Customer lifetime value app survival analysis


US-based media conglomerate with interests primarily in cinema and cable television.


The client had a pressing need to minimise churn rate of customers on the app and identify pain points. The client aimed at releasing the right content at the right time, while considering customer behavior attributes and video engagement metrics. The client was also facing issues with manual intervention in the data capture process.


Using our experience and deep understanding, we implemented a three-step approach to the above problem:
  • SG Analytics collected clickstream data to understand user behavior and performed basic EDAs to gauge the app’s performance.
  • We employed the Kaplan–Meier estimator to provide descriptive statistics and study the effect of different attributes on the survival of the customers. The key predictors used included behavioral attributes such as visits, return frequencies, etc. used in a semi-parametric Cox model in R.
  • Our model helped identify how the effect of different predictors combined together affects the survival of different customers. Based on the effect of these predictors, the team was able to predict the lifespan of a new/existing active customer.


Helped identify optimum length of trial period, cadence for video release and non performing shows.
Improved customer retention rate by 25% leading to significant revenue growth.
Provided for widespread adoption of the model for use across other business divisions to minimise customer churn.