How user clusters are changing game design: after the Game Design Conference

• Tweet
• Pin It

This is a post from editorial assistant Zoya Street, who is spending three months in the San Francisco bay area

I spent yesterday at the Game Design Conference in San Francisco as a media partner. When I first signed up to attend, I had a certain idea in my head of what kind of content it would attract – similarly to GDC, I was expecting post-mortems of level design, scenario design, audio and AI. I was expecting a bit of discussion of metrics and user data, but I didn’t think it would be a major focus.

I was surprised. While the program was full and varied, you could easily spend most of your time there hearing various points of view on how to track user data, how to optimise kpis, which service is best for improving your metrics and how analytics is changing game design.

Bigpoint’s presentation was a highlight of the conference, offering a revealing glance at the different forms of analysis they use to inform iterations on their existing games. They use the metrics we help you to optimise with the free-to-play spreadsheet, but they also look at their virtual economies to draw conclusions about what is in demand with players, and they use segmented analysis of user behaviour to make predictions about which features will appeal to the highest-value users.

It’s the last form of analysis that has really struck me as influential at the moment. Someone at the Cloud Gaming Conference last week called them ‘clustomers’, when specifically segmenting paying users. It’s a form of data analysis that spots similarities between users based on certain behaviours, identifies which behaviours co-occur with spending, and allows you to make predictions about how much certain users are willing to spend within a few minutes of them starting your game.

I spoke to Brendan Burke from Playnomics, a company that specialises in this kind of segmented user analysis combined with demographic data. They call it ‘psychographics’ – gathering insights into player behaviour to make predictions about their engagement, retention and spending. Their goal is to allow game developers to create custom experiences for each player’s taste, using predictive data about what kind of content is going to enthuse them the most. This will mean that not everybody experiences an online game in the same way – instead, the game detects what kind of player you are, and makes it easier for you to access the kind of content you are likely to enjoy the most.

At the Children’s Media Conference last Spring, a similar discussion happened regarding TV on demand. With this kind of analysis, on-demand services are getting better and better at showing you content that you are likely to enjoy. The end goal for some is to create a constant stream of personalised content with no friction. There will be no need to users to find what they want to watch – the software will already know.

The implications of this sort of thing for revenue are huge. At the moment, everybody who plays, say, CSR Racing, get the same series of messages asking them to spend virtual money. Some people hate it, and some don’t mind – the former group complain loudly about it on Twitter, and the latter group are responsible for the astonishing $12 million monthly revenue the game is generating. Imagine if the game knew within a few minutes – basically after your first play session – whether or not you were likely to respond well to requests to spend money. It could tailor the game to suit each player, so that the freeloading twitterati would be impressed by the game’s simple interface and enjoyable gameplay, while the high-rollers would have every opportunity to spend money to improve their gaming experience.

The impression I got from the Game Design Conference is that the kind of metrics we track with the spreadsheet – ARPPU, conversion, retention etc. – have become a necessity for developers. Nobody can afford to neglect these metrics, because most of your competitors are optimising their game along those lines. I suspect that segmented user data and dynamic game systems could go the same way – why tolerate a one-size-fits all game experience when somebody else can offer you something tailored to your own play style?

Related Posts

• Bigpoint gets bought, but for half a billion dollars less than its most recent investment valuation, a fall of over 80% in 5 years
  Bigpoint gets bought, but for half a billion dollars less than its most recent investment valuation, a fall of over 80% in 5 years
• 

  Understanding games designed for retention
  Understanding games designed for retention
• 

  Retention design in different cultures: know your players
  Retention design in different cultures: know your players
• 

  [Gamesbriefers] Is it dangerous to focus on whales?
  [Gamesbriefers] Is it dangerous to focus on whales?