Problem

An E-commerce company is facing intense competition in the market, making it difficult to retain existing customers. The company wants to develop a churn prediction model that can identify potential churners and provide segmented offers to prevent them from leaving. Losing one account could mean losing multiple customers, so account churn is a significant concern for the company.

The challenge is to develop a campaign suggestion that is unique and clear on the offer while being careful not to make a loss for the company. The recommendation will go through the revenue assurance team, who will reject any proposal that gives away too much free or subsidized stuff.

Learn more about our solution here.

Solution

We developed a churn prediction model using Python and presented our findings using PowerPoint. Our analysis included basic EDA, univariate and bivariate analysis, outlier treatment, missing value treatment, variable transformation, SMOTE, business insights, logistic regression model, LDA model, KNN model, Naïve Bayes Model, Bagging and Boosting- Ada.

Our unique campaign suggestion involved providing personalized offers based on customer behavior data. We identified potential churners by analyzing their past behavior patterns and used this information to create targeted campaigns with clear offers that would not result in losses for the company. By doing so we were able to provide valuable recommendations that were approved by the revenue assurance team.

Impact

Increase customer retention rate by X%
Reduce account churn by X%

Our solution had an incredible impact on the company's customer retention rate and account churn, resulting in significant revenue growth.

Team

Team Member #1 - Data Scientist
- Developed the churn prediction model using Python and performed data analysis.
- Provided valuable insights to create targeted campaigns based on customer behavior patterns.
Team Member #2 - Business Analyst
- Analyzed business requirements and provided recommendations for campaign offers that would not result in losses for the company.
- Liaised with the revenue assurance team to ensure that our recommendations were approved.

Churn Prediction Model Boosts Customer Retention