GermanyMedia and entertainment2020 – ongoingKolibri Games
Machine learning model to predict player spending and lifetime value
Kolibri Games raised player lifetime value by 30% with a spending-prediction model
Problem
Kolibri Games, the Berlin maker of Idle Miner Tycoon and Idle Factory Tycoon, wanted to calculate and raise the lifetime value of its players while staying player-led. Its data team had a two-week release cycle, and engineers and analysts could only share data locally over Slack. The games had more than 10 million monthly active players.
Approach
The team moved game-play data processing to Databricks on Azure and built a machine learning model that predicts how much a player will spend. MLflow shortened the path from model to production. Algorithms now also target advertising at player segments, such as those who do or do not use ads or in-app purchases.
Results
Takeaway
Kolibri tied a spending-prediction model to marketing and in-game decisions. The company did not say how it measured the 30% gain.
Categories
- Industry
- Media and entertainment
- Business function
- Marketing and sales
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and ML
- Adoption
- Off-the-shelf tool or platform
- Evidence
- Vendor case study
Sources
Source summaries
Databricks reports that Kolibri Games built an ML model to predict player spending and lifetime value, and moved its data processing to Databricks on Azure. It gives a 30% rise in lifetime value per player and a 3x gain in data team productivity estimated by the CTO. It also says cost per acquisition and cost per install stayed efficient.
Full details are in the original sources.
