Problem Description:
TeleMarket, a company poised for growth and diversification, is considering expanding its services to include Excel assignment assistance. This report focuses on leveraging data analytics strategies to address a critical research question: which customers should TeleMarket target to help with Excel assignment effectively? Providing support for Excel assignments, a crucial aspect of academic success can substantially contribute to TeleMarket's overall growth. The report delves into the evolving landscape of online Excel assignment assistance, emphasizing the importance of precise customer targeting and tailored support in the digital age.
Methodology:
Data Collection: A secondary historical sampled dataset, consisting of responses from 4500 TeleMarket customers, was employed. Key factors considered included job category, employment years, income, churn, default history, and debt-to-income ratio.
Modeling Techniques: K-Means clustering and logistic regression were selected to identify distinct customer segments and predict loan default likelihood.
- K-Means Clustering:
- K-Means divided customers into groups based on shared traits.
- Cluster analysis revealed two main clusters, offering insights into customer distinctions.
- Logistic Regression:
- Logistic regression analyzed factors affecting loan default.
- Identified variables included job category, employment years, debt-to-income ratio, income, and churn.
Results and Limitations:
Descriptive Statistics:
The table below shows descriptive statistics for the variables of the study.
| Variable | Mean | SD | |
|---|---|---|---|
| Employment Years | 9.14 | 9.3 | |
| Household income | 51.88 | 50.11 | |
| Debt to Income ratio | 9.93 | 6.4 | |
| Variable | Frequency | Percentage | |
| Income Category | Under $25 | 1253 | 27.8% |
| $25 - $49 | 1652 | 36.7% | |
| $50 - $74 | 730 | 16.2% | |
| $75 - $124 | 570 | 12.7% | |
| $125+ | 295 | 6.6% | |
| Job Category | Managerial and Professional | 1253 | 27.8% |
| Sales and Office | 1507 | 33.5% | |
| Service | 563 | 12.5% | |
| Agricultural and Natural Resources | 193 | 4.3% | |
| Precision Production, Craft, Repair | 387 | 8.6% | |
| Operation, Fabrication, General Labour | 597 | 13.3% | |
| Churn | No | 3326 | 73.9% |
| Yes | 1174 | 22.1% | |
| Defaulted Loan | No | 3407 | 75.7% |
| Yes | 1093 | 24.3% | |
| N = 4500 | |||
Table 0.1: Descriptive statistics on different variables
- The analysis provided insights into customer characteristics, revealing patterns in employment, income, and customer behaviour.
K-Means Clustering:
- Two clusters were identified, demonstrating the diversity of TeleMarket's customer base.
- Some variables, such as debt-to-income ratio and loan default history, showed no significant impact on cluster differentiation.
Logistic Regression Analysis:
- The logistic regression model highlighted significant predictors of loan default.
- Job category and income categories 3 and 4 were less impactful, with other factors proving to be strong predictors.
Limitations:
- Potential sampling errors in the secondary dataset may introduce biases.
- The study relies on observational data, subject to biases and limitations inherent in retrospective analyses.
Business Implication:
Recommendations:
- TeleMarket should prioritize customers with a history of not defaulting on loans for their digital lending program.
- Factors such as job category, employment years, income, churn, and debt-to-income ratio should guide loan consideration.
Considerations:
- Careful credit risk analysis remains crucial for successful loan approval.
- Targeting average-income customers is advisable to maximize profits and ensure sustainable growth.
In conclusion, this analysis equips TeleMarket with actionable insights, guiding its foray into the dynamic landscape of digital lending. By understanding customer segments and predictive factors, TeleMarket can make informed decisions, mitigating risks and ensuring a successful transition into the loan business.