Case Study: Halfbrick achieves a 20% boost in 7-day retention with Firebase Predictions

A Firebase Case Study

Preview of the Halfbrick Case Study

Halfbrick uses Firebase Predictions to boost retention by 20%

Halfbrick Studios, the Brisbane-based developer behind hits like Fruit Ninja and Dan the Man, wanted to improve re-engagement by intervening before users churned but lacked the time and resources to build and maintain their own predictive models. They were interested in targeting users based on predicted behavior rather than historic heuristics to improve retention.

Using Firebase Predictions with their existing Remote Config in Dan the Man, Halfbrick ran a 3‑variant experiment (control, heuristic gift after beating level 3, and gift for users predicted to churn). Serving the gift to predicted churners raised 7‑day retention from 25.35% to 30.24% — a 5 percentage‑point, 20% relative boost — and the promotion (2000 gold coins + 25 gacha tokens) was rolled out to all users identified as “will churn.”


Open case study document...

Halfbrick

Miguel Pastor

Developer


Firebase

43 Case Studies