Analysis Of Furphy Beer Customers And Production Forecast

Description of the Data

To accomplish allocated tasks, you need to examine and analyse the dataset (Furphy.xlsx) thoroughly. Below are some guidelines to follow:

Task 1 – Summarising Dependent Variables

The purpose of this task is to analyse and explore key features of these two variables individually. At the very least, you should thoroughly investigate relevant summary measures of these two variables. Proper visualisations should be used to illustrate key features.

Your technical report should describe ALL key aspects of each variable.

Task 2.1. – Identifying relevant factors for predicting repurchasing intention Analyse the relevant dependent variable against other variables included in the dataset. Your job is to decide which variables to include here. Use an appropriatetechnique to identify important relationships.

The outcome of this task is a list of variables that should be included in the subsequent analysis. Your technical report should describe why some variables were selected while others were dropped from subsequent analyses.

Task 2.2. – Model building (predicting repurchasing intention)

You should follow a model building process. All steps of the model building process should be included in your analysis. You can have as many Excel worksheets (tabs) as you require to clearly demonstrate different iterations of your predictive model.

Your technical report should clearly explain why the model may have undergone several iterations. Also, you must provide a detailed interpretation of ALL elements of the final model.

Task 2.3. – Interaction effect

To accomplish this task you need to develop a regression model using ONLY the factors discussed in the meeting (Task 2.3). In other words, this section of analysis is separate from the regression model constructed in Task 2.2.

Your technical report should clearly explain the role of each variable included in the model. A proper visualisation technique should be used. Make sure you interpret all relevant outputs in detail and provide managerial recommendations based on the results of your analysis.

Task 3.1. – Model building (likelihood of recommending Furphy)

You should follow a model building process. All steps of the model building process should be included in your analysis. You can have as many Excel worksheets (tabs) as you require to clearly demonstrate different iterations of your predictive model.

Your technical report should clearly explain why the model may have undergone several iterations.You must provide a detailed interpretation of ALL elements of the final model.

Task 3.2. – Visualising and interpreting predicted probabilities

Your technical report must include the predicted probability visualisation and be supplemented by practical recommendations to Furphy’s Management. These recommendations should answer the following question:

“How change in perceptions of quality and brand image  may affect the predicted probability of recommending Furphy by two customer segments (i.e. those purchasing directly, and those purchasing through sales representative)”

Task 4. – Forecasting production

Furphy’s quarterly beer production from the second quarter of 2008 until the end of the 2017 Financial Year are given in the Furphy_Product worksheet. Your job is to develop a proper forecasting model to predict turnover for the next four quarters.

In your technical report, you must explain the reason for selecting the forecasting method to predict future beer production. The report also must include a detailed interpretation of the final model (e.g. a practical interpretation of the time-series model, choices about smoothing techniques etc.).

Task 5. – Technical report

Your technical report must be as comprehensive as possible. ALL aspects of your analysis and final outputs must be described/interpreted in detail. Remember, your audience are well-experienced in analytics and expect nothing but perfection from your report. Perfection means quality content (demonstrated attention to details) as well as an aesthetically appealing report.

Description of the Data

An Australia born micro – Brewery Company named Furphy Beer has an experience of less than fifteen years in the brewing ale. The company has limited sales in Melbourne and reginal Victoria but is fast expanding in the rest of the parts of Australia in terms of its production and sales. This growth has been quite significant in the previous two years. The company has met its increasing demand in its pale ale beer by increasing its brewing capacity to 3 million litres per year. Their products are usually sold in pubs, bars, restaurants and bottleshops. The products are sold either directly or by a sales representative.

Though they have been successful in these operations and have experienced a good amount of financial turnovers over the years, they want to change the business to a new direction in the upcoming five years. This decision was made because of the increasing popularity of the brand. Thus, the management now is in need of a good relationship with its customers. They are also planning to forecast their production of beer in the next four quarters. With the help of this, the company will be able to understand their future demand and supply. Thus, they can plan their needs of production accordingly. To do this analysis, the company has asked a market research company named BEAUTIFUL DATA. The main aim of the company is to analyze the sales of the brewery company Furphy, understand the customers’ characteristics and their conditions for going for a repurchase and to predict the future sales for the next four quarters. The analysis will be done using the appropriate statistical techniques.

To analyze the situation and conditions described in the previous section, the company has made contact with another market research company named Beautiful-Data. This market research company in order to conduct the analysis has collected data from the clients of Furphy. The clients to fill up an online survey form. 200 valid responses were considered from all the responses received for the analysis. The dataset consist of 9 characteristics of the perceptions of the customers. The perceptions of the customers were recorded in a scale of 0-10. Other variables include the information on purchase outcomes and business relationships. Furphy’s data warehousing provided the market research company with another type of data. This data includes information of the quarterly beer production. The analysis will be done using the statistical software MS Excel.

The two dependent variables considered in this case are the “customers’ intentions to repurchase products from Furphy” and “whether a client would recommend Furphy to others”. From the analysis it is clear that most of the clients of the company have given ratings between 7 and 8. Most of the clients have given high ratings to the product. Thus, it can be said that most of the people go for repurchasing the product. Figure 3.1 shows the distribution of opinions of the clients for repurchasing intensions of the clients.

Distribution of ratings for repurchasing intensions of the clients

Data Analysis

The second dependent variable that has been considered is recommendations. It can be seen from the data collected that most of the clients recommend Furphy to others. From this data, it can be seen that the percentage of clients who recommend Furphy is very close to the percentage of people who do not recommend Furphy. Thus, it is not right to conclude anything from this information derived in this section. The information has been expressed graphically in figure 3.2.

In this part, the identification of the potential variables influencing the intention to repurchase Furphy Beer has been done. For the identification purpose, regression analysis has been used.

In the given dataset on Furphy Beer, a lot of variables are included. It is important to identify which variables are more important in influencing the intention to repurchase the brand of beer. To check this, a regression analysis has been done considering all the other variables included in the dataset predicting the dependent variable repurchasing intention (Draper and Smith 2014). The regression analysis will show very clearly which of the variables are significant in interpreting the dependent variable and which are not significant (Montgomery, Peck and Vining 2015). The regression analysis will give the p-values (significance values) of all the independent variables (Chatterjee and Hadi 2015). The test will be done at 95 percent level of significance. Now the variables with p-values less than 0.05 (95 percent level of significance) are termed as significant variables and with values greater than 0.05 are termed as insignificant (Kleinbaum et al.2015).

Here, the variables such as Loyalty, type of customer purchasing the product, the customer location and how the Furphy’s products are distributed have all been decoded to run the regression analysis. For “customer loyalty”, less than one year have been denoted as 1, between 1 to 5 years have been denoted as 2 and longer than 5 years have been denoted as 3. For The variable “customer type” the response Bottle Shops have been recorded as 1 and Pubs, bars and restaurants have been denoted as 2. For the variable “customer region”, Melbourne has been denoted as 1 and Outside Melbourne has been denoted as 2. Then regression analysis has been done on all the 13 independent variables to predict the dependent variable repurchasing intention of Furphy Beer.

From the analysis (Table 5 in Appendices) it can be seen that the P-values of the variables Loyalty, Distribution Channel, Quality and Brand Image of the product are less than 0.05. Thus, these are the only significant variables in predicting the repurchasing intention of the product.

To predict the intention to repurchase Furphy Beer, regression analysis has been done again with the significant variables only. From the analysis it is clear that the identified variables can predict 44 percent (R Square value, Table 6, Appendices) of the repurchasing intention of the customers to buy the beer (Cameron and Trivedi 2013). The predicted value or rating for the repurchasing intention can be given by the following relationship. The relationship is obtained from table 8 (Appendices).

Dependent Variable Analysis

Repurchasing Intention = 5.31 + (0.42 * Loyalty) – (0.29 * Distribution Channel) + (0.12 * Quality) + (0.21 * Brand Image).

It has been said that the research team manager has done a separate analysis and found out that the perception of beer quality is a significant predictor of repurchasing intentions. With his findings, it has been noticed that customers has a tendency to relate brand image and product quality. Thus, according to the manager, a good prediction can be made for the repurchasing intention with these two variables. The results of the analysis show that brand image and perception of beer quality can predict 34 percent (Table 9, Appendices) of the repurchasing intentions. It can be seen from the P-value (Table 11, Appendices) that both the variables quality and brand image are significant to predict the repurchasing intentions. The relationship can be given by

Repurchasing Intentions = 3.59 + (0.31 * Quality) + (0.31 * Brand Image)

The research team manager now wants to predict the likelihood of recommending Furphy to others. The specific interest of the manager is to understand the probability of clients who feel neutral about Furphy’s speed of delivery, with varying levels of perception towards product quality and brand image and clients who purchase directly and purchase via sales representative.

To understand this likelihood, the ratings given by the clients have been rounded off to the next whole number. To understand the likelihood of recommending Furphy to others by the clients who feel neutral about Furphy’s speed of delivery, the recommendation status of the clients giving a rating of 5 to the speed of delivery of the product has been collected. From table 12 it can be seen that the likelihood of recommending Furphy to others by these customers who feel neutral towards the speed of delivery of the product is 29 percent.

With varying levels of perception towards product quality, it can be seen clearly from table 14 that the likelihood of people recommending Furphy to others is 0.505. It can be seen that the lowest rating given by the clients is 6. Thus, it can be said that the product has quite a good quality.

Table 16 shows clearly the likelihoods of recommending the product with negative brand image, neutral brand image and positive brand image. It can be seen from the table that, 18.5 percent of the clients recommend the product with negative brand image, only 1 percent of the customers with neutral brand image recommend the product and around 31 percent of the customers with a positive brand image give recommendation to others for purchasing the Furphy brand of beer.

Table 18 shows the likelihoods of recommending the product to others by the customers who purchase the product directly and by the customers who purchase the product via sales representatives. It can be seen that 34 percent of the customers who purchase the product directly recommend the product to others and around 16.5 percent of the customers purchasing the product with the help of sales representatives recommend the product to others. Thus, from the nature of these likelihoods it can be said that the product is quite popular among the people of the country despite of the negative or positive ratings given by the clients. Irrespective of the ratings given by the customers (positive or negative), it can be seen that the clients do recommend the product to others.

Task 1 - Summarising Dependent Variables

From the data collected by the market research company from Furphy’s datamart on the sales of the product per quarter, it has been predicted that the sales of Furphy in the first quarter of 2018  will be around 1699.40 litres pale ale. In the second quarter of 2018, the sales will be around 1714 litres pale ale, in the third quarter of 2018, the sales will be somewhat around 1656.55 litres pale ale and the fourth or the last quarter the sales has been predicted to be around 1688.71 litres pale ale. The moving average method has been used to forecast the future sales is because a lot of past data points are available (Brockwell and Davis 2016). This will help in understanding the type of the trend clearly (Box et al. 2015). Other reasons for selecting this method in this case is because it is very easy to compute and is understood very easily (Montgomery, Jennings and Kulahci 2015). This method also provides stable forecasts to the data (Granger and Newbold 2014).

Conclusion

The analysis shows that most of the people have given high ratings to the product in terms of repurchase unit. Most people have recommended the product to others. From the regression analysis, it has been identified that the most significant variables to predict the repurchase intensity are loyalty towards the product, distribution channel, perceived quality of the product and the brand image of the product. These characteristics can predict the repurchasing intensity of the customers 44 percent correctly. From the significant variables identified by the manager of the market research company to predict the repurchasing intensity of the product, it has been observed that the variables quality and brand image can predict the repurchasing intensity 34 percent correctly. From the distribution of different variations of the ratings given by the customers it is evident that the product is quite common among the people of the country and is highly recommended. From the prediction of the future sales, it is evident that the sales of the product will increase in the year 2018.

Thus, it can be said that the significant variables such as customer loyalty, distribution channel, perceived quality of the product and the brand image of the product are responsible in the increase in the future sales of the product of the company. Thus, the company should take specific measures to develop these four conditions in order to increase the sales of Furphy beer more.

References

Box, G.E., Jenkins, G.M., Reinsel, G.C. and Ljung, G.M., 2015. Time series analysis: forecasting and control. John Wiley & Sons.

Brockwell, P.J. and Davis, R.A., 2013. Time series: theory and methods. Springer Science & Business Media.

Cameron, A.C. and Trivedi, P.K., 2013. Regression analysis of count data (Vol. 53). Cambridge university press.

Chatfield, C., 2016. The analysis of time series: an introduction. CRC press.

Chatterjee, S. and Hadi, A.S., 2015. Regression analysis by example. John Wiley & Sons.

Draper, N.R. and Smith, H., 2014. Applied regression analysis. John Wiley & Sons.

Granger, C.W.J. and Newbold, P., 2014. Forecasting economic time series. Academic Press.

Kleinbaum, D., Kupper, L., Nizam, A. and Rosenberg, E., 2013. Applied regression analysis and other multivariable methods. Nelson Education.

Montgomery, D.C., Jennings, C.L. and Kulahci, M., 2015. Introduction to time series analysis and forecasting. John Wiley & Sons.

Montgomery, D.C., Peck, E.A. and Vining, G.G., 2015. Introduction to linear regression analysis. John Wiley & Sons.

Nelson, E., 2016. Radically Elementary Probability Theory.(AM-117) (Vol. 117). Princeton University Press.