Modeling Game Monetization
This is not a major feature release of my EUREKA model for social game virality but it does include some important enhancements for financial forecasting. First and foremost the calculation of a users Average Lifetime Value should now be correct and fully automated. (Getting it right was easy, getting it automated was a big PITA) This new version also includes a cohort slider. In this case I have made it possible to examine the behavior of each new group of “Engaged Users” on a day by day basis. If you move the cohort slider, the cohort related statistics under the graph will update themselves to reflect that batch of users who become “Engaged” with the game on that day in the games forecast life-cycle. Recall from my earlier articles on the subject that an “Engaged User” is defined to be a user who is frequently playing the game, most often it represents users who have transitioned from free trial players to paying users. In this model the $/DAU field can be set with a slider and the model will calculate the games lifetime revenue potential, the Average User Lifetime, and the Average Users Lifetime Value.
Link to a new release of EUREKA featuring cohort analysis of “Engaged Users” and average user LTV calculations.
The dashed purple line in the lower left hand corner of the graph shows the life-cycle of the cohort of users who become “Engaged Users” at the time set by the cohort slider. Technically this graph should start and decay at the time set by the cohort slider, but I normalized it to zero because it is usually such a small area compared to the overall game life-cycle graph that it makes it easier to zoom in on using the Scale slider if it stays pinned to the left side of the graph.
In the example pictured above, on Day ZERO, the game being modeled acquires 95.16 new “Engaged Users”. With an Average Lifetime Value of $100, these 96.16 users will generate $9616.00 dollars in commerce revenue over their life-cycle. If you move the cohort slider to the right, you will see the number of daily new users drop as the game chews through it’s potential audience over time, until it has finally churned through its entire potential user base by roughly the 400th day.
One of the major challenges with applying EUREKA to real world analytics is that most users and analytics systems see data in its raw linear sampled from, whereas EUREKA is a set of first order differential equations that represent gameplay in terms of rates of change between states. To really be practical for business modeling and financial forecasting I need to provide hooks into the model that allow users to enter analytics data as it is typically captured and compute the virality parameters for EUREKA automatically for them. This version of EUREKA accomplishes this transformation internally for Total Revenue Potential, Average LTV, Cohort Revenue and Cohort Size. With these calculations in place it should now be possible to build more sophisticated financial forecasting and optimization features into subsequent versions.
If you have live game data that you would like to be able to analyse with EUREKA feel free to send me a sample of it and any supporting explanation of what the data represents. It will help me develop better ideas on how to rig EUREKA to be much more practical for real world applications. I am presently considering including a second parallel graph with EUREKA that transforms the entire model into a financial forecast space in hopes that it will help people gain a better intuitive understanding of the intimate relationship between organic virality and monetization.
Leave a Reply