Executive Summary
This report aims to answer three questions, including what the company’s most valuable customers are, which customers are more inclined to respond positively to promotional offers, and which customers TeleMarket should target if it considers launching a new credit card. The report provides an overview of TeleMarket’s customer demographics to assist the company in identifying key traits of its typical customers and target promotions effectively. The findings demonstrate that the most valuable customers are younger males with increased higher education levels, substantial household income, and long-term use of TeleMarket’s services.
The report examines the influences on customers’ likelihood of accepting marketing offers. The results show that higher education levels and an increased number of TeleMarket’s services contribute to the probability of accepting promotions. Lastly, the report evaluates which customer segments should be targeted for credit card promotions, and the results suggest that the company should target older people with higher household income first.
Introduction
This report focuses on analyzing TeleMarket’s customers using statistical analysis of a subset of 4,500 customers. TeleMarket is a firm that operates in the United States. Despite its moderate success, the company aims to expand and diversify its services. To achieve this goal, TeleMarket has generated several concepts for further expansion and seeks to utilize marketing analytics to turn these ideas into reality. These proposals include creating a new credit card integrated into a mobile app, introducing a mobile app-based loan service, and providing car insurance. This report is expected assist the company in making informed decisions about its future through data analysis.
First, this report outlines TeleMarket’s customer demographics based on gender, age, household income, duration of service usage, and education level. This endeavor is expected to help the company understand the characteristics of its average customer so that it can target promotions successfully. Additionally, the report discusses which customers are most valuable to the company’s success.
Second, the report discusses factors that affect the likelihood of a customer accepting a marketing offer. In particular, the report discusses which of the characteristics, including gender, age, household income, number of years with TeleMarket, education, and number of services used by the customer are likely to stimulate a customer to accept a promotion. This information is crucial, as it can help target customers more likely to accept the offers. Finally, the report analyzes which customers should be targeted to promote credit cards. This is crucial, as one of the possible future development for the company is a new credit card integrated into a mobile app, which needs to be marketed effectively.
Methodologies
Different methods were used to achieve three objectives. The first objective was to describe TeleMarket’s customer base in terms of gender, age, household income, number of months with TeleMarket, and education and determine which of the customers are most valuable. Descriptive statistics was used to overview the crucial characteristics of the firm’s customers, including means, medians, modes, standard deviations, skewness, and kurtoses. Frequencies were used to describe categorical data. Descriptive statistics refers to the set of tools and techniques used to summarize, organize, and describe the key features of a set of data (McClaive, Benson, & Sincich, 2018). Its purpose is to provide a clear and concise summary of the data that can be easily understood and interpreted (McClaive et al., 2018). Thus, the use of descriptive statistics was appropriate for the purpose.
In order to answer the question of what the characteristics of the most valuable customer for the company are, it was first necessary to define the term “most valuable customer.” It was assumed that customers who use the most TeleMarket services are of the most value to the company. Therefore, a new variable was computed as the count of services a customer uses, including toll-free services, wireless services, rental equipment, multiple lines, caller ID, call waiting, call forwarding, conference calls, movie streaming service subscription, satellite TV subscription, and broadband internet use. The broadband internet variable was recoded to a binary variable with “0” for “none” and “1” for any type of internet connection. The new variable, “Number of Services Used,” was computed as a sum of the 12 mentioned above.
After the preliminary work, multiple regression analysis was performed to examine how demographic factors influence the number of services utilized. Multiple linear regression is a statistical method used to examine the relationship between two or more independent variables and a dependent variable (Newbold, Carlson, & Thorne, 2022). It is used when there is a need to predict or explain a dependent variable using multiple independent variables that are thought to be related to the dependent variable (Newbold et al., 2022).
In other words, multiple linear regression is used when there is a need to understand how a dependent variable changes as multiple independent variables change simultaneously. Since the analysis aimed to determine how several demographic variables affected the number of services used by the customer, multiple linear regression analysis appeared appropriate for the purpose. The following model was assessed:
Number of Services = β0 + β1 * Age + β2 * Years of Education + β3 * Months with Service + β4 * Household Income + β5 * Gender + ∈
The study’s second objective was to determine what factors had a significant effect on accepting the previous promotions of TeleMarket. Six variables, including age, gender, years of education, months with service, number of services used, and household income, were tested to impact the reply to the three promotions. Three binary logistic regression models were created to measure the effects of these variables on three dependent variables (Response to product offer 01, Response to product offer 02, and Response to product offer 03).
Binary logistic regression is a statistical technique used to model the relationship between a binary dependent variable and one or more independent variables (Saunders, Lewis, & Thornhill, 2019). It is used when the dependent variable is binary or dichotomous, meaning it can only take on two possible values (Saunders et al., 2019). Thus, the use of binary logistic regression was appropriate for the purpose of the analysis, since the responses to product offers were dichotomous variables with possible answers of “no” and “yes.”
The following models were assessed:
Response to product offer (01 – 03) = β0 + β1 * Age + β2 * Years of Education + β3 * Months with Service + β4 * Household Income + β5 * Gender + β6 * Number of Services Used + ∈
The third objective was determining which customers should be targeted for credit card promotion. It was assumed that the more items a person purchases using a credit card. A new variable was computed by adding the number of purchases on the primary credit card to the number of purchases on the secondary credit card. After that, a multiple linear regression model was created to evaluate the impact of demographic factors on the frequency of credit card purchases. The following model was assessed:
Number of purchaeseson credit cards = β0 + β1 * Age + β2 * Years of Education + β3 * Months with Service + β4 * Household Income + β5 * Gender + ∈
Results and Limitations
Most Valuable Customers
The sample analysis demonstrated that 49.1% of the customers of TeleMarket were males and 50.9% were females. The mean age of the customers was 46.1, with a standard deviation of 17.66. The mean number of years of education was 14.56, with a standard deviation of 3.29. The mean household income of the participants was $51,880, with a standard deviation of $50,110. The average number of months with TeleMarket was 36.42, with a standard deviation 22.34. Finally, the generated variable of the number of services used was 5.87, with a standard deviation of 2.98. Detailed descriptive statistics of the selected variables are provided in Table 1 below.
Table 1. Descriptive statistics.
Regression analysis was conducted to assess the impact of these variables on several services used. The results demonstrated that all of the variables were statistically significant, which implies that gender, age in years, household income, number of months with service, and years of education had a significant effect on the number of services used. The detailed analysis of coefficients is provided in Table 2 below.
Table 2. Regression model 1 coefficients.
While all the coefficients were statistically significant, the model’s predictive ability was very low. In particular, the determination coefficient was low, R2 = 0.148 (Adjusted R2 = 0.147). The model could justify only 14.8% of the fluctuations in the dependent variable.
The screenshots of the SPSS output for the analysis are given in Appendix A.
Factors Affecting Response to Promotion
The tested factors expected to affect the probability of responding to promotion included gender, age, years of education, household income, months with service, and quantity of services used. The dependent variables were Responses to Product Offer 01.. 03. The results are summarized in Table 3 below.
Table 3. Results of binary logistic regression analysis.
* – Coefficient is significant with p < 0.05
The results indicate that years of education and the number of services used significantly predict positive responses to all three product offers. However, despite two variables being statistically significant, all three models had very low predictive abilities. In particular, Nagelkerke R was 0.03 for Model 1, 0.012 for Model 2, and 0.35 for Model 3. This implies that the models could explain 3%, 1.2%, and 3.5% of changes in the dependent variable.
The screenshots of the SPSS output for the analysis are provided in Appendix B.
Factors Affecting the Number of Items on Credit Cards
Regression analysis was conducted to determine which factors affected the number of items on primary and secondary credit cards. The results of the analysis are summarized in Table 4 below.
Table 4. Regression model 2 coefficients
The results demonstrate age and household income were significant predictors of how many items the participants bought last month. However, the model’s predictive ability was very low, with the coefficient of determination being as low as R2 = 0.007 (Adjusted R2 = 0.006). This suggests that just 0.7% of the fluctuations in the dependent variable may be explained by the model. Appendix C contains the screenshots of the analysis’s SPSS output.
Limitations
While the study uses statistical methods, there are some possible limitations that must be acknowledged. The list of possible limitations is provided below.
Limited Scope
The study only focuses on the analysis of customer demographics and their response to promotions. Other factors influencing customer behavior, such as market trends, competition, and macroeconomic factors, are not considered.
Self-Reported Data
The data may be subject to self-report bias, where respondents may not provide accurate information. Additionally, the data may not capture all relevant information, such as customer behavior outside TeleMarket’s services.
Lack of Information on the Proposed Marketing Strategies
While the study analyzes customer characteristics and their response to promotions, it does not provide information on the feasibility or effectiveness of the proposed marketing strategies. Therefore, the study may not fully inform TeleMarket’s decision-making process regarding these strategies.
Business Implications
Descriptive statistics demonstrated that the number of males and females among TeleMarket’s customers is almost equal. The average customer is 45 years old, with 14 years of education and a household income of $37,000. The person has been with the company for 36 months and uses 5-6 services the company provides.
The first regression model analyzed what factors affect the number of services a customer uses. The results demonstrated that the coefficient for gender was negative, which implies that males are more likely to use increased number of services of TeleMarket. Similarly, the coefficient for age was negative, which implies that younger customers were more likely to use increased number of services. Since the coefficient years of education was positive, the customers with more years of education were more likely to use more services.
Household income had a significant positive correlation with the number of services used by the customers, meaning that higher household income was associated with increased customer service. Finally, the number of months with the service also positively impacted the number of services used, meaning that the longer the customer is with the company, the more services it uses. Therefore, it may be concluded that the most valuable customers are younger males who have spent more years in education, have a high household income, and have been using the services of TeleMarket for a long time.
The binary logistic analysis’s findings showed that the number of services utilized and years of schooling were important indicators of favorable reactions to each of the three product offers. This implies that the more years of education customers have, the more likely they respond to the offers. Similarly, the more services TeleMarket customers use, the more likely they will respond to the product offers. All other coefficients were non-significant. This implies that in the future, the company should target customers who use an increased number of services of the company and who have higher education. This may mean that customers who are more likely to accept promotions are those who are highly educated and can tell that the offer is beneficial for them. Additionally, this may mean that customers that use increased number of services enjoy them and believe that TeleMarket provides high-quality services.
The second regression model assessed which factors that affected the number of products bought using a credit card. The results demonstrated that household income and age had a significant impact on the dependent variable, which implies that older customers with higher household incomes were more likely to purchase items using credit cards. Thus, if the company decides to market a new credit card service, it should target older people with increased household income first.
References
McClaive, J., Benson, G. & Sincich, D. (2018). Statistics for business and economics. London, UK: Pearson.
Newbold, P., Carlson, W., & Thorne, B. (2022). Statistics for Business and Economics, Global Edition. London, UK: Pearson.
Saunders, M. N. K., Lewis, P., & Thornhill, A. (2019) Research methods for business students. Pearson.