The Problem with Mobile Game Data — and How to Fix It
Any mobile developer today can explain what terms like MAU and ARPDAU mean (monthly active users and average revenue per daily active user, for the uninitiated). But can they explain how to use it? In many cases, no — and that’s a flaw with some history.
Before 2010, data analysis wasn’t so widely known in games. A few large companies did it, but most developers — even those with free-to-play games — saw data on player habits as an optional second thought. Data infrastructures were difficult to implement, and there was little common knowledge available.
Fast forward to today: it’s easy to plug in off-the-shelf analytics, and we’re drowning in terms like retention and LTV. Devs can easily see if their game has players, if those players stick around, and how much money they’re making. But according to Andrew Turner, marketing manager at deltaDNA, these at-a-glance measurements “hide as much as they reveal. They can’t tell you how to solve a problem or what you should do to optimize your opportunity.”
So data analysis is at an inflection point. “Big data” and easy-to-use tool have opened the floodgates, but resulted in a great deal of gathering data for data’s sake. Developers will need to learn to be more thoughtful to make effective use of their data.
Picking apart big data
The shift in progress is from “big data” to an understanding of “deep data”. Big data tells devs what’s happening across large chunks of their player base — usually at too high a level to inform design decisions.
But, says Turner, “hidden behind these KPIs are small clusters of user interactions that are either killing or making your game.” These are known as deep data, which delve to the level of individual players to glean insights such as skill levels, interaction with the UI or interaction with other users. A growing number of companies like deltaDNA and Datameer are using visualization, cohort analysis and funnel analysis at the segment level to access this type of information.
“In a market where it’s increasingly hard to retain players long enough to monetize, what we are starting to see is the focus shifting toward the right data to generate actionable insights,” Turner says.
The secret to engagement
In past years some developers spent their energies tuning monetization; later the focus moved to retention.This move to more segmented, profound data coincides with a new focus on players: engagement. Engagement encompasses multiple measurements of player satisfaction. It’s a tougher target, but developers are learning that great engagement trumps (and improves on) both retention and monetization.
So how does data inform engagement? Say 12 percent of players are leaving at level 3 of a particular game. With segmentation, the developer could determine who those players are. Maybe the majority are skilled players leaving because of an overpowered weapon acquired too soon or soft enemy AI.
Segmenting according to skill set is just one of the capabilities of current advanced analytics tools. Another example is catching individual players churning out of the game. A point may be too hard for specific players, or some players may feel bored faster than others. This is where game developers can adjust game balance for these players, minimizing churn and improving the chances of monetizing a segment that would have been lost otherwise. The most advanced tools may detect the churn before it happens, proactively sending a message to the player in an effort to keep them from leaving.
Monetization methods can also be optimized using engagement metrics. Many games show the same ad at the same place to all their players. With segmented data, devs can make sure some group (such as paying players) don’t see the ads. Devs also have the option to trigger different types of ads to individual players who may be receptive, or adjust the appearance of the ad in-game to prevent churn.
The future of mobile game data
Advanced data is a return to better days for developers, back to a time that individual player voices mattered. With developers able to focus on the needs of individual players through data — an impossibility just a few years ago — we can only expect to see better iterations of current mobile game designs.
It might not be long until even the process of customizing the game becomes semi-automated. Turner thinks data analytics trends are bringing us toward automated optimization, with tools that will process the data and adjust player pathways accordingly.
John Morrell, senior director of product marketing at Datameer, offers another prediction for advanced data analysis. “Next we’ll see micro-segmentation for multiplayer games,” Morrell says. “A gaming company can use micro-segmentation based on behavior to match players in games, which matches that will last longer, and create more opportunities for monetization.”
