I’ve just read a very interesting article, Diffusion dynamics of games on online social networks, from the Usenix conference Workshop on Online Social Networks WOSN 10. (If you want to read it, too, there’s a .pdf of it here. ) The paper, by Wei, Yang, Adamic, de Araújo and Rehki, looks at game invitation behaviour and outcomes in two different Facebook social games, Yakuza Lords (YL), and Diva Life (DL), both from developer LoLApps. There is a lot of interesting detail in this work, which I hope to talk more about in future, but the biggest fattest screamingest red-top headline is this:
- users who were recruited by friends played longer
For both games, recruitment by a friend was not the way that most people started to play the game (an invitation from a friend preceded download for only 37% of new YL players, and 25% for new DL players).
However those users who were recruited by friends became more engaged. For example, 80% of non-invited new players left the game within the first day, but over 50% of invited new players played for more than a day. (Mind you, these are both pretty big loss ratios, but the difference between them is worth paying attention to. )
There are lots of possible reasons why friends make better recruits. The authors point out several of them. My own pet hypothesis, which is sitting here wagging its tail at me, is that, for most people, having people you know ready to play with you with when you first join up makes games more attractive.
If this is really the case, then there should be an effect on engagement of the size of in-game network you join up to – unless the most important effect is a threshold effect at ‘1’. (Slide rules at dawn, anyone??) And what’s even more promising, commercially, is a way of identifying, early on, people for whom this is not the case. Those are the ones you want to reach, when looking beyond your existing user base.
As I’ve said before, I’ve heard a fair amount of vibe-ing that viral marketing on Facebook just ain’t what it used to be. But friends of users should still be very highly sought after. Not, as some people seem to think, because they are free. Rather, because they are are more valuable.
So much so, that Facebook is offering a way to charge for them, via its Ads for Applications feature. However it’s worth testing exactly what it yields for you. One of the study’s many other interesting findings was that although onboarding via a friend invite is a good predictor of engagement, friend invites sent out en masse had the worst success rate in terms of uptake.