Tuning the Money Funnel: Customer and Process Metrics in Online Games
Moderator: Larry Mellon, Larry Mellon Consulting
Panelist: Marty Poulin, Social Sage
Panelist: Darius Kazemi, Orbus Gameworks
Panelist: Brian Hafer, TubeMogul, Inc.
Any industry that can measure the behavior of its recurring customers and processes benefits from doing so only if changes are taken and the results measured as well. Iterative improvements are possible in game play, community, economy, system stability, production iteration rates and recurring production or operational tasks. This panel looks at metrics that map drive continual process and player experience improvements. To make such metrics accessible and actionable, automation of the data acquisition and distribution process is critical, so data mining across game databases, user communities, forums, and the production to operations pipeline is also discussed. Some sample data sources are listed below.
Event-level tracking of players by category and by economy will also provide optimization information for increasing retention rates and focusing production dollars where they are the most profitable. Embedded profilers provide detailed views of user actions against game resource consumption: change the game play to reduce the cost or optimize the code against that feature, the choice is yours, if you can see where the hit is coming from in a complex distributed system…
Pipeline metrics, such as build failure rates, end to end cost of standing up a stable build, and common developer / operator tasks that take too long or fail too often allow you to make your entire team more effective: your bug DB, task management system, automated build and test system, pipeline tools and frequent talks with your production team yield the required data here.
Test-Driven Development is a vital source of information, providing a true, measurable view of, right now, what works in a system, and what does not. Extending pure class-level functional testing into system-level testing and projecting trends off of historical data gives an excellent view into overall progress and recurring trouble spots. This is also an excellent source of runtime resource utilization for optimization of recurring costs such as bandwidth or DB accesses by game feature.
Intended Audience:
Prerequisites:
Format: Panel
Track: Programming
Date/Time: Thursday, 2:00 PM
Room: Harbor
