COMM1190 Data, Insights and Decisions
Assessment 1: Initial report
TelcomCo churn rate project
The General Manager (GM) at TelcomCo is mandated to deliver a customer retention update to the Board of Directors. To prepare, the GM initiated a pilot study led by a junior analyst using sample data on customer service usage patterns.
The junior analyst’s initial findings, documented in a memo (Appendix A), were based on pilot data from July 2024. As a Business Analyst at TelcomCo, you have been asked to conduct a deeper analysis using an expanded dataset to investigate churn rate factors and offer actionable recommendations for improving customer retention. Additionally, the GM raised concerns about the memo’s quality, requiring a revised report.
Appendix A: Analysis of customer pilot data
MEMORANDUM
DATE: 12 September 2024
TO: Board of Directors
RE: Marketing Analytics Report
Introduction
This report presents an initial statistical analysis of customer data from the company’s central database. The dataset includes information on customer demographics, account types, service usage, churn rate rates, and spending patterns. The analysis focuses on pilot data from July 2024 and is structured into four sections: customer demographics, monthly charges, customer satisfaction, and recommendations.
Customer Characteristics
Figure 1 and Table 1 summarise key customer characteristics from the pilot sample, including demographic information and service usage insights.
Figure 1: Customer Characteristics
The gender distribution among customers is evenly split. The average age of the customers is 46.64 years, with an average tenure of 32.25 months. Approximately 75% of customers do not have dependents, and over 60% are not living with a partner.
Figure 2: Age
According to Figure 2, the customer age distribution peaks in the 19-25 age group, with about 15% of customers being senior citizens. The distribution is not symmetrical, indicating a concentration in younger age groups.
Table 1: Customer Characteristics
|
Statistic
|
Tenure
|
Age
|
|
Mean
|
32.25
|
46.64
|
|
Median
|
29
|
46
|
|
Mode
|
1
|
21
|
|
Standard Deviation
|
24.84
|
19.20
|
|
Skewness
|
0.26
|
0.38
|
|
Range
|
72
|
71
|
|
Minimum
|
0
|
19
|
|
Maximum
|
72
|
90
|
After analysing customer characteristics, we explore the financial impact through transaction sales data.
Transaction sales
The monthly charges represent the transaction sales, indicating the amount customers pay for the company’s telecommunications services. Table 2 shows summary statistics for these charges. Based on a sample of 999 customers, the average spending is $66.48, while the median is $74.25, suggesting a negatively skewed distribution. A mode of $19.90 indicates a significant portion of customers pay low charges, possibly reflecting a large group of customers subscribing to lower- tier service packages.
Table 2: Transaction Sales
|
Statistic
|
Monthly Charges
|
|
Mean
|
66.48
|
|
Standard Error
|
0.95
|
|
Median
|
74.25
|
|
Mode
|
19.90
|
|
Standard Deviation
|
29.93
|
|
Sample Variance
|
896.07
|
|
Kurtosis
|
-1.20
|
|
Skewness
|
-0.30
|
|
Range
|
97.30
|
|
Minimum
|
18.95
|
|
Maximum
|
116.25
|
|
Sum
|
66,413.50
|
|
Count
|
999
|
A correlation analysis was conducted, and the results are shown in Table 3. There is a moderate positive correlation between tenure and monthly charges, indicating that customers who stay longer tend to spend more. However, tenure has a weak negative correlation with age, suggesting that age and tenure are not closely related.
Table 3: Correlation Matrix
|
Variable
|
Age
|
Monthly Charges
|
Tenure
|
|
Age
|
1
|
0.176
|
-0.028
|
|
Monthly Charges
|
0.176
|
1
|
0.252
|
|
Tenure
|
-0.028
|
0.252
|
1
|
Customer churn rate
The overall churn rate is 25%, as depicted in Figure 3. Figure 4 shows that female customers are slightly more likely to churn rate than male customers. Additionally, Figure 5 reveals that customers not living with a partner are more prone to churn rate, as illustrated by the percent stacked bar chart.
Figure 3: Churn rate
Figure 4: Churn rate by Gender
Figure 5: Churn rate by Living Status
Conclusion
• A high proportion of customers are in the youngest age bracket, contributing to the lower mean spending than the median.
• Female customers and those not living with a partner show a higher propensity to churn rate.
• A positive correlation exists between tenure and monthly spending, indicating that customers who stay longer tend to spend more. TelcomCo can leverage this information to reduce churn rate (e.g., loyalty programs or incentives for long-term customers).
Recommendation
The small pilot dataset limits the reliability of these insights. More customer data should be collected to generate a more robust analysis, and additional variables should be explored.
To enhance the reliability of the analysis, TelcomCo should expand the dataset by collecting additional customer data, such as customer engagement metrics, service quality feedback, or payment history. This would enable more comprehensive insights into churn behaviour. Additionally, future analyses should explore new variables like customer satisfaction overtime and the impact of promotional offers. A more extensive analysis could involve segmenting customers by service package or region to tailor retention strategies effectively.
Appendix B: Data dictionary
|
Variable
|
Description
|
|
CustomerID
|
Unique identifier for each customer.
|
|
Gender
|
Gender of the customer (Male, Female).
|
|
Age
|
Age of the customer in years (18+).
|
|
Partner
|
Whether the customer lives with a partner (Yes, No).
|
|
Dependents
|
Whether the customer has dependents (Yes, No).
|
|
Tenure
|
Number of months the customer has stayed with the company. For analytical purpose, the tenure can be classified into three broad categories: short-term (<9 months); medium term (between 9 and 18 months) and long-term (greater than 18 months)
|
|
InternetService
|
Type of internet service (Fiber optic, 5G, DSL, No service).
|
|
OnlineProtect
|
Level of online security/back-up (0: None, 1: Security, 2: Backup, 3: Both).
|
|
TechSupport
|
Type of tech support (1: Chatbot, 2: Email, 3: Phone).
|
|
Streaming
|
Streaming subscription status (0: None, 1: TV, 2: Movies, 3: Both).
|
|
Outage
|
Frequency of service outages (Occasional, Frequent).
|
|
ContractType
|
Type of contract (1: Month-to-month, 12: One year, 24: Two year).
|
|
MonthlyCharges
|
Monthly amount charged to the customer.
|
|
MultipleLines
|
Whether the customer has subscribed to more than one phone line (0: no phone line; 1: one phone line and 2: more than one line)
|
|
Churn
|
Whether the customer churns (Yes, No).
|
|
Pilot
|
Indicates whether data is from the pilot study (1: Yes, 0: No).
|