I am very close to launching a test version of my mobile game, Froogies At this point I would like to add stats mechanism that will allow me to adjust and improve the user experience and the game's "stickiness" and "virality".
As I started to plan my stats mechanism, two fundamental questions came up:
1. Stats Logical Structure
In my game the player has to make the "Froog" jump and eat "vitamins". As the game goes along:
a. the score goes up upon collecting vitamins.
b. the "froog" might die (3 lives per game, and you may collect hearts that will get you additional life)
c. new levels are opening up (with new types of vitamins)
The questions that I'm trying to answer:
1. How many times will the average user launch my game before leaving it behind for good, and over what period of time? (Retention data)
2. What's the average session length?
3. How many games are played over an average session?
4. What's the average score on the 1st/2nd/3rd/10th game? What's the average level achieved on the 1st/2nd/3rd/10th game? (Progression/Time graph)
5. How many lives are actually played on an average game?
6. What's the average number of "special vitamins" that are collected on an average game?
7. If the Share button is clicked - after how many sessions/games did it happen? what's the user's highscore at this point?
What should a proper set of game stats include? On one hand, I want to know as much as I can about many different aspects of the game usage. On the other one, I don't want tons of data that is too complicated to analyze. Does anyone have a good example of a set of game stats that is proved useful? Is anything of importance missing on my list?
2. Stats Technical Structure
I read the related (limited) materials I found on Scirra's site and the conclusion is that I'm going to use Google Analytics plugin by Sun Temple.
I'm expecting that most of my users will be kids 6-12 y.o. Kids at young age don't have their device online at many times (since they mostly rely on available wi-fi). This means that in many cases the data will be lost (if base on realtime mechanism only).
I was thinking about a mechanism that stores the stats to a local webstorage when the browser object is not online, and sends the data to GA when online next time. The problem is that with such mechanism I will be getting two sets of data ("Realtime Data" and "Delayed Data") and these are going to be quite complicated to combine.
Did anyone experience a similar mechanism? Any other ideas how to work around the "delayed data" issue?
I would love to hear any comments and ideas regarding Game Statistics and Analytics, and turn this thread into a viable source on the matter. Proper implementation and utilization of Game Statistics is a key factor to success. unfortunately, most of the theory behind it is kept by the large gaming companies, and not available to Indie game developers. Lets change this!