Big Data in Mobile Gaming: Optimizing the User Experience on iOS and Android – Eric Seufert

Cheap storage and a general lack of data often leads to less sophisticated statistical techniques when looking for insights while more data makes it easier to spot true trends and anomalies.

Lets talk about big data. Having a lot of different data points means you won`t have to normalize your findings and make any assumptions. Having less data will force youto make a lot of hard to verify assumptions.

Example - AB testing
 

• 200 data points, but you don`t know if size reflects the population and the distribution of the data is a mystery
• 2 million data points both makes distribution easier and requires less testing and normalization
   

What is data driven design

It is an iterative design process which is responsive to various different metrics. In a nutshell:
 

• MVP > behaviourial feedback > iterations > updates > repeat
   

Data driven game design can be thought about as half intuition and half data science. The fundamental game mechanics are established and then tweaked, based on player behavior, and the whole game is constantly updated according to the metrics.

Data driven game design can also be a creative process and it gives players much more of what they want. Giving consumers what they want is always more useful than just inflating the ego of yourself or your supervisor.

There are tons of console titles out there that just plain suck, so let players tell you what they like and don`t like!
 

Four key groups
 

• Retention
• Monetization
• Engaement
• Virality
   

Retention

Most important set of metrics to track as it communicates delight - extent to which players enjoy the game, which shows you the extent to which they`re willing to return to it.

It basically tells if your game sucks or not and gives you the elements of lifetime customer value. 
 

• Retention profle: D1, D7, D28
• Follows general decay pattern - D1*.5 = D7, D7*.5 = D28
• Used to calculate player lifetime
• Look for D1 -- 40% for launch
   

Don`t make your game shallow and boring

In order to make the project last and not run out of breath, you have to communicate long term values from the game itself. If the game is completed in a few days then the time of your player is pretty much wasted because they are looking for a deep experience that lasts more than a few hours.

They could have just downloaded a different game with deeper gameplay and heaps more awesome content.

Short games are not worth the time investment and players know this! Gamers want as much out of their games as possible, which is very reasonable when you look at the price tags of different console and PC titles, and also mobile. Some games on Google Play for example cost 17 EUR and up. 

 Monetization 

• Conversion rate: by day. % of users that ever spend on in-app purchases. 3-4% is good, 5+ is high 
• ARPU: average revenue per user/player. Average amount of money users spend in-game and varies from genre to genre and from game to game. 
• ARPPU: average revenue per paying user, a daily metric
• Catalog distribution - good to know where the bulk of review comes from. Small or large purchases? How do the users engage with the product catalog. 
   

Don`t push in-game purchases

In-game purchases are not for the first time players of your game. Putting too much emphasis on it could lead to alienation. Instead, target players who have played for a longer period of time. There is a also a chance that smaller sums leads to a bigger revenue. 
 

Virality

• K-factor - number of users on average that a single user introduces to the game
• Hard to track on mobile
   

Analytics strategy

Monetization and retention are easy, while engagement is medium and virality is bloody damn hard to achieve. For the very basics, tracking should be oriented toward the first session and last session. For the hard stuff tracking must be far more extensive. 
 

The point of analytics

• Collect and process data that can be used to derive insight and helps to increase revenue
• Your analytics should deliver $$$
• Analytics is never an intellectual exercise 
   

 Analytics is a vague term

• Analytics engineer - back-end guy, usually a system architect 
• Analyst - handles the reporting, dashboard design, implementation
• Data scientist - does research and algorithmic prediction, helps to drive data proven product design. 
   

Analytics is perfect for:

• Optimizing the user experience
• Giving players the optimal experience based on the information they give you through their behavior 
• Improving the product catalog
• UI/UX improvements through A/B testing
   

F2P design in a nutshell

• It`s a pain in the butt
• It requires a different approach to design
• I has a continous monetization curve
• High retention
• Low conversion
   

Overall it`s a useful tool for making priority decisions on game development, but it is not a one stop solution for every problem you might have.