Sales Revenue in Return of Facebook Marketing - Essay Example

? Making decisions in Business Introduction Research scholars such as Saunders, Lewis and Thornhil (2009) pointed out that business reach is important part of strategic decision making for companies. Sekaran and Bougie (2009) also supported the fact that business decision makers need to analyze the past data in order to make valid future forecast. With the growing of technology and software classification, companies have started making decisions by optimizing historical data (Gray, 2009). Hence, the business world is moving from content driven decision making to data driven making. Now, the fact is that certain level of mathematical skill set is required in order analyze data and research techniques are also needed in order to gather the right data (Creswell, 2009). In such context, the researcher has decided to go deep in the topic of business research in order to understand how data is being collected and how this data can be analyzed in order make valid business decision. The essay will also take example of case study regarding sales revenue in return of Facebook marketing in order to shed light how statistical calculation is used for analyzing data. But, in the first section, the essay will shed light on data collection procedure. 1.1 Collection of Primary and Secondary Data Saunders, Lewis and Thornhil (2009) pointed out that researchers should be careful while collecting data regarding given business problem because business research is significantly different from academic research. For example, business researchers should focus more on real time data such as sales revenue after integrating certain advertising measures while academic researcher focus more on historical data in the form of literature review. For a business problem, researchers can use two types of data source such as primary and secondary. 1.1.1 Plan According to the instruction, I have been approached by marketing department of Blue Marketing in order to help them to analyze the success of the new social media (Facebook) marketing campaign which has been designed for their client cocktail cosmetics. Data sources for analyzing the success of marketing campaign can be following types; Data Sources Characteristics Validity Secondary Data Previous result of such kind of marketing that has been prepared by the company for other clients. Historical data from the company database regarding industry trend on social media marketing trend. Available business report on global social media marketing trend. Financial report of cocktail cosmetics which can serve as basal plan for understanding the resource capability of client for funding digital marketing campaign. No. of likes is Facebook page of the social media marketing campaign of cocktail cosmetics. Data sources should not be old (not before 2011) while each data should be collected after cross checking the authenticity. Data should be collected from updated databases and previous social media advertising campaigns should be adjusted as per present condition and size of cocktail cosmetics. Number of likes in Facebook page can help the Blue Marketing to assess the popularity of the campaign. These data are being collected in real time manner hence there is no question on data validity. Primary Data Sources Daily sales revenue data, Facebook Posts (FP) and Fan interactions (FI) are three major primary data sources for Blue Marketing to analyze the performance of the designed social media campaign. Survey method can also be used in order to collect data from Facebook users regarding their opinion about the campaign. Reliability statistics should be used in order to test the autocorrelation between the sample data and data having low responses having low should be deleted in order to increase validity of the data. Secondary data should be collected by accessing internet sources and paid business databases while primary data can be collected by visiting the Facebook page of the campaign and customer interaction page in Facebook. Sales data can be collected from the accounting department of cocktail cosmetics. Now, the question is done how the primary data through survey method can be collected? 1.1.1.1 Sampling Strategy All the users of Facebook do not have equal chance to view or like the posts regarding cocktail cosmetics hence probabilistic sampling cannot be used. In such context, I will recommend Snowball sampling should be used as part of non-probabilistic sampling for selecting the respondents. In Facebook, communities are created for members who share equal likings for a particular phenomenon hence by using Snowball sampling (one participant refers the other); I will get pool of respondents who have seen the social media advertising of cocktail cosmetics in Facebook page. A sample size of 1000 respondents will be good enough to understand the consumer behaviour towards Facebook campaign of cocktail cosmetics. 1.1.1.2 Data Collection Instruments Close ended questionnaire should be designed in order to collect the data. For nominal data such gender, no. of likes in Facebook Post etc, dichotomous questions should be designed while for measuring ordinal data, Likert rating questions should be designed. Questionnaire will be posted on the Facebook wall page and respondents will be asked to tick options in the questionnaire. A pilot study should be done in order to check durations of filling up the question and feedback from respondents will be used to decrease the numbers of questions so that respondents will not need more than 5 minutes to fill up the questionnaire. Open ended questions will not be used during survey because such types of question will take more time to fill up and many respondents will not be ready to provide that much time. Response from respondents will be collected in separate database and only valid data will be selected for data analysis after filtering the database as per validity, originality and completeness of the answer sheet. 2.1 Information Creation Hair et al. (2010) rightly pointed out that data should be presented in representative otherwise it becomes difficult for the respective researchers to find out right data set to analyze. Hence summarization of the data set is very important for the researcher and summarization of representative values of data will also help me working on behalf of marketing department of Blue Marketing to identify distribution pattern. Summary of the data set regarding Facebook advertisement of cocktail cosmetics can be presented in the following manner. Now, it is evident from the pattern of data that they are continuous in nature and they can take any value within the range of 0 to ?. Only data of days is nominal in nature and other three variables such as Facebook posts, Fan interaction and Sales data are continuous in nature. Let’s take the example of Sales data; the value can lie within the range of 0 to ? (theoretically). Sample space for the sales data, (S) = [200,......, 698]. There are thirty values and the probability density function can be expressed in the following manner 698?200P(x) dx=1 (Theoretically) From the integral model, it can be surmised that all the data are normally distributed hence statistical treatments such as correlation analysis and regression analysis can be done without any central value estimation. It is also evident from the data summarization that, Face book posts and Facebook Interaction has somewhat symmetric relationship with Fan interaction but same cannot be said for relationship between Face book posts and Sales or Facebook Interaction and Sales. In such context descriptive statistical measure should be used in order to bring more information from the data set. 2.2 &2.2 Data Analysis & Result Hair et al. (2010) pointed out that descriptive statistics help researchers to understand the nature of data and researchers can also understood which necessary statistical measures that can be used to analyze the business problem. Using mean, median, and mode can help the researcher to understand the average of data and identification of the spot where majority of value lies while variance and standard deviation can be used to understand the dispersion in the data set. Descriptive analysis of the data set can be presented in the following manner; Table 1: Descriptive Analysis Descriptive Statistics Facebook Post Fan Interaction Sales(?) Mean 18.3 254.1333333 450.9667 Standard Error 1.038511313 16.49156757 30.17026 Median 16 246 508.5 Mode 16 195 426 Range 21 336 598 Minimum 10 126 100 Maximum 31 462 698 Sum 549 7624 13529 Count 30 30 30 Confidence interval for the data set is being selected as 95% with significance level of 0.05 and as the response distribution has been selected as normal hence the response curve for variable such as Facebook Post, Fan Interaction and Sales(?) will be bell shaped. It is evident from the central tendency value, average face book post is little over 18 in a day 16 posts in a day was being repeated for more than once in the month. 195 fan interaction was being done maximum time in the month while average fan interaction in the month was within the range of 254 to 255. Average daily dales during the campaign were ?450.96 and daily sales of ?426 were repeated more than once in the month. It is also evident from the descriptive statistical table that maximum 31 Facebook posts were given during 2nd day of campaign and it has created 314 fan interactions which is way below than the maximum Fan interaction of 462. Maximum Fan interaction was 462 during 9th day when only 14 Facebook posts was given but surprisingly sales volume was lowest ?100 during that day. Now, such trend is creating the doubt that fan interaction might not ensure the sales revenue generation rather there should be other factors which has influenced the sales revenue generation. It is evident from the daily sales revenue figure from last 10 days that sales revenue continuously has continuously increased irrespective of the frequency change in Facebook posts or Fan interactions. Hence, relationship between Facebook posts and daily sales revenue of Fan Interaction and daily sales revenue is needed to be tested in order to make any conclusive statement. Quartile Analysis can be done in order to show the percentage of data lies in the particular region. Table 2: Quartile & Percentile Analysis Facebook Post Fan Interaction Sales(?) 1st Quartile=14.25 1st Quartile=190.5 1st Quartile=343.75 2nd Quartile= 16 2nd Quartile= 246 2nd Quartile= 508.5 3rd Quartile=23 3rd Quartile=307.75 3rd Quartile=558 20th Percentile= 13.8 20th Percentile= 179.2 20th Percentile= 258 60th Percentile= 17.4 60th Percentile= 288.4 60th Percentile= 524.2 For Facebook posts, 25% cases exist where 10 to 14 Facebook posts were posted while Blue Marketing posted the 10 to 23 Facebook Posts in 75% time. Hence it can be said that generally 10 to 23 Facebook posts are sufficient to create 126 to 308 fan interactions (range of 75% quartile). In 75% time in the month, sales revenue was in the range of ?100 to ?558. Same conclusion can be drawn by using the percentile method such as 60% Facebook posts were in the range of 10 to 17.4 while 60% fan interactions was in the range of 126 to 308 and 288.4 and 60% times, daily sales volume was in the range of ?100 to ?524.2. Dispersion of the dataset cannot be understood without identifying the normal distribution curve foe responses. In such context, the following Histogram can be used in order to understand nature of dispersion in the data set. Figure 1: Facebook Post Figure 2: Fan Interaction Figure 3: Daily Sales (?) Table 3: Measures of Dispersion Dispersion Facebook Post Fan Interaction Sales(?) Standard Deviation 5.688160723 90.32803564 165.2493 Sample Variance 32.35517241 8159.154023 27307.34 Kurtosis -0.719249866 -0.306531313 -0.72481 Skewness 0.568058565 0.552329997 -0.53958 It is evident from the table that the nature of the normal curves for Facebook Post, Fan interaction and daily sales are pretty much symmetric and due to low value of Kurtosis and skewness ( Read More