Unity Focus: How twimler kept Fruit Bump sweet with analytics

Unity Focus: How twimler kept Fruit Bump sweet with analytics
Unity

Presented by Unity

October 6th 2016 at 12:56PM

Majid Khosravi reveals why the studio decided to utilise Unity Analytics and the ways in which the data-tracking platform allowed its match-three title to thrive

As any good developer knows, nowhere are analytics more vital than on mobile. With factors such as session length and average revenue per user (ARPU) correlating to the sink-or-swim nature of most smartphone titles, tracking and understanding data is absolutely key.

When Twimler decided to launch its next title in one of the mobile sector’s most highly-contested genres – the match-three game – founder Majid Khosravi knew that picking the right analytics platform could make or break Fruit Bump.

“Unity Analytics not only offered a range of free, powerful tools which worked seamlessly with the Unity game engine, but also boasted a user-friendly, fully comprehensible interface with which to analyse the data,” he recalls. “Given the importance of said data in today’s game industry, choosing an appropriate analytics provider was a decision I took very seriously. However, where other analytics providers were concerned, I often found them to be lacking in at least one or more of Unity Analytics’ aforementioned strengths, which in turn led me to choose the platform.”

KNOW YOUR STUFF

Of course, finding the right service is just the start – devs must then ensure they are keeping on top of the right metrics in order to fully appreciate their players’ trends.

“The session length and ARPU metrics, combined with the custom reports functionality were of particular value, as they allowed me to identify important information regarding which game levels, on average, players were reaching before dropping out,” Khosravi reveals. “I could even identify which levels appeared to encourage the sole IAP Fruit Bump offered and boost revenue.

“Utilising this data, in combination with Unity Ads, saw our daily average users grow exponentially from 500,000 to 800,000 since mid-January – creating 300,000 new daily average users in just two months – whilst resulting in over 70 per cent increase in sales and over 60 per cent increase in ad revenue.” 

DIVIDE AND CONQUER

Analytics can track a huge number of factors, all of which are useful both independently and in relation to other elements, and which can be used to group similar users together in order to refine an effective strategy for each type of player.

“By using custom events such as levels completed, total session time and level completion rate, we were empowered to, effectively, segment our users into skill levels,” Khosravi recalls. “This allows us to better monitor how players are progressing and alter level designs to maintain an optimum user progression rate and overall user engagement.

“By taking IAPs into account, we also segmented users into ‘buyers’ and ‘non-buyers’, ensuring that the experience each segment of player enjoys is better tailored to them. Both groups are now offered different IAP special offers, with ‘buyers’ being presented with package deals focused on customisation or time-saving, whereas ‘non-buyer’ packages may focus on limited-time, high-value offers to better encourage that first purchase.” 

DESIGN ASSISTANCE

Coming to understand its players better has also affected Twimler’s development methods for the future, Khosravi adds, with analytics extending to influence core design facets.

“Preparing custom events appropriately, depending on your game, can really help in analysing difficulty levels and assessing how players are engaging with your game,” he advises. “Achieving cognitive flow, where gameplay strikes that perfect balance between difficulty and skill can be, in my opinion, vital to any game’s success.

“As experience has taught us, certain gameplay can seem simple in a developer’s mind but, for whatever reason, either fail to draw mass appeal from users or even confuse them. We now use Unity Analytics in combination with beta testers to better assess Minimum Viable Product performance before additional investment is considered.

“Essentially, by adopting a data-driven design process we are better able to understand user preferences and in turn provide a richer overall experience and subsequently create a better end product.”