For those who cannot make my presentation on Wednesday at Casual Connect (4:30 pm) about analytics (separate from my other talk on building a great game company), I wanted to summarize it here. First, I want to thank our Director of BI, Aren Arakylan, and our VP Operations, Ana Echeverri, who provided invaluable help in preparing this presentation.
The presentation will focus on the opportunities to use analytics beyond what is commonly being done in the social gaming space, a topic I have blogged about a few times. First, I will look at the current state of analytics in the social gaming space.
Current state of analytics
We are a data-driven industry and I do not want to trivialize the tools currently being used. These tools have allowed social games to leapfrog the traditional gaming space in profitability, growth and reach. They include products like Kontagent, Mixpanel, and Honeytracks, which provide a great picture of the current state of a game. Products such as these do a great job of measuring the key components of a game’s lifetime value: monetization, virality and retention. These tools allow:
- Executive dashboards. These provide easy visualization of high-level metrics so all stakeholders can quickly assess the performance of your game(s).
- Ad-hoc reports. The above tools allow you to address immediate concerns.
- Query/drilldown. The tools help you determine the cause of problems, such as issues with monetization or retention.
- Alerts. The current tools help you see quickly when action is needed in your game.
Although incredibly valuable, most companies are only touching the tip of the iceberg when it comes to how analytics can help your company. There are many techniques and tools that can improve your performance. The ones I consider most valuable, and what they can tell you, are:
- T-tests and ANOVA. Are there statistically significant differences among groups in usage or monetization? If the answer is yes,it can affect not only who you advertise to but what features and content you build out.
- Time-series analysis. How much revenue will you bring in next quarter and how many users will you have? This information is quite important both in determining resources to devote to your game and cash flow planning.
- Logistic regressions and decision trees. Who is more likely to monetize and how will they react to in-game messaging? This analysis helps you build your game to ensure you capture and optimize your high value players (what our competitors call “whales”).
- Survival analysis. How long will it take users to monetize, what impacts their retention and ultimately what will be the real lifetime value of a player? This data is crucial to determining the health of a game and long-term revenue potential.
- Clustering. Are there customer segments that could be approached differently? This analysis allows you to customize your game so it optimizes profitability from all customers, not only one subset.
- Association analysis. Are there items that sell together? This research is done all the time by retailers but social game companies have not used it rigorously. By pairing items (think Amazon), you can generate more revenue from your players.
- Text mining. How do users feel about your game and are comments on Facebook and Twitter trending positively or negatively? This tool allows you to see how your players are reacting beyond the immediate metrics and thus predict what long-term impact changes in the game will have.
- Monte-Carlo simulation. What happens if we tweak the mechanics in the game and how long a play session will last? This analysis allows you to judge what changes to make in a game.
- Linear programming. What is the optimal allocation of resources for supporting the game? This allows you to determine the team size and make-up for live game support.
There are tools on the market, though not focused on the gaming space, that help you conduct the above types of analysis. I will be discussing these at the presentation but feel free to reach out to me for information.
The key message from my presentation is not to become complacent because you have implemented an analytics package and are using it to optimize your game. You must look beyond your current tools and see what techniques and products can take your social game to the next level.