Your question is Handling Imbalance in Churn Models. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
You're training a churn classifier and only a small fraction of customers actually churn. A model that looks good on accuracy can still miss most of the at-risk users, so you need an approach that handles the skewed target properly.
How do you handle class imbalance in churn prediction?