Design a customer churn prediction system
PremiumML systemEasy
- Churn prediction
- Label window
- Label leakage
Quick answer
This ML system interview question asks you to design a customer churn prediction system. Aim for a design that turns data into useful predictions in production, then explain the data, training, serving and monitoring. Interviewers want to see how you frame the problem, measure success and keep the model healthy after launch.
The problem
Design a customer churn prediction system for a subscription music streaming service like Spotify: a model that ranks paying subscribers by their risk of cancelling in the next 30 days so the retention team can send offers. The hard parts are defining churn without leaking the future and proving the offers change behaviour.
Every week the model scores all subscribers. The retention team has a budget for a fixed number of offers per month, such as one month at half price. An offer sent to someone who would have stayed anyway wastes money; an offer sent too late reaches someone who has already left.
What must a customer churn prediction system do?
- Weekly risk score: Score every paying subscriber each Monday with a probability of cancelling in the next 30 days.
- Targeting list: Hand the retention team the top subscribers by risk, within a budget of 200,000 offers per month.
How well must it do it?
- Honest labels: No feature may use information from after the scoring date.
- Calibrated scores: A score of 0.2 should mean about 20% of such users cancel, so finance can forecast churn.
What scale should you design for?
Use these numbers in your estimates.
The scale numbers unlock with this questionUnlock
Key concepts
- Churn prediction
- Label window
- Label leakage
- Uplift modeling
- Precision at k
- Batch scoring
How you’ll practise
- Problem framing
- Data and labels
- Features and model
- Training and evaluation
- Serving and monitoring
- Wrap-up
Lala, personalized to you, gives a hint when you ask and reviews your design when you are ready.
What do interviewers look for?
In any ML system design round.
- Frames the product goal as a clear ML problem
- Defines offline and online success metrics
- Explains where the data and labels come from
- Separates training from serving and keeps features consistent
- Plans for drift, retraining and rollbacks
- Weighs model quality against latency and cost
Questions people ask about this interview question
- Which design type is this question?
- It is an ML system question. You design a customer churn prediction system and explain your choices the way you would in an interview.
- Is this question asked at entry level?
- It is marked Easy, so it suits engineers who are new to ML system interviews. It is also a good warm-up before harder questions.
- What should I cover first?
- Start with the product goal, how success is measured and what data is available. Agree on scope with the interviewer before you design anything.
- Can I practise it for free?
- It is a premium question. Sign up and unlock it once for 10 credits; it then stays unlocked. When you finish, Lala, your personalized AI mentor, can review your design for 30 credits.
Related questions
Updated · Written by the SevyDevy team