🎙️ Mastering Retention Podcast: The Role of Data Science in Game Development with Yoni Ruuskanen
Host: Tom Hammond, Co-founder of UserWise
Guest: Yoni Ruuskanen, Data Scientist at Metacore
Introduction
Data Scientist at Metacore, working on merge games. Over six years in the gaming industry. Began career as a QA tester, transitioned into data science. Previous roles at Frozenbyte and Ubisoft. Journey into Data Science in Gaming
🎮 Early Passion for Gaming: Started playing games at age four. First game was a 2D side-scroller called Prehistorik on PC. 📚 Educational Background: Studied psychology but left after six months. Later pursued business and finance. Sent open applications to game studios in Finland; joined Frozenbyte as a QA tester. Worked in finance but returned to gaming at Ubisoft. Transitioned from project management to data roles. Joined Metacore as a Data Scientist over a year ago. Role and Importance of Data Scientists in Game Development
🔍 Responsibilities at Metacore: Performs data analysis using SQL and Tableau. Assists with data engineering tasks for new game events. Develops data science models for player segmentation and LTV prediction. 🏗️ Versatility in Smaller Companies: Acts as a "jack of all trades" in data-related tasks. In larger companies, roles may be more specialized. 💡 Impact on Game Development: Provides insights that influence game design and live operations. Helps teams understand player behavior and preferences. Utilizing Data Science Effectively
🧠 Problems Suited for Data Science: Predicting player lifetime value (LTV). Building predictive models using gathered data. 🚫 Problems Less Suited for Data Science: Understanding player motivations and reasons behind behaviors. Requires qualitative methods like surveys and direct player engagement. 🤝 Importance of Collaboration: Data scientists should work closely with game designers and other teams. Domain knowledge enhances the relevance of data analysis. Approaches to Data Analysis and Problem-Solving
Analyzing unexpected patterns can reveal underlying issues. Example: Detecting negative virtual currency balances leading to code fixes. 💬 Combining Qualitative and Quantitative Data: Engaging with players and playing the game provides context. Surveys complement data analysis to understand player needs. Player Segmentation Strategies
Dividing players into groups based on various criteria. Helps tailor experiences and offers to different player types. 🔧 Methods of Segmentation: Using clustering algorithms for unsupervised machine learning. Features might include spending habits, play styles, and game progression. 🎯 Applications of Segmentation: Informing game design and feature development. Targeting players with personalized offers and events. Enhancing monetization through tailored experiences. 🧠 Incorporating Psychological Motivations: Including qualitative data like player motivations can enhance segmentation. Tools like surveys help gather this information. Challenges and Best Practices in Data Science
⏳ Timing of Data Scientist Involvement: Early involvement helps design effective data collection systems. Prevents technical debt and data comparability issues later on. 🔨 Designing Reliable Data Systems: Preference for in-house solutions over third-party tools for transparency. Ensures trust in data accuracy and reliability. 📈 Managing Segments Over Time: Regularly review and adjust segments as the game evolves. Collaboration with live ops managers and designers is essential. Leveraging Player Behavior and Seasonality
📆 Understanding Player Patterns: Analyzing when and how players engage with the game. Adjusting live ops events based on daily and weekly player activity. 🛠️ Using Models to Enhance Live Ops: Incorporating seasonality into segmentation models. Designing content and events that align with player availability. Conclusion and Retention Strategies
🔑 Key to Increasing Retention: Understanding how retention metrics are calculated. Being cautious with third-party retention figures. Using additional metrics like session and match retention for deeper insights. Combining data science with qualitative insights leads to better outcomes. Data scientists play a crucial role in modern game development. 
