Discovery-driven Design in Social Games: Techniques, Processes, and Problems

The field of gaming is changing lately, with concepts like social, mobile and free-to-play, this has led to success for some like OMGPOP, however, recently the ‘poster child’ of social gaming, Zynga, took a big hit.

Successful companies in the field of social gaming are doing data analysis. It is said that Zynga hired 10 people (assumable data guys) before they hired their first game developer. What exactly they are doing is not exactly known. You can take courses in game design, but not (yet) in game data  analysis.

So a glimpse into how people are doing it:

  1. Monitor everything
  2. Try concurrent versions (Zynga recently told that they are currently testing 5000 versions of their games)
  3. Adapting the design dynamically
Basically, the same story as the previous talk on A/B testing.
KPI’s to monitor: virality, retention and value. It is very important to know how your users behave. People who are referred to a game by a friend are way more likely to stay, so people who are not spending a lot, but bringing on friends can be very valuable to a game.
Data mining for games is a growing market “like selling shovels from the gold rush”. The problem with game analytics is that is it easy to find problems, like the fact that players are not advancing quick, but it is hard to come up with a good solution. Just making players advance quicker might not be the best option.
We cannot expect that game designer will do the data mining. Normally, in companies product manager are concerned with the numbers, but should they do the data mining?
The Heather Stark’s suggestion is to move to discovery-driven design, where the core mechanic is to ask questions that lead to other questions.