companies in the long run (Hoffman & Fodor, 2010).
11
The structure of this research is as follows. First, extant literature on social media content and customer engagement is discussed to build a conceptual framework and form hypotheses. Second, the methodology of this research is described. Third, the results are discussed in the sequence of the hypotheses. Forth, the results are deliberated on in the discussion section, limitations are uncovered, and theoretical and practical implications are revealed. To conclude, the main research question is answered in the conclusion and directions for future research are proposed.
2. Theoretical Framework
2.1 Social media use in business
Social media allows firms to communicate and engage more extensively online with customers than in the past. Social media marketing offers a solution for the declining effectiveness of traditional marketing techniques (Holliman & Rowley, 2014). Despite the potential, social media is not as much used as traditional digital tools like email and websites in business, and is still underutilized by industrial companies (Karjaluoto, Mustonen, & Ulkuniemi, 2015).
The extent to which companies and employees use social media for professional purposes depends on several factors: technology savviness, personal use and the use of social media in work processes. Taking insights from different studies together suggests that the steps to increase social media usage in business are, firstly, to let employees get social media savvy personally, e.g. encourage them to make a personal account (Keinänen & Kuivalainen, 2015; Schultz, Schwepker, & Good, 2012); secondly, to implement social media network usage in the internal organization; after which, thirdly, the reach can be expanded externally to customers (Karjaluoto et al., 2015). The literature that provides these insights does not distinguish between
12
different business models. However, B2B and B2C companies are likely to act differently on social media.
2.2 B2B vs. B2C
Extant research has studied the differences in customers, companies and sales people between B2B and B2C. First, customers of B2B companies have different motivations to use social media than customers of B2C companies. Customers of B2B companies are professionals who represent their employer. The customer and company in the B2B context build a professional relationship. Customers of B2C companies, on the other hand, represent themselves. They buy products or use services for their own personal needs. The objectives of the customer to use social media are to find product and service information, share their experiences or join in co-creation on social media (Constantinides, Schepers, & De Vries, 2015; Moore, Hopkins, & Raymond, 2013).
Second, B2B and B2C companies themselves differ. The general objectives of companies for using digital channels are to create awareness, enhance customer relationships and support sales (Karjaluoto et al., 2015). These key objectives are different for B2B companies than for B2C companies. For B2C companies, the most important objective to use social media is to create brand awareness by consumers (Michaelidou et al., 2011). Furthermore, providing information, having interaction and doing sales is also important for B2C companies (Constantinides et al., 2015). On the other hand, attracting new customers and cultivating customer relationships are the most important objectives of social media for B2B companies. While creating brand awareness is the main objective in the B2C context, its only third for B2B companies (Michaelidou et al., 2011).
Third, B2B sales people seem to use social media more than B2C sales people. A study found that when comparing social media usage of B2B and B2C sales people, the utilization
13
percentage of B2B salesmen was significantly higher than the number of B2C sales people on many social media platforms (Moore et al., 2013). This would imply that the adoption rate of social media is higher for B2B sales people than for B2C sales people. It can be explained by the relationship-oriented selling practices of B2B sales people. For B2B sales people, quality of relationships is more important, while B2C sales people use social media to reach a large quantity of potential customers (Moore et al., 2013).
2.3 Customer engagement
These differences between B2B and B2C markets in customers, companies, and its sales people are expected to endure in customer engagement. But what is customer engagement? Scholars do not seem to agree on one general definition as customer engagement is a very broad concept (Brodie, Hollebeek, Jurić, & Ilić, 2011; Malthouse, Haenlein, Skiera, Wege, & Zhang, 2013; Sashi, 2012). The much used definition of customer engagement proposed by Brodie et al. (2011) is: “a psychological state, which occurs by virtue of interactive customer experiences with a focal agent/object within specific service relationships”. It emphasizes the interactive nature of social media on which both sides (company and customer) can create content and interact.
To examine the effectiveness of the interactions, social media is becoming more data driven. Marketers utilize social Customer Relationship Management (CRM) systems to collect data on customer engagement, from which they can gather insights on how to reach new customers and enhance relationships more efficiently and effectively (Trainor, Andzulis, Rapp, & Agnihotri, 2014). These insights improve relationships and facilitate integration of the obtained information in the sales process (Rodriguez, Peterson, & Ajjan, 2015). Keeping consumers engaged is crucial to increase the value of interactive firm-consumer relationships.
14
Barger & Labrecque (2013) identified four types of social media users based on their customer engagement: bystanders, followers, participants and advocates. Bystanders stumble upon posts or mentions of a brand, but do not interact. Followers neither engage with a brand, but do search for them on social media and actively follow their posts. Participants engage and interact with a brand by liking or commenting on their posts. Lastly, advocates both interact with and actively promote a brand by creating and uploading their own positive content (Barger & Labrecque, 2013).
Customers of B2C companies have found their way to social media and are more likely to be participants or advocates. According to Keinänen and Kuivalainen (2015), B2B customers are still fairly passive on social media, as they use it as a source of information and do not actively comment nor open a discussion. Therefore, most B2B customers can be classified as bystanders and followers. Companies should think about how to activate business customers on social media and turn them into participants and advocates to get the true value from the platforms.
Cuillierier (2016) argues that customer engagement is more important for B2B firms, because client bases are smaller and customers buy in large volumes or do larger transactions, as compared to B2C firms. Maintaining good relations with customers and retaining customer engagement should therefore be more critical to B2B firms than to B2C firms. Also, B2B sales people are more focused on building relationships with the help of social media and are utilizing the platforms more for these purposes (Moore et al., 2013) than B2C sales people. It is expected that customers reciprocate the efforts of B2B salespeople. Therefore, it is hypothesized that:
Hypothesis 1: Customer engagement is higher on B2B company posts than on B2C company posts on LinkedIn.
15
2.4 Content type
Besides the objectives of firms to post on social media networks, content of posts (firm-generated content) is controlled by firms and might have an influence on customer engagement. Content on social media is created by corporate, employee, professional and civilian users. Firms have limited direct control what customers see and do on social media, but firms do have an (indirect) influence on the customer engagement of a post. B2B companies influence customer engagement directly with corporate user accounts by adding new content, participating in discussions, and removing content with corporate accounts (Huotari et al., 2015). So, the content companies post is important for customer engagement.
Some types of content might work better than others to involve customers. There haven’t been many studies on which content works on social media and which doesn’t. In other disciplines of digital marketing has been some more research. For instance, in search engine advertising (SEA) stating a call to action in an search engine advertisement’s content has a positive effect on the engagement (click through rate) (Rutz & Trusov, 2011). Another study found that increasing the interactivity of a post was found to have a strong effect on customer engagement on posts, measured by likes, comments and shares (Luarn, Lin, & Chiu, 2015). Semantics, writing style and evidence type of content also influence engagement (Haans, Raassens, & van Hout, 2013), but are out of the scope of this study. This study is primarily focused on subjects of content.
The individual content types can be divided into three broad categories: promotion, reputation, and events. The content types in the promotion category promote products and services from the companies. They are directly related to sales. The promotion category includes the content types promotion/discount, contests, and products and services. With the second category, reputation posts, a company can influence the image a customer has of the company, and present the company as an industry expert. The reputation posts include posts
16
about awards, CSR, knowledge sharing, reports, industry news, company information, and workforce/recruitment. In the third overarching category, content types are related to events, which include content types like presence at internal or external events, conferences or award shows. As there is no prior study which provides a categorization, the following categorization is based on assumptions.
The type of content used may be dependent on the business model. For example, B2C firms may make more use of certain types of content than B2B firms, like content to enhance awareness among customers. Because the objectives for using social media differ for B2B and B2C, it is expected that business to business and business to consumer companies emphasize different types of content.
For instance, I expect that B2C companies engage more in promotional activities like (publicly announced) discounts and contests. B2B companies are less likely to engage in these types of promotions, as their individual customers have to some extent more bargaining power than B2C customers, because the transactions in the B2B industries are larger and the products and services are more personalized on the client’s preferences (Cuillierier, 2016). B2B marketers rely more on (personal) emotional appeals than functional appeals (Swani et al., 2014). On the other side, B2C customers have less bargaining power as prices and product specifications are commonly fixed. Also, the relationship with a brand is less personal (Homburg, Klarmann, & Schmitt, 2010) and B2C companies must rely more on mass communication channels, like social media, to promote their products, because the quantity of potential customers is larger than for B2B companies (Moore et al., 2013). Therefore, I hypothesize:
17
Customers may be more responsive to a “1+1 for free!” promotional post than to “Did you know [fact] about our product?” as a product introduction post. The type of content of posts may influence the amount of likes and comments on posts. It is expected that some types of content interact better with either B2B or B2C customers than other types. I expect that the content types that are popular with B2B or B2C marketers, are also more engaging for their customers. The business model works as a moderator on the relationship between content type and customer engagement. So, in line with the previous hypothesis that B2C companies post more promotional content, it is expected that these posts generate higher engagement for this target group. Therefore:
Hypothesis 3a: B2C customers are more engaged in promotional content than B2B customers.
Content types in the reputation category are important for both B2B and B2C companies (Smaiziene & Jucevicius, 2009). Building a positive brand reputation on social media causes an increase in firm performance (Swani et al., 2014), which is an objective of all companies. However, having a good corporate reputation and posting about this might be slightly more important to B2B companies. Especially in the professional services industry, reputation is very important to retain loyal customers (Walsh, Beatty, & Holloway, 2015). Though, I expect the difference between B2B and B2C to be only small.
Hypothesis 2b: B2B companies post slightly more posts to increase reputation than B2C companies.
18
Hypothesis 3b: B2B customers are slightly more engaged in reputation posts than B2C customers.
Furthermore, as B2B companies are more focused on individual relationships with customers and can adapt their efforts to some degree of personalization, B2B companies are more involved in organizing, joining and promoting events. During events, they can meet and engage with customers on a more personal level. On the other hand, B2C companies have a less personalized relationship with their customers, and due to the relatively large size of their customer base, they are less likely to post about events to reach their individual customers. Therefore, it is expected that B2B companies post more event-related posts to engage customers than B2C companies.
Hypothesis 2c: B2B companies post significantly more event-related content than B2C companies.
Once again, in line with the reasoning of the hypothesis about content, I expect the following:
Hypothesis 3c: B2B customers are more engaged with event-related posts than B2C customers.
Throughout the theoretical framework of this paper, the conceptual framework will be build up out of smaller parts. The full conceptual framework is given in the end. Figure 1 shows a conceptual model that connects the (sub) hypotheses 1, 2 and 3. The moderating effect of
19
business models on the relationship between content type and customer engagement is left out of this figure.
Figure 1 Conceptual model of variables business model, content type and customer engagement
2.5 Relationship phases
As mentioned before, firms post content on social media to support their relationship objectives: raising (brand) awareness, doing customer acquisition or enhancing customer retention (Michaelidou et al., 2011). These objectives form a customer’s simplified journey (see Figure 2). Companies can create content on social media targeted at customers in one of the three phases of the customer journey. A relationship that is still in the awareness phase targets potential customers who do not know or barely have knowledge about a brand. In the acquisition phase, potential customers are encouraged to buy a product or use a service. The potential customers might never have used the brand before or only incidentally. In the retention phase, relationships with customers are present, and companies like to invigorate the relationships and encourage customers to re-use the brand. Below, all relationship phases are further explained.
Figure 2 Simplified customer journey in relationship phases
20 2.5.1 Awareness
Customers gain awareness through exposure to a brand or product. Hoffman and Fodor (2010) describe brand awareness in the social media environment as: “every time a person uses an application designed by or about the company, the company gains increased exposure to its brand, often in highly relevant contexts.” The exposure enhances and strengthens associations of the brand in the customers’ minds. Awareness posts are not necessarily personalized, at best segmented, because in general the first awareness messages are meant for a wide audience. The goal of raising awareness is to inform customers about the existence of a brand, product or service, not yet actively acquiring or converting them as customers. That is the next step in the customer journey. In traditional marketing, television advertisements or billboards are examples of ways to create customer awareness. On social media, this could be online advertisements, (sponsored) posts, mentions or events.
In B2C markets, social media are used for one-to-many communications and posts are quite general. In B2B markets the focus is a slightly more personal, even in this first stage where potential customers get acquainted with brands and products. The social media channels can be used for one-to-many, but also one-to-one communications. It is not likely that B2C marketers contact individual customers to let them know about a certain new product. The return on (time) investment is too low for B2C, while in B2B ‘cold calling’ on social media may be worth the time investment. On the other hand, the main objective for B2C marketers to use social media is to create awareness (Michaelidou et al., 2011). Accordingly, they will focus their social media strategy on awareness posts and aim to reach many potential customers with one message (one-to-many marketing). Therefore, I expect that B2C companies post more awareness posts.
21
2.5.2 Acquisition
In the acquisition phase, companies, which already created awareness, start encouraging prospective customers to buy a product or use a service. This may be the first time a customer converts, or a customer has incidentally bought the product or service before, without reaching the retention phase. Online acquisition efforts include running promotions about a product or service or giving discounts on e.g. Facebook or LinkedIn. In traditional marketing, this could be promotions in brochures, on television or in magazines. An important difference between traditional and online marketing is the ability to improve targeting and personalization in online marketing (Malthouse et al., 2013). For instance, a web shop could run personalized ads of products a customer viewed online, but did not buy yet. This is an example of retargeting.
Another distinction between B2B and B2C in the acquisition phase is the difference in the origin of demand. For B2B, demand is derived by a subsequent customer; the buyer is not likely to be the end user, while in B2C demand is driven by the specific tastes, emotions and preferences of the end customer (Lilien, 2016; Mora Cortez & Johnston, 2017). Furthermore, the buying decision process differs between the two markets. Where in B2C the decision maker is most likely the same person as the consumer, in B2B the decision making process is more complex and involves more stakeholders (Mora Cortez & Johnston, 2017). For example, a consumer of a product (an employee) does not make the buying decision, but his/her boss does. This implies that it is more about quantity of demand in the B2C markets and more about quality of relationships in the B2B markets. Therefore, it is hypothesized:
Hypothesis 4b: B2B companies post more acquisition-related posts than B2C companies.
2.5.3 Retention
In the retention phase, customers are aware of a company’s offerings and have been successfully acquired as a customer. The goal of retention is to encourage them to return and
22
remain in an ongoing relationship with the company, preferably becoming loyal to the brand. Firms can retain existing customers by keeping them satisfied and engaged. Customer relationship management is very important in this stage. In online marketing, firms use cookies and re-marketing techniques to remind customers of the brand, product or service (Iankova, Davies, Archer-Brown, Marder, & Yau, 2018). The relationship and communication is becoming even more personal to create a bond with customers and let them know their business appreciated. With the remarketing technologies, it is possible to automate the process, yet make the messages personal. When the customer is successfully retained, the customer enters the loyalty loop and stays with the company. Once the customer has entered the loyalty loop, it is important to maintain the relationship and nurture the customer, otherwise the customer may become unsatisfied and switch to a competitor’s product or service (Thaichon & Quach, 2015). Hence, the retention phase is mainly about keeping the customer satisfied. In B2B, this is done by creating strong buyer-seller relationships. In B2C, buyer-seller relationships are weaker (Homburg et al., 2010), so customers base their choice to stay loyal to a brand on different factors. As B2B marketers are focused on building relationships with customers on social media, more than B2C marketers, I expect that B2B marketers post more retention posts than B2C marketers.
Hypothesis 4c: B2B companies post more retention-related posts than B2C companies.
In the acquisition and retention stages of the customer journey, relationships are established to some extent, and customers might feel more connected to a company’s brand. The relationship is more personal than in the awareness phase. Because of this feeling of connectedness, I expect higher customer engagement on acquisition and retention posts, which implies that B2B customers are more engaged than B2C customers. Therefore, I hypothesize:
23
Hypothesis 5a: Acquisition and Retention content generates higher Customer Engagement than Awareness content.
If hypotheses 4b, 4c, and 5a are true, this might indicate a moderating relationship of business model on the connection between relationship phase of the content and engagement. Therefore, I additionally hypothesize:
Hypothesis 5b: Relationship content posted by B2B companies create higher engagement than relationship content by B2C companies.
Figure 3 shows a conceptual model that connects the (sub) hypotheses 1, 4 and 5. The moderating effect of business model on the relation between relationship phases and customer engagement is left out of this figure.
Figure 3 Conceptual model of variables business model, relationship phase and customer engagement
2.6 Relationship between Content Type and Relationship Phase
The two variables relationship phase and content type are studied separately in this study. Although the relationship phase and content type of posts might match in some instances, like a ‘product introduction’ promotion always targets the acquisition objective, there may also be promotions that rewards only existing customers, and therefore targets the retention of
24
customers. This way, the same content category promotion can target two different relationship phases. Interestingly, which types of content and relationship phases are connected, might deviate between the B2B and B2C markets. What content types can be linked to certain relationship types in which markets? Marketers could use this information to determine what content to create when they aim to influence customers in a certain relationship phase, according to their business model.
Few scholars studied the connection of content types to the relationship phase, so there is little academic knowledge in this area. This study is the first to address this question. Using the two business models (B2B and B2C), the three broad content categories (promotion, reputation and events) and the three relationship phases (awareness, acquisition and retention), I framed hypotheses based on intuition.
I expect the types of promotional posts that the two different business markets use to differ. B2C companies probably use promotional posts to incentivize potential customers to buy the products by offering discounts, targeting the acquisition phase. B2B companies are less keen on giving discounts, as mentioned before. B2B companies probably use promotional posts to introduce and inform customers about products or services, targeting the awareness and retention phases. Therefore, I hypothesize:
Hypothesis 6a: Promotion posts in the B2C context mainly target the acquisition phase.
Hypothesis 6b: Promotion posts in the B2B context mainly target the awareness and retention phases.
Next, reputation posts are presumably used by B2C companies to raise awareness about the company’s products and capabilities, mainly creating trust in the brand. Therefore, reputation
25
posts by B2C companies are likely to be focused on the awareness phase. On the other hand, B2B companies use reputation posts for raising awareness with potential customers, but also use reputation for retention purposes. To retain customers and keep the company-customer relationship positive, B2B company would want to profile themselves as experts in their field of work by sharing reports and industry news. These content types would be less relevant for B2C customers to increase a B2C company’s reputation. Therefore, I hypothesize:
Hypothesis 6c: Reputation posts in the B2C contexts mainly target the awareness phase.
Hypothesis 6d: Reputation posts in the B2B context mainly target the awareness and retention phases.
Finally, event-related posts could also have different purposes in the B2B and B2C contexts. If a B2C company would, for example, organize an event for its customers, it is probably to increase awareness about the brand by e.g. handing out free samples. On the contrary, B2B companies would organize events to invite existing customers for retention purposes, for example inviting customers to conferences. Therefore, I expect that B2B and B2C companies would use event-related posts to reach customers in different relationship phases. I hypothesize:
Hypothesis 6e: Event posts in the B2C context mainly target the awareness phase.
26
Figure 4 shows a conceptual model of hypothesis 6 that connects the variables business model, content type and relationship phase. It is expected that there is a sequential relationship between the three variables.
Figure 4 Conceptual model of variables business model, content type and relationship phase
2.7 Conceptual Framework
The conceptual framework below (Figure 5) shows all the variables and hypotheses (excluding the moderating effects) schematically. The three previous conceptual models are merged into one. This research ultimately studies the relationship between B2B and B2C business models and customer engagement on social media. It also looks at the different types of content and how these types of content are related to relationship phases. Next, the method of how these subjects were investigated is discussed.
27
3. Methodology
The purpose of the study is to assess the influence of business model and content on customer engagement. This study had an explanatory, cross-sectional research design. The data was measured at a specific point in time, and the sample was selected based on existing differences rather than random allocation (Labaree, 2009). For obtaining data on content of social media posts, this study used the methodological framework of content analysis (White & Marsh, 2006).
3.1 LinkedIn
The social medium assessed in this study is LinkedIn (www.linkedin.com). LinkedIn claims to be world’s largest professional network with more than 546 million users in more than 200 countries and territories worldwide. The mission is to connect the world’s professionals to make them more productive and successful (Linkedin Corporation, 2018). Individuals can display their resumes publicly online and link with other professionals. Companies can create corporate pages where they can upload posts (‘updates’), communicate vacancies and display information about the company. Members can opt-in to follow company pages.
Companies can also create corporate accounts on other social media platforms like Facebook, Twitter and Instagram. In this study, LinkedIn was chosen as the focal social media platform, because members represent both themselves (targeted by B2C’s) and companies (targeted by B2B’s) as they use the medium for professional purposes and communication. LinkedIn is the most important social media platform to follow a company professionally (Keinänen & Kuivalainen, 2015). Also, B2B marketers indicated that LinkedIn was the most important platform for them, however, B2C marketers preferred Facebook (Social Media Examiner, 2017). Since part of the contribution of this study is to narrow the B2B knowledge gap, the preference of B2B marketers was given more weight. Furthermore, the number of
28
studies on LinkedIn is limited, so our academic contribution would be more relevant when studying LinkedIn instead of Facebook.
3.2 Sample
The LinkedIn posts (‘updates’) of 6 companies were assessed. The selection included 3 B2B and 3 B2C firms (see Table 1). The companies were selected according to several criteria. Firstly, the company had to report at least an annual revenue of €100 million. Because of this revenue level, it is assumed that these companies have professional digital marketing employees/departments, so professional marketers create content on the LinkedIn page.
As B2B and B2C companies mostly operate in different sectors, getting all the sampled companies from only one sector is not feasible. Therefore, the companies in the sample all operated in different sectors. An advantage of this method of sampling is that the sample was not biased by one specific sector and the results are generalizable across sectors. On the other hand, a disadvantage of this sampling method is that it is harder to find significant results.
The LinkedIn profiles of the companies also had to meet several criteria. The selected companies had to be active on LinkedIn for at least one year and posted at least 27 times in the last six months (average of once a week). This would imply that the marketers had experience on writing content for LinkedIn. Furthermore, the company’s LinkedIn page had to have at least 10,000 followers, so there were enough LinkedIn members to engage. The content of company pages was also assessed to check if they used LinkedIn as a marketing and sales tool, not merely a recruitment tool, so maximum 25% of the lasts 50 posts should could have recruitment purposes (i.e. the page cannot exclusively contain content about vacancies and recruitment activities). The posts had to be in either Dutch or English. Lastly, if the company had both B2B and B2C divisions, it was checked that the LinkedIn content was not targeted at both B2B and B2C customers, but exclusively at one group. Following these requirements and using the
29
LinkedIn search tool, the following six companies were selected: Accenture, Maersk Line, Adyen, KLM Royal Dutch Airlines, Citroën and Louis Vuitton (Table 1).
Table 1 Sample of selected companies
Company Name
Industry Country Business Model LinkedIn Page Name Number of followers (10-4-2018) Accenture Management Consulting
Netherlands B2B Accenture the
Netherlands
19460
Adyen Financial Services Netherlands B2B Adyen 29482
Maersk Line Logistics and
Supply Chain
Denmark B2B Maersk Line 283988
Citroën Automotives France B2C Citroën 47485
KLM Royal Dutch Airlines
Airlines/Aviation Netherlands B2C KLM Royal
Dutch Airlines
437128
Louis Vuitton Luxury Goods &
Jewelry
France B2C Louis Vuitton 623467
3.3 Measurements
3.3.1 Business ModelDifferent business models determine in which business companies are and who they sell to. In the sample, only business to business (B2B) and business to consumer (B2C) are compared. A company’s business model is identified simply by looking at who their customers are: individual consumers (B2C) or other businesses (B2B)?
3.3.2 Content Type
The type of content of a post can differ in many aspects. Previous content analyses came up with categorizations of social media posts, for example Parsons (2011) identified and used 20 different categories of content posted by brands on Facebook. However, content is specific per social media platforms, therefore, the coding sheets of other researchers could not be reused. Based on adjusts of the existing coding sheets of Parsons (2011) and Ashley and Tuten (2015), the coding sheet in Table 2 was created. Each post received the best fitting content code.
30
Table 2 Coding sheet Content Types
Content type Broad category Description
1 Awards and
Congratulations
Events Notification of award won by company
2 Company Information Reputation History, announcement
3 Contests Promotion Call to action, do [this] and win …,
4 CSR Reputation CSR initiative by company/employee
5 Events and Conferences Events Post about event/conference company
representatives are attending
6 Industry news Reputation Happenings in industry
7 Knowledge Sharing Reputation Sharing knowledge, but not directly related to
product/service
8 Other Other Holiday greetings
9 Products and Services Promotion Introduction of product/service
10 Promotion and discount Promotion 2+2 for free, 20% discount if…
11 Reports Reputation Announcement report by company or
industry-related
12 Workforce/Recruitment Reputation Vacancy announcement or business course, or
reasons why the company is a good employer
3.3.3 Relationship Phase
The next variable, relationship phase, is a more abstract variable. Some guidelines have been made to identify the phase of a post and to ensure reliability in future studies. The descriptions are also included in the coding sheet in Appendix 8.1 Table 16.
• An awareness post is targeted to introduce customers to a brand, product or service. The content is non-specific and does not encourage explicitly to buy the brand/product/service.
• An acquisition post is targeted at potential customers to buy a product or service for the first time. The customer may be encouraged by a (first time) promotion or a discount.
31
• A retention post is targeted at existing customers. The post could mention a loyalty program or promotion, an invitation to events for customers, complementary services or the name of an existing customer as a best practice or case study of their work. The phases of a post were sometimes mutually inclusive, for example when a post targeted both new and existing customers (both acquisition and retention). However, only the best fit was assigned to a post for analytical purposes. Furthermore, posts about recruitment and job opportunities were excluded in this measurement, as they do not target customers in any phase of the customer journey. These posts received the code None.
3.3.4 Customer Engagement
The customer engagement measurements are based on existing social media metrics. The number of likes is a metric for low engagement and number of comments is medium engagement (Neiger et al., 2012). There are many other social media metrics for engagement, but only these two are used, because they are publicly available and retrievable. To get one general engagement score, the number of comments were weighted twice as heavy as the number of likes. To control for the number of followers per company page in the sample, as more followers increases the number of impressions and thus the possibility to engage, the number of likes and the number of comments were divided per 10,000 followers to standardize the metrics. Thus, the engagement score of a post was calculated as follows (formula 1):
(1) 𝐸𝑛𝑔𝑎𝑔𝑒𝑚𝑒𝑛𝑡 𝑆𝑐𝑜𝑟𝑒 = 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑙𝑖𝑘𝑒𝑠 + (𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑜𝑚𝑚𝑒𝑛𝑡𝑠 × 2) (𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑓𝑜𝑙𝑙𝑜𝑤𝑒𝑟𝑠/10000)
Number of shares is also a widely used metric for customer engagement across social media platforms, but as of February 2018 LinkedIn does not longer show the number of shares on posts or provide the count. LinkedIn states: “The share count on its own doesn’t fully reflect the impact that a piece of content delivers, and we encourage publishers and other content creators to leverage the [share tool] as a way to drive conversation and engage with members
32
on LinkedIn.” Therefore, only the number of likes and comments were used as measurements for customer engagement.
Logically, the longer a post has been online, the more exposure it could get and the more time there was to like or comment on the post. However, the time a post has been online was not considered in the data analysis of engagement. LinkedIn only displays recent posts on the timeline of users and it is assumed that users mostly engage with posts that appear on their timeline. From own experience, most likes and comments are given within a few days after posting the message. Research on Twitter messages confirmed that the majority of shares happens within one week (Priem & Costello, 2010). Therefore, it is unlikely that a three-month-old post received significantly more likes and comments than a one-month-three-month-old post. To control for recently posted updates which are still appearing on timelines, the post sample excluded posts younger than one week.
3.4 Procedure
3.4.1 Data CollectionAfter the sample was successfully composed, the collection of data started. The unit of analysis was the independent posts. For each company in the sample, the last 100 qualifying posts including the engagement data were collected. A post qualified when it met all the following criteria:
1. The post should contain at least three words (besides optionally an ‘attachment’ like e.g. a website link, report, photo, video).
2. The post must be generated by the official account of the company (firm generated content).
33
3. Posts posted in the seven days prior to the date of data collection were excluded. Customer engagement is not reliable yet, because not all followers had the chance to see the post, as explained before.
The collection was done either by using the LinkedIn API which collected several statistics for each post separately, or it was done manually by copying/pasting the data in Microsoft Excel. The first method had preference, but the LinkedIn API had the limitation that the data collector must be an administrator of the LinkedIn company page. Not all companies in the sample had provided this access to the company’s LinkedIn page, so these companies’ data had to be collected manually. After the content of the post, date, number of likes, number of comments, and number of followers of the company page were documented, the content analysis could begin.
3.4.2 Content Analysis
To obtain data on each post about the variables Relationship Phase and Content Type, each post had to be assessed on its content. According to Riff, Lacey and Fico (2014), “Quantitative content analysis is the systematic and replicable examination of symbols of communication, which have been assigned numeric values according to valid measurement rules, and the analysis of relationships involving those values in statistical methods, to describe the communication, draw inferences about its meaning, or infer from the communication to its context, both of production and consumption.”
Due to the manageable size of the sample of posts, the content of the LinkedIn post was coded and analyzed on content type and then categorized in a relationship phase, both manually. As the content of the post in the data file (Excel spreadsheet) only captured text and not interactive content, the content analysis was done based on the online version of the posts, and not just the text in the data file. To check whether the post was not altered in the time between
34
the data collection and the content analysis, the text of the post in the dataset and the text online were compared. If the post was altered, the post was excluded from the dataset.
The data was collected in different formats for the purpose of data analysis. For the business model variable, each company and corresponding posts received a categorical value of 0 (B2B) or 1 (B2C). For the relationship phase variable, each post was assessed on each category and received a 1 (Awareness), 2 (Acquisition) or 3 (Retention). For the content type variable, each post was assessed on each type and received the corresponding number 1 to 12 (see Table 2). Engagement was measured as an interval variable, and the number of likes and comments were recorded as the number displayed on the LinkedIn website.
3.5 Reliability
Before testing our hypotheses, the reliability of the acquired data was tested. The whole dataset was coded by one researcher (also the author of this paper). To guarantee reliable results and reduce the possibility of bias, one additional researcher was trained on the coding method and was asked to code 50 random posts on content type and on relationship phase. According to intercoder reliability research, the twice coded sample had to be at least 10% of the full sample (±30 in this study) or 50 units (Lombard, Snyder-Duch, & Bracken, 2002). To assess intercoder reliability, Krippendorff’s Alpha had to be at least 0.67 (De Swert, 2012; Krippendorff, 2004). This reliability test was done to ensure that the coding of the main researcher isn’t biased and the research can be replicated.
3.6 Data Analysis
The data was quantitatively analyzed in the IBM SPSS software program. To analyze hypothesis 1, the relation between business model and customer engagement, an independent t-test was conducted. The hypotheses 2a, 2b and 2c about the relation between business model
35
and content type were assessed by conducting a 2X12 Pearson’s chi-square test. However, when the assumptions of the chi-square test were violated, the test would be replaced by a Fisher’s exact test. This would be the case if counts were less than 5. After the 2X12 chi square, each category would be analyzed separately with a 2X1 chi square test. Hypotheses 3a, b and c about the relation between content type and customer engagement were tested with a one-way independent ANOVA and a factorial ANOVA to add business model as a moderator. Hypotheses 4a, b and c were analyzed with a 2X3 chi-square test to look for a statistical link between business model and relationship phase and 2X1 chi square tests for individual differences in business models. Hypothesis 5a about the association of relationship phase of posts to the engagement was analyzed with a one-way independent ANOVA. Hypothesis 5b, which studies the moderating effect of business model on the relation between relationship phase and customer engagement, was assessed by conducting an independent factorial ANOVA. Lastly, the influence of the business model on the relationship between content type and relationship phase was analyzed by conducting loglinear analysis.
Furthermore, as content analysis is also seen as a qualitative research method (White & Marsh, 2006), the quantitative results were complemented to offer additional insights with qualitative results such as frequencies, means and proportions. This was especially relevant for the descriptive statistical results of the chi-square tests. Descriptive statistics indicate an association, but are not modeling techniques and can therefore not predict values of the outcome variable. The reader should keep this in mind when interpreting the results.
4. Results
4.1 Reliability
Before starting statistical analyses, the reliability of the coding was assessed. The results of the Krippendorff’s Alpha showed that intercoder reliability was moderate for the content type
36
coding (α = 0.72) and on the border of acceptable for relationship phase coding (α = 0.69). As a reference, a Krippendorff’s Alpha of 0.67 is acceptable and 0.80 is high (Krippendorff, 2004). Thus, the dataset was reliable enough to proceed with the analysis according to the standards of Hayes and Krippendorff (2007).
4.2 Result per hypothesis
Hypothesis 1First, it was tested whether there is an association between the business model type and the level of customer engagement. In the independent t-test, Levene’s test showed equal variance could not be assumed (F (1, 513) = 28.06, p = .00). On average, posts from B2C generated higher engagement (M = 16.90, SD = 12.04) than B2B companies (M = 11.73, SD = 08.13) (see Table 3). This difference, 5.17 BCa 95% CI [6.95, 3.39], was significant, t (446.77) = -5.70, p = .00. It represented a medium-sized effect, Cohen’s d is 0.50. Hypothesis 1 stated that customer engagement is higher on B2B company posts than on B2C company posts on LinkedIn. However, the contrary was found. B2C customers were significantly more engaged.
Table 3 Descriptives of the engagement scores per business model groups
Business Model N M SD SE
B2B 259 11,734 8,1284 0,5050
B2C 256 16,903 12,044 0,7527
Note: 5 outliers were removed.
Hypothesis 2
To research the sub hypotheses 2a, b and c about the relationship between business model and content type, a Chi Square test was conducted over the counts of the separate content types and an additional Chi Square test over the broad content categories.
37
First, the Chi Square with the twelve content types was conducted. According to the descriptive of frequencies (Table 4), the category Promotion and Discounts had a value of 0 for the B2B business model. This broke the assumption that each value should be at least 1 and no more of the expected counts should be less than 5 (Field, 2013). To avoid losing a radical reduction of test power, this category was excluded from the Chi Square test. B2B companies never posted about this and B2C companies 8 times. We can infer from this that B2C companies used promotional posts more than B2B companies, but it could not be statically tested. Also, the category Other was excluded from the Chi Square test. The category was not relevant as it did not represent a single type of posts, but contained the random posts that could not be categorized. The number of posts labeled Other was low (total count is 5), so most posts of the data set could be classified in one of the other categories.
The 10X2 Chi Square test results showed that there was a significant association between business model and content type (χ2(9) = 65.06, p < .001). None of the assumptions were
violated. Cramer’s V had a value of 0.341, which means that the effect size is large. The conducted z-test within the 10X2 Chi Square showed that the proportions of some of the individual categories were significantly different between the two business models.
To specify which content types were used significantly more by one type of company, 2X1 Chi Square tests were conducted over the remaining ten categories (Table 4). Significant associations were found for the categories Contests (χ2(1) = 9.00, p = .003), Industry News (χ2(1) = 27.46, p < .001), Knowledge Sharing (χ2(1) = 6.63, p = .01), Products and Services
(χ2(1) = 13.26, p < .001), and Workforce and Recruitment posts (χ2(1) = 7.69, p = .006). B2B
companies posted significantly more about industry news and shared more knowledge than B2C companies, while B2C companies posted significantly more contests, products and services, and workforce and recruitment posts.
38
There were no significant associations for Awards and Congratulations (χ2(1) = 0.00, p = 1.00), Company Information (χ2(1) = 0.02, p = .89), CSR (χ2(1) = 0.57, p = .45), Events and
Conferences (χ2(1) = 0.05, p = .81), and Reports (χ2(1) = 2.31, p = .13). This implied that either
B2B or B2C did not post more about awards and congratulations, company information, CSR, events and conferences, or reports than the other company type.
Table 4 Descriptives and chi square results of content types
Content Type Broad Category Counts Chi Square Tests
B2B B2C Total χ2 df p-value
Awards and Congratulations Events 9 9 18 0.00 1 1.00
Company Information Reputation 25 26 51 0.02 1 .889
Contests Promotion 2 14 16 9.00 1 .003**
CSR Reputation 16 12 28 0.57 1 .450
Events and Conferences Events 39 37 76 0.05 1 .819
Industry News Reputation 33 2 35 27.46 1 .000***
Knowledge Sharing Reputation 64 38 102 6.63 1 .010**
Products and Services Promotion 51 95 146 13.26 1 .000***
Promotions and Discounts Promotion 0 8 8 - - -
Reports Reputation 22 13 35 2.31 1 .128
Workforce/Recruitment Reputation 36 16 52 7.69 1 .006**
Other None 1 4 5 1.80 1 .180
* p < .05, ** p < .01, *** p < .001
Second, the content types were lumped into the broad content categories and an additional chi square tests were conducted. There were significant associations found for the broad content categories Promotion (χ2(1) = 24.09, p = .00) and Reputation (χ2(1) = 26.14, p = .00), but no
39
significantly more promotion posts, B2B companies more reputation posts, and that there was no significant difference in counts for the event category (see Table 5).
Table 5 Descriptives and chi square results of broad content categories
Broad Category Counts Chi Square Tests
B2B B2C Total χ2 df p-value
Promotion 53 117 170 24.09 1 .000***
Reputation 196 107 303 26.14 1 .000***
Events 48 48 94 .04 1 .837
* p < .05, ** p < .01, *** p < .001
With these results, the sub hypotheses were assessed. Hypothesis 2a stated that B2C companies post more promotional content than B2B companies. The types Contests, Products and Services, and Promotions and Discounts fall in this category. The findings presented that contests, and products and services showed significant results that B2C companies indeed posted more promotional content. It seemed like the same applies to promotions and discounts, but this was not significantly tested. The broad content category confirmed that B2C companies posted significantly more promotional content than B2B companies. Overall, the hypothesis is accepted.
Hypothesis 2b stated that B2B companies post slightly more posts to increase reputation than B2C companies. The types CSR, Knowledge Sharing, Reports, Workforce and Recruitment, Company Information and Industry News fall in this category. Indeed, the hypothesis was accepted for the categories industry news and knowledge exchange. No significant results were found for CSR, Reports, and Company Information. However, Workforce and Recruitment was significantly more posted by B2C companies. The broad content category Reputation confirmed that B2B companies posted more reputational posts.
40
Given that the hypothesis mentioned that B2B’s post slightly more than B2C, this hypothesis is accepted, keeping in mind that this did not account for Workforce and Recruitment.
Hypothesis 2c stated that B2B companies post significantly more event-related content than B2C companies. The types Events and Conferences, and Awards and Congratulations fall in this category. However, the results showed no significant results for these two content types. Neither did the broad content category Events. Therefore, the hypothesis is rejected.
Hypothesis 3
To test the sub hypotheses of hypothesis 3, whether there is a relationship between content types and customer engagement and a possible moderating effect of business models, a factorial ANOVA was conducted. However, Levene’s test showed a significant result (p = 0.004), so the assumption of homogeneity of variance was violated. Therefore, first a one-way independent ANOVA was conducted with a Welch test to look at the relation between content type and customer engagement. The results (Welch’s F (10, 72.5) = 26.68, p < .001) showed there were significant differences in customer engagement between the content types. The Games-Howell Post Hoc test exposed that some categories generated either significantly higher or lower engagement than other types of posts. See Table 6 for the results over all categories. The most notable results were that the content category promotions and discounts generated to lowest engagement, and knowledge sharing posts didn’t perform very well either. Posts with company information performed the best.
To find out whether there is an interaction effect between the business model and content types, a two-way independent ANOVA was conducted. To account for the violated homogeneity of variance, the data of the dependent variable data had to be transformed by taking the square root of the engagement score to meet the homogeneity of error variance assumption. After this, the result of Levene’s test was that homogeneity of error variance could be assumed (p = .46). There was a significant main effect of the business model on customer
41
engagement, F (1, 499) = 12.93, p = .007. Just like in the one-way independent ANOVA, there was also a significant main effect of the content type of a post on customer engagement, F (10, 499) = 11.69, p = .00.
Table 6 Games Howell Post Hoc Test: Mean differences
Pr o mo ti o n a n d D isc o u n t C o n te st s E v en ts a n d C o n fe re n ce s A w a rd s a n d C o n g ra tu la ti o n s C S R Pr o d u ct s a n d S er v ic es Kn o w le d g e S h a ri n g R ep o rt s C o mp a n y In fo rma ti o n In d u st ry N ew s Ot h er
Promotion and Discount 0 -13,08 -13,59*** -18,58*** -9,10*** -13,81*** -6,72** -14,11*** -22,22*** -10,65*** -19,88 Contests 0 -0,51 -5,50 3,98 -0,73 6,36 -1,03 -9,13 2,43 -6,80
Events and Conferences 0 -4,99 4,49 -0,22 6,87* -0,52 -8,63 2,94 -6,29 Awards and Congratulations 0 9,48 4,77 11,86* 4,47 -3,64 7,93 -1,30
CSR 0 -4,71 2,38 -5,01 -13,12* -1,55 -10,78
Products and Services 0 7,09** -0,30 -8,41 3,16 -6,08 Knowledge Sharing 0 -7,39* -15,50*** -3,93 -13,16 Reports 0 -8,11 3,46 -5,77 Company Information 0 11,56 2,33 Industry News 0 -9,23 Other 0 * p < .05, ** p < .01, *** p < .001
However, no interaction effect was found between the content type of a post and the business model of the company placing the post, on the customer engagement (F (9, 499) = 1.02, p = .42). This indicated that it did not matter for customer engagement whether a B2B or B2C company posted a certain type of content. This means that hypotheses 3a, 3b and 3c are not supported, as there was no interaction effect and we could not distinguish between B2B and B2C. Figure 6 displays the interaction plot.
