game analytics challenges and visions of. what lies beneath?
TRANSCRIPT
Game AnalyticsChallenges and Visions of
What Lies Beneath?
Definitions
Analytics
Game Analytics
Game Telemetry
Game Metrics
Analytics
The process of discovering and communicating patterns in data towards solving problems in business
Supporting enterprise decision management Driving action Improving performance
Or for purely frivolous and artistic reasons!
Game Analytics A specific domain of analytics: game development and game
research
The game as a product: user experience, revenue …
The game as a project: the process of developing the game
Game telemetry Quantitative, unprocessed data obtained over any distance,
which pertain to game development or game research.
Describes attributes about objects
Many sources: Installed clients, game servers, mobile units, user testing/playtesting
Game metrics Interpretable, quantitative measure of one or more attributes
of one or more objects – operating in the context of games
Object: virtual item, player, user, process, developer, forum post ....
Attribute: an aspect of the object
Context: tied to process, performance or users of games.
7 challenges
1. Standards
Lack of standards
Makes it hard to communicate and share knowledge
Need a ”game analytics association” – to develop standards of terminology, practices and ethical guidelines
2. Unique beasts Games are not websites
Goal of games: user experience – not selling running shoes (virtual shoes maybe)
Games can be immensely complex information systems 100+ possible user/system and user/user interactions Extended periods of user-game interaction From 1 to lots of people interacting in-game
Hard to directly import methods from other IT-fields – adaptation needed
3. Social online focus Most advanced analytics currently in social online games/F2P
– and focused on monetization A/B Classification Prediction Segmentation Etc.
Rest of industry ”mostly” basic behavior analysis
Need analytics to improve UX, not just sell Farm Potions +5
Knowledge transfer image
4. Knowledge transfer
What is going on?
Minimal knowledge flow about methods, algorithms, ideas
No dedicated conferences or workshops
Presentations at events high level Not oriented towards application More high-level, marketing and ”bragging” than helping ...
4. Knowledge transfer Analytics is business intelligence – holds direct monetary
value A strong predictive algorithm can make a game Value: therefore kept confidential
Problem: re-inventing the deep platter
Need the front-runners to take charge: everybody benefits from knowledge transfer
5. Knowledge gulf Knowledge gulf: academia – industry
Academia provides a strong partner in analytics 1000´s of specialists in dozens of fields Can do explorative/blue sky research
Zynga, Wooga, Blizzard, EA ... – can build the expertise in-house – what about small/medium devs? – collaborate to innovate!
6. Lots´n lots of data Even a mid-size game can generate TBs of data per week –
> storage/processing
Reporting needs to be fast -> rapid analysis
Bandwidth vs. data coverage -> feature selection
Coverage vs. speed -> sampling
"You are no longer an individual, you are a data
cluster bound to a vast global network" –
7. Unrivaled power
”Never before have so few
known so much
about so many”
Unrivaled power2 powerful tools for monetization:
User knowledge
Analytics
Unrivaled power
User knowledge In-game Purchasing From game platforms (Facebook etc.) From Net tracking (Google etc.) Clickstreams From mining the Net (social mining) Geodata (mobile phones) National person databases ...
In the future knowledge of users will increase
Unrivaled power
Analytics & user researchLarge-scale, data miningPrediction, clustering, etc. Behavioral BiologyBehavioral PsychologySocial/community behavior science
When playing games, the barriers are down
Unrivaled power
User knowledge
Analytics
Great games
Luke skywalker image
Unrivaled power
User knowledge
Analytics
Revenue requirement
(potential for) Great evil
Darth vader image
The future
Game data mining
Huge untapped potential in dozens of fields/sectors: Human behavior analysis Spatial analytics Behavioral economics Insurance, banking and finance Social and community research Ecology and large-scale biological modeling ...
Game data mining
3 high-potential areas of game data mining:
Prediction: inform about future behavior of users
Behavioral clustering: making high-dimensional behavior datasets accessible
Association and sequence: finding the patterns and associations in how games are played
Behavioral clustering
SIVM: finding extreme profiles
Assassins Veterans Target dummies
Assault-Recon Medic-Engineer Driver Assault wannabee
Behavioral clustering
Each different playstyles, and different things that keep them in the game
”Driver”: drives, flies, sails – all the time and favors maps with vehicles
”Assassin”: kills – afar or close – no vehicles ”Target dummies”: unskilled newbies
Behavioral clustering
Use behavioral clustering to find profiles, then cater to them – in real-time
Monitor players´ profiles to track behavior changes: target dummy -> veteran
Spatial analytics
Games are experienced spatio-temporally All games require movement All games take time to play
Why is analytics then mainly temporal?
Beyond the heatmap
(Images: Ubisoft, Microsoft, Square Enix)
Spatial analytics Spatio-temporal analytics
Does not reduce the dimensions of game metrics data Deals with the actual dimensions of play.
(Image: Ubisoft)
Spatial analytics
(Image: Square Enix)
Spatial analytics
Decades of knowledge in spatial analytics outside of games – ripe for harvesting Trajectory analysis (how do users play the game? Move
in 3D?) Spatial outlier detection (finding exploitation spots, bugs) Spatial clustering (are players distributed across maps?) Spatial co-location patterns/trends (army composition in
RTS)
Adaptive games
Games that respond to the actions of the user in order to maximise UX (and/or revenue)
Left 4 Dead, Borderlands, Terraria, Virus ... – these relatively primitive but powerful – tip of the iceberg
Sizeable European/US community of researchers working for a decade on adaptive games
Future: Real-Time Analytics driving the game experience, within pre-planned frame (think pen-and-paper RPGs)
Automatization
Problem: time consuming analysis and reporting
Huge potential for automating analysis and reporting, interactive reports, etc.
Future: More effective analytics
Future: More interactive, tailored reports
Diversification
Currently focus on: Player behavior and monetization
Game analytics is much more: (Almost) all aspects of a game development can be measured Integrating and synchronizing data and sources Do not regulate the creative process! Games are diversifying! – analytics must follow suit
Knowledge sharing Game Analytics – maximizing the value of player data
50+ experts from industry and research
2 intro/foundation chapters (on website below): Game Analytics: The Basics Game Data Mining
IGDA GUR SIG
Slides from presentation will be available on: www.andersdrachen.wordpress.com
Blogs: blog.gameanalytics.com, engineroom.ubi.com, www.gamesbrief.com etc.
Contact: [email protected]
