[DSC Europe 23][Pandora] Siyu SUN Data Science Enter The Game.pptx
Download as PPTX, PDF0 likes40 views
Siyu Sun presented on applying data science in the gaming industry. Some key applications discussed include optimizing monetization through improving in-app purchases, ads, and subscriptions. Specific use cases covered predicting payers versus non-payers, ad trigger points, and subscription tier and pricing optimization. Challenges with A/B testing large changes over long durations were also noted. Reinforcement learning applications and the future of generative AI and text-based games were briefly discussed at the end.
More Related Content
Similar to [DSC Europe 23][Pandora] Siyu SUN Data Science Enter The Game.pptx (20)
More from DataScienceConferenc1 (20)
![[DSC Europe 24] Anastasia Shapedko - How Alice, our intelligent personal assi...](https://cdn.slidesharecdn.com/ss_thumbnails/anastasiashapedkoyandexaliceforpeoplewithspeechimpairments-250218225516-7720cf27-thumbnail.jpg?width=560&fit=bounds)
![[DSC Europe 24] Joy Chatterjee - Balancing Personalization and Experimentatio...](https://cdn.slidesharecdn.com/ss_thumbnails/joy-chatterjee-ab-testing-friday-final-250218225454-d92d0731-thumbnail.jpg?width=560&fit=bounds)
![[DSC Europe 24] Pratul Chakravarty - Personalized Insights and Engagements us...](https://cdn.slidesharecdn.com/ss_thumbnails/pratulchakravarty-personalizedengagements-250218225450-d3180df3-thumbnail.jpg?width=560&fit=bounds)
![[DSC Europe 24] Domagoj Maric - Modern Web Data Extraction: Techniques, Tools...](https://cdn.slidesharecdn.com/ss_thumbnails/domagojmaric-modernwebdataextractionfinal-250218225444-c7bcad20-thumbnail.jpg?width=560&fit=bounds)
![[DSC Europe 24] Marcin Szymaniuk - The path to Effective Data Migration - Ove...](https://cdn.slidesharecdn.com/ss_thumbnails/marcinszymaniuk-250218225444-7999b477-thumbnail.jpg?width=560&fit=bounds)
![[DSC Europe 24] Fran Mikulicic - Building a Data-Driven Culture: What the C-S...](https://cdn.slidesharecdn.com/ss_thumbnails/franmikulicic-250217194605-ae6b204e-thumbnail.jpg?width=560&fit=bounds)
![[DSC Europe 24] Sofija Pervulov - Building up the Bosch Semantic Data Lake](https://cdn.slidesharecdn.com/ss_thumbnails/sofijapervulovbosch-250217194555-1bec347a-thumbnail.jpg?width=560&fit=bounds)
![[DSC Europe 24] Dani Ei-Ayyas - Overcoming Loneliness with LLM Dating Assistant](https://cdn.slidesharecdn.com/ss_thumbnails/daniei-ayyasovercominglonelinesswithllmassistant-250217194553-85be0404-thumbnail.jpg?width=560&fit=bounds)
![[DSC Europe 24] Ewelina Kucal & Maciej Dziezyc - How to Encourage Children to...](https://cdn.slidesharecdn.com/ss_thumbnails/ewelinakucalmaciejdziezyc-worldreader-250217194551-28a1a23a-thumbnail.jpg?width=560&fit=bounds)
![[DSC Europe 24] Nikola Milosevic - VerifAI: Biomedical Generative Question-An...](https://cdn.slidesharecdn.com/ss_thumbnails/nikolamilosevicverifai-250217194548-af2e8290-thumbnail.jpg?width=560&fit=bounds)
![[DSC Europe 24] Josip Saban - Buidling cloud data platforms in enterprises](https://cdn.slidesharecdn.com/ss_thumbnails/josipsaban-buidlingclouddataplatformsinenterprises-250217194546-5568421d-thumbnail.jpg?width=560&fit=bounds)
![[DSC Europe 24] Sray Agarwal - 2025: year of Ai dilemma - ethics, regulations...](https://cdn.slidesharecdn.com/ss_thumbnails/srayagarwal-250217194532-d8132a8a-thumbnail.jpg?width=560&fit=bounds)
![[DSC Europe 24] Peter Kertys & Maros Buban - Application of AI technologies i...](https://cdn.slidesharecdn.com/ss_thumbnails/peterkertysmarosbuban-aisolutionsslovenskeelektrarne-250217194529-8ed93d8d-thumbnail.jpg?width=560&fit=bounds)
![[DSC Europe 24] Orsalia Andreou - Fostering Trust in AI-Driven Finance](https://cdn.slidesharecdn.com/ss_thumbnails/orsaliaandreoufosteringtrustinai-drivenfinance-250216113615-6aa23936-thumbnail.jpg?width=560&fit=bounds)
![[DSC Europe 24] Arnault Ioualalen - AI Trustworthiness – A Path Toward Mass A...](https://cdn.slidesharecdn.com/ss_thumbnails/arnaultioualalen-250216113612-9b92043d-thumbnail.jpg?width=560&fit=bounds)
![[DSC Europe 24] Nathan Coyle - Open Data for Everybody: Social Action, Peace ...](https://cdn.slidesharecdn.com/ss_thumbnails/nathancoyle-peacetechbuzzwordorbreakthrough-250216113604-3c8d0e95-thumbnail.jpg?width=560&fit=bounds)
![[DSC Europe 24] Miodrag Vladic - Revolutionizing Information Access: All Worl...](https://cdn.slidesharecdn.com/ss_thumbnails/miodragvladic-revolutionizinginformationaccessallworldknowledgeinbillionsofconnectednuggets-250216113002-bf2aab42-thumbnail.jpg?width=560&fit=bounds)
![[DSC Europe 24] Katherine Munro - Where there’s a will, there’s a way: The ma...](https://cdn.slidesharecdn.com/ss_thumbnails/katherinemunro-intentdetection-250216113000-adcae328-thumbnail.jpg?width=560&fit=bounds)
![[DSC Europe 24] Ana Stojkovic Knezevic - How to effectively manage AI/ML proj...](https://cdn.slidesharecdn.com/ss_thumbnails/anastojkovicknezevic-250216112956-9624ef04-thumbnail.jpg?width=560&fit=bounds)
![[DSC Europe 24] Simun Sunjic & Lovro Matosevic - Empowering Sales with Intell...](https://cdn.slidesharecdn.com/ss_thumbnails/simunsunjiclovromatosevic-250216112916-e19beeba-thumbnail.jpg?width=560&fit=bounds)
[DSC Europe 23][Pandora] Siyu SUN Data Science Enter The Game.pptx
- 1. Data Science: enter the game
- 2. Siyu Sun (She/Her) Global Director Data Science and Analytics [email protected] sunsiyu Data Science Marketing Analytics +5 years in gaming industry: ex-Blizzard, Gameloft Has always been in the data science domain but across industries: FinTech, digital travelling (Booking.com) etc. Game Analytics
- 3. Agenda 1. Background Introduction 2. Categorizing DS Application in Gaming 3. Top Use Cases 4. Special Focus on Monetization 5. Help with a DS Puzzle 6. The future trends
- 4. MONTHLY ACTIVE USERS WORLDWIDE ACROSS ALL PLATFORMS 4.5 STARS AVERAGE PLAYTIME PER USER 5 SESSIONS/DAY 27 MINS/DAY GAMELOFT GAMES ARE DOWNLOADED EVERY DAY +80M +1.5 MILLION AVERAGE GAMES RATINGS ON APP MARKET PLACES AT LAUNCH GAMING ENTERTAINMENT COMPANY, PUSHING THE BOUNDARIES 275 GAMES SINCE 2000 LEADER IN THE DEVELOPMENT AND PUBLISHING OF GAMES Paris Barcelona Montréal Toronto Bucharest Budapest Sofia Kharkov Lviv Saigon Hanoi
- 5. A PASSION FOR GAMES A RICH AND UNIQUE GAME PORTFOLIO COVERING THE MOST POPULAR GENRES Asphalt 9 Disney Dreamlight Valley Disney Speedstorm Disney Magic Kingdoms Minion Rush (Despicable Me 2) etc. OVER 200+ GAMES
- 6. +1 BILLION DOWNLOADS …AND GROWING! FRANCHISE BENCHMARK FOR MOBILEARCADE RACING ASPHALT 8: AIRBORNE AMONG THE WORLD’S TOP 10 MOST DOWNLOADED MOBILE GAMES OF ALL TIME +1 BILLION DOWNLOADS …AND GROWING! AMONG THE WORLD’S TOP 10 MOST DOWNLOADED MOBILE GAMES OF ALL TIME +200 MILLION DOWNLOADS …AND GROWING! ALWAYS AIMING FOR THE BEST INDUSTRY AND PLAYERS RECOGNIZE OUR ART AND CRAFT +3 MILLION DOWNLOADS …AND GROWING! #1 ON STEAM, GAME PASS AND NINTENDO STORES AT LAUNCH PART OF THE BEST GAMES OF 2022 FOR KOTAKU
- 7. FEATURE PHONES +200 CARRIERS IN THE WORLD MOBILE WIDE RANGE OF BUSINESS MODELS TRY & BUY PREMIUM FREE TO PLAY SUBSCRIPTIONS WE WANT EVERYONE TO PLAY GAMES FOR ALL PLATFORMS CONSOLES & PC
- 8. ~42400 Jobs/day +250 TB of data ~1700 Jobs/hour 25 to 30 N. of threads +300 GB Increase/day ~1020 Heavy agg jobs A snapshot of our data:
- 9. “ Provide usable data products based on trustable and actionable data to enable game and support teams to take the best decisions and improve our products. DS&A Team Mission
- 10. Categorizing DS Application in Gaming Game Analytics Marketing Analytics DS / ML / AI Predictive Analysis Prediction as Product / System AI for Gaming Dev Business Verticals: Level of DS: “Traditional” Machine Learning / Deep Learning / Reinforcement Learning
- 11. Data Science Use Cases in Gaming
- 12. Optimize Monetization F2P: Free-to-Play How games generate revenue? 3 Commonly used Monetization Methods in F2P Gaming: 1. In-App Purchase (IAP) 2. Ads 3. Subscription
- 13. IAP Optimization Predict PAU vs NPAU Analyze top conversion point, and predict next conversion point Which offer category? Any purchase strategy? Where is the entry point of in app promotion pop-ups? Dynamic pricing for different players with different content How DS can help? Who, when, what, where, how much. Why
- 14. Challenges of IAP Optimization Predict PAU vs NPAU Analyze top conversion point, and predict next conversion point Which offer category? Any purchase strategy? Where is the entry point of in app promotion pop-ups? Dynamic pricing for different players with different content How DS can help? Who, when, what, where, how much. Why When to make the prediction? Some players pay early, some pay very late. What actions to take given the predicted result?
- 15. Ads Optimization Predict PAU vs NPAU Predict ads sensitive players Analyze top ads trigger point (placement) Which ad format? Which ad content? How frequent? capping, ads fatigue How DS can help? Who, when, what, where, how often. Balance
- 16. Subscription Optimization Battle Pass Season Pass Event Pass Subscription-like services that comes with a seasonality, and an obvious time-based or event-based pattern and rhythm. Subscription in F2P Gaming
- 17. Subscription Optimization Battle Pass Season Pass Event Pass Subscription-like services that comes with a seasonality, and an obvious time-based or event-based pattern and rhythm. Subscription in F2P Gaming A lot of design decisions: - Different tiers of subscription, cater to different level of needs - Pricing vs Rewards - Double optimization: monetization & engagement - Hard to A/B test due to the lengthy duration
- 18. A/B Testing to prove the value A/B Testing is the golden standard to quantify the value brought by DS products (e.g. prediction as segmentation, or predicted actions) And DS products do not always beat our existing baseline.
- 19. Beyond A/B Testing What if we cannot A/B test? (usually it’s because we lose control of true randomization) - Causal inference - Synthetic control
- 20. Help with this puzzle We want to predict churners. Among our current active players, who will churn. Our steps: 1. Define churn: consecutively inactive for X days 2. Define the ML problem – supervised learning (binary classification) 3. Gather past dataset and feature engineering in- game player behavior data 4. Train the model and perform prediction 5. Check the performance and top features
- 21. Help with this puzzle We want to predict churners. Among our current active players, who will churn. We ended with 2 top features: 1. how many recent inactive days 2. crashed sessions But… crashed sessions is negatively correlated with churn score. (aka, the higher the crashed sessions, the lower the player is likely to churn), we asked ourselves why?
- 22. What about Reinforcement Learning in Game? Cons?
- 23. In Summary Segmentation Action A/B Testing
- 24. The Future of AI in Gaming From traditional Machine Learning, to AI and beyond.
- 25. ChatGPT + Generative AI (MidJourney) The future of game development? Text-Based Adventure Game + Image Generation
- 26. THANK YOU Questions / Ideas?