Game Analytics 101 Part 2: Basic Definitions

This post is part two of a five-part series on analytics in the gaming industry.

The first goals are typically to understand “how are we doing?” and “are we getting better or worse?” The simple answers come from looking at your--pardon the jargon, but this is really what people say--KPIs, or Key Performance Indicators. These are basic metrics that you should be tracking over time. It’s good to know their values today, but you want to see that compared to yesterday, last week, last month, etc. Over-time graphs or tables with day-over-day/period-over-period are especially helpful.

These are the most common, along with a definition:

DAU (Daily Active Users)  -- the number of users who engage with the game on a daily basis. What does engage mean? Usually at least one logged action such as logging in. You also have to define what “day” means. It’s 24 hours, but which ones? I suggest using GMT given the increasingly global nature of the business.

MAU (Monthly Active Users) -- the number of users who engage with the game on a monthly basis. MAU is a useful metric because it often ties into financial reporting and business models. Many games have shifted from monthly subscriptions, but CFOs still use monthly as a timeframe for financial reporting.

DAU/MAU -- Daily Active Users divided by Monthly Active Users. This is one of the few basic statistics that starts to give you a sense of proportion. Imagine you have 100k MAU. If your DAU yesterday was 10k, then you had about 1 in 10 of your regulars online. If your DAU was 50k, you had about 1 out of 2 of them online (which would be unusually high). This metric gives you a small sense of the intensity and repetition of use.

Concurrency -- how many players are online not just during the day but at any one time. That obviously impacts operations since a big spike of traffic may be great news to the marketing department until they learn that it crashed the server and created ill will and lost potential revenue.

Retention -- how many players come back within a given timeframe. Developers can look at retention on a daily basis, weekly basis, monthly basis, etc. Typically, I see 1-day, 2-day, 3-day, 7-day, 28 or 30-day, and then 3, 6 and 12 months.

K-Factor -- A very basic stat to show how viral your game is (the exact formula is number of invitations sent multiplied by the percentage success rate). Over 1 and your game is generally growing through social channels.

ARPU (Average Revenue Per User) -- the spending across your player base, including users who pay and those who don’t. If the game is free, this is generally small, especially because a small fraction of players typically pay, and the larger pool of free-riders pull the average down. That makes ARPU as somewhat dangerous number to rely on. ARPU can be reported daily, weekly, etc. Monthly is most common, but daily views can show interesting patterns like spikes on weekends, etc.

ARPPU (Average Revenue Per Paying User) measures how much players who pay spend, on average. It’s going to be higher than ARPU, often many times. Like ARPU, ARPPU is usually viewed monthly, but can be used in any time frame.

Average Session Length -- the amount of time players spend in a game, on average. Short bursts or long engagement? And does your title auto-log out after some period of inactivity? This can be a useful metric so long as you have that clearly defined.

These are just a few KPIs, but should give you a picture of what you can measure from player data. From KPIs, developers can then wade into deeper waters and more advanced predictive analytics, which include lifetime total value (how much a player is expected to spend in-game before they churn out) and social value (which measures player worth in terms of how they influence others in the game).

The next installment will focus on the social interactions of players--why you should care, how you can use it, etc.