Rebecca Castellani

In this MGS Europe 19 session, Martin Jelinek, Head of Marketing for AppAgent, delivers an informative and insightful overview of lifetime value (LTV) predictions.


Jelinek introduces LTV by discussing the nature of prediction and the use of historical data to determine a set of rules. He explains how these rules can create a predictive model that enables additional inputs to become future conclusions and highlights the three most common models:

  • retention x ARPDAU
  • spending ratio
  • behavior

Jelinek notes that predictive outcomes are always a percent chance or range, but can still be used to improve user-acquisition rates or re-engage existing users. Jelinek then gives some practical examples of how LTV models are applied to subscription apps, in-app purchases, and e-commerce apps. After some thoughts on how a team should drive the LTV prediction process, Jelinek summarizes the session with key takeaways.