3 Ways Predictive Analytics Will Change Mobile Games: Part II

Predictive analytics is positioned as a force that will realign mobile game supply with player demand and make commercial success less random. This second installment of a two-part series, published by GameRefinery in February 2015, focuses on benefits to players and smaller developers, following a first part that addressed more efficient use of development resources.

The first argument is that more games will match what players actually want. A GameRefinery survey of US players, paired with a scan of the App Store, revealed a notable mismatch. Some 23% of US players said they want to spend time on RPGs, a share forecast to grow, yet fewer than 10% of App Store games are RPGs. Conversely, about 40% of App Store games demand sensomotoric skills such as fast reactions, while only 7% of players prefer them. Beyond market-level forecasting, the author argues that games themselves will adapt to individual players, for example by treating in-app-purchase-averse players differently from heavy spenders. Such behavioral prediction is already used by large studios, and simpler tools are expected to reach smaller companies.

The second argument concerns fairness in who succeeds. Because investors and publishers struggle to assess a game's potential before launch, they fall back on proxies such as a team's PC or console experience, which GameRefinery's database of hundreds of games and developers suggests is a poor predictor of mobile success. The company's study found that over 10% of the 300,000+ App Store games show hidden potential, equating to more than 3,000 overlooked "hidden gems," only a fraction of which large publishers have found. Wider access to predictive tools would increase transparency and reward merit.

The piece closes with preparation advice: build a fact-based decision culture without sacrificing creativity, and outsource analytics when no full-time analyst exists. It promotes GameRefinery's pre-soft-launch prediction services and mentions deltaDNA and Omniata as other tools worth evaluating.