Whilst many mobile game developers are finding it difficult to monetise their apps, mobile game publisher Deemedya has spent 18 months to develop a solution. Called Game of Whales, the app monetisation tool combines artificial intelligence and machine learning to learn player behaviours and predict the in-app offers that are most likely to work for an individual gamer.
Deemedya tested the solution during its trial period and found that revenues increased by as much as 25%. The platform is automated, which means that once the SDK is integrated as part of a game, a quick learning phase is initiated and then the artificial intelligence takes over.
Deemedya named the solution Game of Whales after a nickname given to high-spending mobile game players – Whales. These are the players most mobile app developers are trying to reach.
Doron Kagan, CEO of Game of Whales explained:
“As a game publisher ourselves, we are only too familiar with the ongoing stress and complexity of keeping people playing, and getting those players to spend something – anything! – in our game. Game of Whales happened because, frankly, there was nothing out there that did the things we wanted.”
The platform simply micro-segments users into smaller groups based on usage and behaviourr. It then calculates when and how certain incentives should be offered in order to increase purchases. The predictive machine learning component is able to predict which players are more likely to spend.
Results from initial testing also found that player churn jumped 10%, whilst engagement increased 9% and conversations of non-paying to paying players was over 15%.
Meanwhile, Deemedya noticed that Whale gamers which had spent more than $300 playing a certain game, were keen to hang around and spend more.
“Making successful games should be about fun and creativity, not spreadsheets and data analytics. So we thought, why not build something that takes care of all that? That’s where the AI approach came from. The algorithms can analyse and make decisions faster and better than we ever could, and the proof is in the results it’s generated. You literally plug it in to your game, and leave it alone,” Kagan added.