Tableau: Analyzing Customer Churn

Background

This case study is from Carl Rosseel’s course on DataCamp. It analyzes fictitious data from a US-based telecom provider called Databel. Each row of the dataset represents an individual customer at Databel with 29 informative columns. More details on these fields can be found here. This data was collected from a report of the company database at the end of their most recent fiscal quarter. Customer churn occurs when a customer ends their subscription to Databel services. Churn rate refers to the percentage of churned customers amongst total customers that subscribed to Databel services this past quarter.

Databel needs help identifying what is driving customer churn at their company. This project allowed me to explore data visualization in Tableau while discovering data-driven insights!

Analysis

The strategy team at Databel wanted to take a closer look at certain demographics in their customer base. The company recognizes a fine balance between service profitability and what they charge their customers. Databel has asked for some new ideas regarding where to modify costs in their data plans to reduce customer churn. The storyboard below contains four interactive tableau dashboards. Each dashboard was designed to answer the following questions.

What are the most common reasons for churning?

Navigate to the Overview tab in the storyboard. Looking at the Churn by Category pie chart, 44.82% of churned customers left Databel last quarter due to competitor-related reasons. This is reflected in the Churn Reasons chart, where we can see that 2 of the top 3 churn reasons are in fact related to competitors. We see that 16.87% of churned customers left because a competitor made them a better offer. Similarly, 16.54% of churned customers left because a competitor had better devices.

What is the national churn rate for Databel?

In the Overview tab, we can see that the national churn rate for Databel in the previous quarter was 26.86%.

Which state has the highest churn rate?

Looking at the Churn by State map in the Overview tab, California had the highest churn rate out of all US states at 63.24%. This is quite high! Click on California in the Churn by State map and observe how the Churn Reasons, Churn by Category, and Customer Contract visuals change. Note that the majority of these churned customers did not provide a reason for churning. Databel could lower the California churn rate by conducting further research – perhaps through surveying – to discover the needs and frustrations of their California demographic.

How does contract type affect the reason for churning?

The Customer Contracts pie chart in the Overview tab reveals that 51.01% of Databel’s customers had month-to-month contracts last quarter. We can click on this pie slice to filter the data for these customers. Month-to-month contract customers had a 46.29% churn rate! Of these churned customers, 45.41% of them left Databel for a competitor. One might expect lower-commitment contracts to have a greater turnover rate. This customer demographic should be a primary focus for Databel when modifying their customer policies.

Do average monthly charges differ by customer age?

Navigate to the Age Brackets & Groups tab in the storyboard. In the purple Pick Metric filter along the top of the dashboard, select Avg Monthly Charge. In the Age of Customer area chart, we can see that average monthly charge (Metric) stays relatively consistent amongst all customer ages below 65. In comparison, customers aged 65+ are charged almost $10 more every month. To reduce churn rate in their elderly customers, Databel might want to consider offering a discount or senior plan to this demographic.

How do group plans impact average monthly charges?

Currently, Databel offers a discount on group plans of 2 people or more. Group plans are incentivized by offering lower monthly charges per individual. Databel hopes this deal will draw more customers in to purchase their services. If we look at the Group Plan area chart in the Age Brackets & Groups tab, we can see that average monthly charge and churn rate decrease significantly for customers in a group. Observe that the average monthly charge per customer in any sized group is fairly uniform. We can also see a slight increase in churn rate amongst groups with 5 people. Databel might want to consider lowering average monthly charges per individual in a more linear pattern as the size of a group increases. This would further reward customers for adding more members to their group plan. In turn, this could boost customer satisfaction and decrease customer churn for this demographic.

Is there any correlation between average customer service calls and churn rate?

In the Age Brackets & Groups tab, set the purple Pick Metric filter to Avg Customer Service Calls. In the Age of Customer area chart, average customer service calls and churn rate share a positive correlation. It is worth mentioning that peaks in both average service calls and churn rate for customers aged 70+ are slightly inflated since there are less senior customers.

Explore filters! How many customers aged 30-35 live in the western region of the US and have an international data plan?

In the Age Brackets & Groups tab, set the purple Pick Metric filter to Number of Customers. To filter by western states in the US, use the yellow State filter to select California and Nevada (CA and NV) from the dropdown list. Next, find the yellow International Plan filter and select Yes. We can now look at the Age of Customer area chart to answer this question. Note that all age values along the x-axis represent 5-year age bins. For instance, the age value 20 represents all ages between 20-24 (inclusive). Thus, we find that there are exactly 3 customers in the demographic posed by this question.

How does the international plan influence churn?

Navigate to the Data Usage & Plans tab in the storyboard. Yellow filters along the top of this dashboard can be used to filter demographics. If we look at the International Activity table, we can determine which customers are internationally active and which customers purchased the international data plan. The appropriate churn rate is displayed in each table entry. We can see that customers who are not internationally active and purchased the international plan have a churn rate of 71.19% (additional details can be found by hovering your cursor over each table entry). Conversely, customers who are frequent international travelers that do not have the international data plan have a churn rate of 40.34%. Both of these results seem intuitive. Customers may be more inclined to churn if they are spending money on a plan they do not need. Similarly, traveling customers without this data plan may be more inclined to churn provided they are receiving high international data fees. To help reduce churn in these two demographics, Databel can improve their marketing strategy around the international data plan. Finding a better way to inform customers about this plan could reduce unwanted fees and improve customer satisfaction.

How might international plans and contract-types influence international charges?

In the Data Usage & Plans tab, locate the yellow Contract Type filter. Click through each filter and observe the change in Extra International Charges (on average). Lower-commitment contracts tend to have lower international fees on average. Databel might want to consider offering a discount on international plans for customers that subscribe to longer contracts. This could increase sales on international plans while reducing customer frustration induced by unwanted international charges.

What relationships do contract type and payment method share?

Navigate to the Contract Type & Payment Method tab in the storyboard. Yellow filters along the top of this dashboard can be used to filter demographics. We can see that month-to-month contracts have the highest churn rate in the Contract Type bar chart at 46.29%. If we click on the month-to-month bar to filter by these contracts, we find that customers in this demographic using paper check payments have the highest churn rate of all payment methods at 57.33%. This can be slightly expected. The lack of customer commitment in month-to-month contracts allows customers more freedom to discontinue their service purchases at Databel. Customers using checks as their primary payment method could be more inclined to miss payments since their monthly charges are not automatic. Databel might consider introducing a mercy-period policy for late check payments to improve customer satisfaction for this demographic.

Skills Applied

This case study allowed me to further explore calculated fields, level-of-detail expressions, dynamic zone visibility, and parameterization by analyzing the Databel dataset from scratch. I discovered how variables like Location, Age, and Contract Plan influenced churn rate using effective data visualization. I then used these findings to provide Databel with suggestions to reduce churn rate and improve customer satisfaction.