Game Development Reference
rest. But careful study and regular iteration based on metric data can act as a
force multiplier, transforming your “good” game into a huge hit.
The relevance of some of these metrics is obvious; you already understand
why you care about the ARPU, for example. The relevance of other types of
data may be less obvious. For example, why do you really care about the cost to
acquire a new user? Won't existing on a social network automatically make your
game “go viral”? The answer, sadly, is “not any more.” The marketplace has been
sufficiently flooded with products, many of which are of low quality or alienate
users by spamming their friends, or worse, by compromising sensitive user data.
This situation has led to a noisy market in which gathering new users is difficult.
Typically, you will be forced to rely on advertising to reach an initial critical mass
of players. The number of users you can attract tends to be directly related to the
amount of advertising sugar you have to spread around. But because these users
don't actually pay you unless they decide they love your game, the only way to
really know how big a bet you should place is to evaluate the LTNV of your users
against your UAC. At the simplest level, the formula you care about is:
Using a formula like this, you can determine which games deserve the big
bets. A high LTV justifies spending marketing dollars to acquire new customers.
It represents the net present value of your customer base (and in some ways,
when viewed as an aggregate of all of your users, reflects the total current value
of your company, currently).
When considering your metrics, here are a few things to keep in mind:
l It's easy to track and gather metrics, but it's famously difficult to interpret
them. So be sure to become someone with—or hire people with—a superb
mind for analytics and understanding statistical data.
l Your game metric gathering should be structured to allow for easy experi-
mentation so that you can easily test out new theories and gather metrics on
new features and content.
l The goal of statistical analysis isn't to prove that your game is good. The
goal should be to show you which levers to pull in order to make your good
game even better.
l Focus on metrics that allow you to test two different scenarios (A/B) and
determine what effect each has on a particular metric. This focus leads
immediately to actionable information. As a simple example, if you're plan-
ning an email reminder to your customers to try to bring them back for
a new session, and your mail will offer them a free gift if they return, try
sending out two versions of the mail to 10 percent of your user base. In one
version, you can offer them two free in-game items. In another version, you
can offer them one in-game item and let them pick a friend to gift as well.