(free pic under CC)
1. Rolling Retention
In my short article on "Healthy Retention Rates", I primarily focused on rolling retention. This is: Take a look at a cohort and track their life time. Averaging this figure gives you the average life time.
2. DAU Frequency Retention
By that I mean a methodology which takes a snapshot from one specific day and takes a look at that daily cohort by measuring how many of those that have been active today have been active on a daily basis over the past 7 days. This also gives you an indication of recency. I have seen top mobile games achieve a 70% DAU Frequency retention. In other words, of that daily logged in users, 70% had logged each day for the past 5 out of 7 days! Of course, the DAU basis should be reduced by the number of new daily logins, so that you really have only loyal DAU.
3. Predictive LTV and Milestone Tracking
Define Milestones which could give you an indication as to whether your mobile game will succeed or not.
These two findings are from Tapjoys research on "predicting the future LTV of your your users."
- "Reaching a critical point of 1,000 users who make > 3 purchases is a good indication that an app will ultimately top $1MM in revenue. 84% of the apps with 1,000 or more users who completed three or more in-app purchases within the first 90 days broke that $1MM threshold.
- 35% conversion rate from 1st to 3rd purchase was the critical number for breaking the $1MM revenue threshold."
I previously mentioned that D1 (40%), D7 (20%) and D30 (10%) is a good Western benchmark for midcore games. In my experience, East-Asians take a a different approach, which is slightly more short term on this. I remember one Japanese executive mention the following retention rates as successful.
D1 > 50%
D3 25-30% !
D7 ~ 20%
In his talk, he stressed D3 which seems to empirically work for him to predict if a game is successful or not.
Anyway, these are just preliminary thoughts. Please comment below for further exchange.