Home About EIS →

Fantasy, Farms, and Freemium: What Game Data Mining Teaches Us About Retention, Conversion, and Virality

This past Saturday I had the pleasure of delivering a keynote presentation at the 2011 Mining Software Repositories (MSR 2011) conference (part of the pleasure being the location, Waikiki beach in Hawaii). My slides are available in pdf (1.3M) and ppt (13.5M).

The talk explores how to use mined gameplay data to reduce the cost of developing games, and to understand the interactions between game design decisions and player retention. It also takes a dive into the design of CityVille, focusing on the techniques it uses for attracting new players to the game. It summarizes research performed by Ken Hullett, Nachi Nagappan, Eric Schuh, and John Hopson on Project Gotham Racing 4, as reported in the paper, Data Analytics for Game Development in the NIER Track at the 2011 Int’l Conference on Software Engineering, and research performed by Ben Weber, Michael John, Michael Mateas, and Arnav Jhala reported in the paper, Modeling Player Retention in Madden NFL 11.

Some tidbits I found interesting while researching the talk:

  • Games (14%), books (13%), and entertainment (10%) are the three largest categories of apps in Apple’s app store. Business applications are 3%.
  • This is part of a trend where software use is becoming volitional. People are using software because it provides an enjoyable experience, and not necessarily because it is solving some specific need.
  • CityVille achieved 6 million daily active users only 8 days after launch. What if this kind of adoption curve was achievable in existing software segments? This could be very disruptive.
  • Retention engineering is a new subfield of software creation that is focused on how to design software so users have high engagement, and continue to use it. That is, if software use is volitional, you can no longer depend on some need to force people to use it. How do you ensure people continue to use software they really don’t need to use?

About the author:  Jim is Professor and Chair of Computer Science at UC Santa Cruz. He has research interests in procedural level generation for computer games, as well as automatic bug prediction. His favorite games are Radiant Silvergun and Civilization IV. Read more from this author

This entry was posted in Academics, Games and tagged , , , , , , , , , . Bookmark the permalink. Both comments and trackbacks are currently closed.


  1. RonSR
    Posted May 29, 2011 at 2:14 PM | Permalink

    Call it retention engineering or just plain psychological influence, but there is definitely a reason for the people producing apps in that lower 3 percent group to integrate volitional software principles into their works.

    You are highly astute to question the potential for a shake-up if a City-Ville-esque adoption curve is achieved outside of the gaming realm.


  2. Asim
    Posted May 31, 2011 at 2:12 PM | Permalink

    14%? Well, that’s good news for Apple’s app store! Anyway, i think the surge of gamers is contributed by the teenagers’ population.

  3. Rommel C. Caibal
    Posted June 13, 2011 at 6:25 PM | Permalink

    Interesting stats on popular apps. Not surprising how well Apple marketed their products to the general consumer, and the first 5 categories can easily fall under Entertainment. Thanks for sharing. :)

  4. David
    Posted July 20, 2011 at 11:27 AM | Permalink

    Actually, one of Zynga’s most popular demographics is middle-aged to older females. Simple freemium games with lots of immediate rewards and a slight social aspect. *click click click click click click click click*