Acquiring new customers costs more than maintaining current customers. When a current customer is lost, we lose future revenue and also the resources and investment that was spent to acquire that customer.
- Our goal is to create a logistic regression model that accurately predicts the potential churn of customers (true positives) and has minimum misidentifications (false positives).
- Once the customers who are most likely to churn are identified, a targeted approach can be used to address the factors that most influence churn.
A preliminary look at the data shows our overall churn rate is 14.5%. Customers with an International Plan have a much higher churn rate than the general population(42.4%).Those with a voicemail plan have a smaller churn rate than the general population (8.68%)
A model is only ever as good as the data that drives it. For that reason, it is essential to spend time investigating the properties of your data and whether any manipulation may be in order prior to beginning the modeling process.Complete Analysis is performed using R
There are so many ways one can access the performance of the logistic regression model. we have considered the ROC curve and AUC
In the ROC plot, you want your points on the curve to get closer to the northwest (0,1) for your model to be more accurate. The closer your points are to the diagonal line, the less accurate your model is. AUC is the area under the curve of ROC.
Focus on factors most influencing churn.Identify areas where we can make improvements that will make our customers want to stay.
