User Research in Fractal Spaces:

Transcription

1 User Research in Fractal Spaces:

2 Behavioral analytics: Profiling users and informing game design Collaboration with national and international researchers & companies Behavior prediction and monetization: Predicting human behavior, user experience, engagement, monetization Spatio- temporal analytics: All of the above, but in 3D (sports analytics) User Research: What is user experience? How do we measure it? Optimizing user testing value, getting at the why of behavioral patterns Digital Humanities: Using big data combined with deep data to answer questions in the Humanities best of both worlds Newspaper archives, Twitter, social network data Opinion mining, text mining, time- series analysis - > need explaining why patterns occur

3

4

5

6 Games are not websites, social networks or dating networks Goal of games: user experience not selling running shoes (virtual shoes maybe) Games can be immensely complex information systems Thousands of possible user interactions Extended periods of user- game interaction From 1 to lots of people interacting in- game and out- of- game Hard to import methods from other fields adaptation needed

7

8

9 In one decade: 200M - > 2B consumers 100 billion USD+ global industry The rise of mobile platforms new business models F2P and beyond More online games Numerous new ways to acquire data: Players, Process, Performance Changes in the development infrastructure and publishing, new tools for developers and researchers

10 Not just gameplay data: 360 degree view on users Profile data from social networks Off- game social networks In- game social networks Micro- transactions Advertising Payment systems Game system behavior/responses Geolocation Psychographic marketing Context data

11 Data acquisition prohibitive cost - > right at our fingertips Lab- based research - > populations of players 1 5 games per study - > thousands Second by second detail in player behavior - > more detailed studies Explore if previous work stands up on the large scale (replication) Testing theories and assumptions

12 Data surge also at small scales ( small and deep vs. big and shallow ) More data sources for GUR work If you have a game, you can get deep data about users - > data collection is (sort of) democratic Middleware tools enable tracking Mobile phones enable geolocation Facebook data from users Payment data...

13

14 You cannot improve what you cannot measure Lord Kelvin

15 "You are no longer an individual, you are a data cluster bound to a vast global network"

16 Analytics is the process of Business Intelligence Analytics is the process of discovering and communicating patterns in data towards solving problems in business Supporting decision management Driving action Improving performance Research & Development - Or for purely frivolous and artistic reasons!

17 Game analytics = subdomain of analytics: game development and game research The game as a product: user experience, behavior, revenue, system The game as a project: the process of developing the game Evidence- driven support for decision making

18 Strategic GA: The global view on how a game should evolve based on analysis of user behavior and the business model. Defining a monetization model, scoping DLC Tactical GA: informs game design at the short- term, A/B test of a new game feature, prediction, profiling Operational GA: Analysis and evaluation of the immediate, current situation in the game. Removing a bug, adapting game to user behavior in real- time, reacting to cheating/piracy

19 Vastly most common data source! User

20 User metrics Metrics related to the users, viewing them as either customers (revenue sources) or players (behaviors) Metrics related to the game system interacting with players (e.g. AI director) Metrics related to artificial agents behaving as players (bots, mobs, etc.)

21 ARPU, DAU, MAU - > customer focus - > revenue goal Avg. Playtime, Completion Rate - > player focus - > user experience goal Navigation, strategies, responses, agent behavior - > system focus - > game optimization

22 Hypothesis testing Confirming ideas/looking for confirmation Example: classification Problem: require existing knowledge/theory Explorative analysis We do not know what is going on/want to find out what is going on Example: clustering Problem: feature creep - > resource demanding

23 In practice often Mixed- Methods approaches Analytics good at WHAT Can only INFER why Need qual/quant to get at the WHY (and sometimes to find the questions to ask in the first place )

24

25

26 Never before have so few known so much about so many

27 User knowledge In- game Social networks Purchasing information From game platforms (Facebook etc.) From Net tracking (Google etc.) Clickstreams From mining the Net (social mining) Geodata (mobile phones) National person databases Traders actively sell and buy information about people In the future knowledge of users will increase

28 Psychology/analytics Behavioral Biology Behavioral Psychology Social/community behavior science Large- scale, predictive data mining When playing games, the barriers are down

29

30 Steam: Online game distribution platform 75 million active users 172 million accounts 3-7 million daily active users games/apps/sdks hosted

31 3000+ games 6 million+ players 5,068,434,399 hours of play - a lot, but: World of Tanks played 3.5 billion hours in 2013 DotA 2 is played 1040 years per day Main behavioral feature: playtime per game

32 Goal: investigate if there are patterns in how much time people spend playing games Result: Weibull modeling works well to model aggregate playtime frequency distributions

33 Implications: Fundamental properties governing playtime and how it evolves over time (?) We refer to this as the Playtime Principle

34 Supporting evidence: Bauckhage et al. [2012]: Tomb Raider, Crysis, Battlefield, Just Cause 2, Medal of Honor - same pattern (non- Steam data) Similar types of patterns reported for game session frequencies for individual games through the years, e.g. Feng et al. [2007] Lots of anecdotal evidence from the industry, e.g. GDC Similar patterns seen in e.g. internet popularity of memes, tech company searches

35 Why is this interesting? First time we have a documented pattern occuring across thousands of games not possible to do previously due to lack of data Strong industry interest: If playtime in games can be modelled using the Weibull distribution, we can predict playtime I.e. for any player, any game we can predict how long you will play the game Associated results in user profiling and LTV prediction: predicting when a player leaves and how much money they will spend: powerful models.

36 Why does playtime follow Weibull? We do not know need psychologists here Hypothesis: Weibull is comprised of two antagonistic growth functions increasing and decreasing interest The combined operations of these two functions may mimic human interest (?)

37