Using Data to Improve Player Experience: Insights from Rovio’s Game Analytics

Game analytics is surely an exciting and sought after career goal but it can be unclear for someone outside the industry to understand what it is really for. Everyone in a game team at Rovio is on a mission to develop the game that players will love and continue to return to for years to come. So it is the player who the game analysts think about when consuming and reacting to data. To this end, there are many ways to use data for improved player experience.

First, one way Rovio’s game teams create the best experience for players is by simply testing different ideas and choosing the best one. Ideas for these features and experiences can be inspired by other games, but quite often the game team draws inspiration from its own previous research. Both options require having an understanding of player behaviour and the intuition to know what is expected to happen when the test goes live. It is not always easy to come up with new ideas that bring significant business impact in a game which has already been optimised for a long time. The choice of the winner after a test poses another challenge for the game analyst because a seemingly good performance can be the result of random variation or novelty effect. In these cases, the observed uplift in the test may not hold in the long run.

To ensure that new features will have a long term impact, it is essential to have an estimation of player experience already when the game team works on a new game feature. This can take the form of simple spreadsheets with a handful of static numbers derived from an average player, to more elaborate stochastic models describing a sequence of possible events in the player journey (a map of how the player interacts with the game over a period of time). One way to assess player experience is to backtest how the new feature would have played out for example one month ago, if it existed then.

These solutions help the game team to configure the new feature so that the players will experience it in the desired way. When the feature goes live, the desired behaviour can be compared with the actual behaviour. If the players react surprisingly strongly to the new exciting feature, the configurations are changed via live operations to bring the desired player experience.

Even when the game team has a good understanding of how players behave, the game might not always work technically the way it has been designed, causing players to not be able to do things that they wanted. Data monitoring and alerting helps the game team to be aware of these technical problems and get the player facing issues quickly under work. The game analysts can go deeper and troubleshoot to discover the reasons and business implications of the problems. It can turn out, for example, that a newly released in-game event gets stuck for some players. In this case the team might have received an alert on a low number of purchases. At times, the impacted players can be compensated for the trouble they had with the game once the issue is discovered.

Another interesting mismatch is a situation where players behave differently than the game team expected. Such a finding is likely to lead to a change of design. Consider a powerful in-game resource which a large share of players fails to understand well. For the game analyst this situation shows up as a bimodal distribution of level retries where one group wins easily with the powerful resource but the other group struggles to beat the level. Seeing this kind of distribution is a key indicator that the team should reexamine the design of that feature, making its usefulness more clear.

Player profiles have large variances in free-to-play mobile games, especially due to the recent changes in privacy settings which make user acquisition less targeted. The game needs to be fun for all players, including a player who has not made purchases and maybe spends more time playing through the content, as well as players who purchase in-game resources that allow them to experience the game content more quickly. The moment when the player makes the first purchase for additional resources must bring significantly more fun and progress so that the player wants to purchase again. The available resources, skill, and daily time spent on the game often vary so much that the game content has to be personalised for a player.

The personalisation typically means segmenting players based on their previous behaviour while playing the game. The same in-game event can require a little more from the most active players, but should also reward them more generously for completion. Players appreciate when the game bundles resources that they are in high demand and serves them this best matching offer instead of the same thing for everyone.

In conclusion, game analytics is a crucial aspect of game development, aimed at improving the player’s overall experience. By analysing player behavior and preferences, game analysts at Rovio can test and refine new game features, troubleshoot technical issues, and personalise the game for a diverse player base. By constantly monitoring and analyzing data, game analysts can help ensure that players have an enjoyable and engaging experience that keeps them coming back, which is instrumental in Rovio’s mission of crafting joy for players in gaming experiences that last for decades.