3.1 Research Design
This study adapted quantitative analysis method to conduct research. To explore the relation
between variables and validate the hypothesis, this study adapted partial least squares (PLS) to help validate.
In the other hand, to determine whether different intensities of content components affect customer engagement, we collected data from brand fan pages and analyzed their effect on customer
engagement. Figures 1 and 2 depict our research model. We used diverse dimensions of content components as research variables to explore their influence on customer engagement and verified whether resonance would affect customer engagement.
14
Figure 1. Conceptual Depiction of Research Model 1
Figure 2. Conceptual Depiction of Research Model 2
First, we collected the contents from the brand fan pages of Marie Claire Taiwan, There are three reasons why collected the contents from Marie Claire. First, Marie Claire is an international
magazine, which includes various topics, and it attracts different kinds of customers engage in their brand fans page. Second, due to our senior who works in Marie Claire, this study is more
understood empirical situation, and attained useful advice from him. Last, the value of customer engagement should be used the benefits and performances of companies to evaluate, which is credential and can’t attain from external. Moreover, with our senior help, this study can be verified the hypothesis by the data and doesn’t leak any business secret. Therefore, it is very helpful to this
15
study for collecting data from Marie Claire to conduct research on examining impact of content on social media.
In this study, our data came from two ways, in the questionnaires part, adopting NVivo software, which is a powerful application for qualitative analysis and provides many tools. One of NVivo’
tools is similar to the web crawler, which can download the data from any brand fans page. Besides, our senior also provides business performance data to evaluate the impact of factors in our study.
After collecting the data, we inserted the data into a database and constructed a website that would assist us in inviting interviewees to judge the intensity of diverse dimensions.
After the first test was finished, we adjusted the number of posts that would be judged by the interviewees and conducted a second test. We compared the tests results to find their common attribute and filtered invalid samples of the second test to be used as research data, which was analyzed by SmartPLS to examine our research model.
3.2 Research Processes:
Step1.
Collect one thousand posts from Marie Claire Taiwan for use as material in our questionnaires
Step2.
Input the posts into our database and construct a website that would help us give questionnaires to interviewees. See Figure 3.
Figure 3. Questionnaires from the website.
16
Step 3.
In the first test, we distributed 1,000 posts to 20 interviewees, so each interviewee had to answer fifty posts about their intension of dimension.
Step 4
Based on feedback from interviewees, we adjusted the number of posts in each interview. Thus, we redistributed the posts to each interviewee who has twenty-five post and totally have forty
interviewee to reduce their workload.
Step 5.
After two runs of tests, the table below shows the results.
Table 1. Description of samples
Test Interviewee Reviewed Posts Sample Valid sample
1 20 50 1000 859
2 40 25 1000 941
Step 6.
We classified the intension of dimension from a number into a level. Our principle of classification is shown below:
1-2 point → Low level 3-5 point → Middle level 5+ point → High level
We compared the results of each of the same posts in the two tests, as shown below.
Table 2. Comparison of two tests Common result Samples
More than 1 412 Completely same 50
Step 7.
The valid samples (941) from the second test were considered our research sample and were used to examine our research model with SmartPLS.
3.3 Method
This study explores the relationships between content development and resonance behavior and how that finally affects customer engagement. We downloaded relevant data from brand fan pages and interviewed fan page members. We then applied Netnography as our research method because it
17
tends to push the researcher to participate in virtual communities and become one of the members, which can help the researcher thoroughly observe the community’s activities and effects.
3.4 Research variables
Because we aimed to examine what factors can stimulate resonance behavior and affect customer engagement, we constructed a series of processes to verify our findings. We combined the concepts of content strategy, customer engagement, and customer intention into our research model. Table 3 specifies each variable in the content.
Table 3. Definition of research variables
Level Low Medium High
Vividness Number of elements is less than 3.
Interactivity Share a link to their website
Give a question or quiz to stimulate fans to respond
Host an activity or
contest to persuade fans
to do somethingRichness Promote the event, activity, and spaces content so people can easily understand the purpose of
Content Elements: Tag, link, picture, emoticons, sufficient sign, title.
18
Example of elements:
To evaluate the popularity of content, we recorded the number of likes and comments, which indicates that the content is successful and influential because customers take the time to express their opinion of the popularity of the content. Both variables have been used extensively to study SNS (Cvijikj, 2013;
L
isette de Vries et al., 2012). A high number of likes indicates that the content is interesting and accepted. Similarly, sharing can be considered to demonstrate high intensity of agreement with the content, as customers share content with their friends. Therefore, we used these variables to judge the intensity of content popularity (Ferran, 2014). In addition, we used the redirect-click-through-rate to examine whether the resonance behavior would affect customer engagement; thus, customer engagement was also one of our research variables.Table 4. Research variables and definitions
Research Variables Description
Likes Number of likes
Comments Number of comments
Shares
customer engagement
Number of shares
Number of redirect-click-through-rates
19
3.5 Data collection and analysis
We focused on Marie Claire Taiwan brand fan pages because they contain various topics, such as fashion, celebrities, and beauty. The brand provides such diverse contexts that they can reach different customer demographics. All details are displayed in Table 5.
Table 5. Description of sample resource
Brand fan pages Description Number of fans
Marie Claire Taiwan A monthly women’s magazine based on the French version that focuses on health, beauty, and fashion
378,327
We chose Facebook as a research platform because it is the largest and most frequently used SNS in the world. Its platform enables customers and companies to easily communicate with each other.
Customers who use Facebook exhibit many different behaviors that can help us evaluate the intensity of content popularity. Many other studies have used Facebook to examine customer engagement (Dholakia, 2004; Hansen, 2011).
Most of interviewee are teenage, because the interviewee are the person who spend much time on social media and notice regarding fashion and celebrity, which is related to our brand fans page, so this study invite many college to conduct questionnaires. Table 6 are more detail information.
Table 6. Demographic Information of Respondents (N = 40)
Measure Items Frequency Percent
Gender Male 14 35%