CC Games is a 20-person gaming studio that started in 2015 with a simple game called “Checkers,” created as a programming exercise. The success of this game exceeded expectations, giving rise to one of the best-known Polish mobile game development companies. Today, with the titles “Checkers” and “Chess,” the company has more than 100 million downloads to its credit.
Company acquires most of its users through User Acquisition (UA) activities. In 2021, 80% of players came from paid campaigns - by 2023, that number had risen to 94%. UA has become a key strategic and financial pillar of the company.
Before partnering with us, campaigns were based on simple metrics - installations, number of games played, etc. However, the company's CEO recognized the need to move away from a mass approach and target marketing efforts to the highest value users. He also noted that tROAS campaigns, despite being the market standard for targeting high-value users - were not producing satisfactory results. Their results were too variable and difficult to predict, limiting the ability to scale marketing efforts.
Project Objective
Increase the effectiveness of marketing campaigns by identifying the most valuable users early in their interaction with the app.
Scope of work
- Analysis of user data
We conducted exploratory data analysis (EDA), identifying user behavior and patterns. - Definition of KPIs and timeframes
We defined key metrics: retention, in-app purchases, and rewarded ads, which indicate whether a user can be considered valuable to a game publisher. We also set timeframes - that is, short time windows after app installation in which the expected behaviors should occur in order to be used in ad campaigns. The shorter the time window, the faster Google's algorithm receives feedback and can learn effectively. At the same time, in order for the campaigns to be effective, we needed to link these quick signals to the long-term value of the user - which required precise identification of behavioral patterns in the first days of use. We performed this analysis using logistic regression models. - Creating a proxy event for Google Ads
We created an artificial event combining various user actions that allowed us to precisely target potential “whales” in campaigns. Instead of relying on single events, which often did not give a full picture of a user's potential, we focused on identifying combinations of behaviors. It was their co-occurrence in the short term after game installation that best predicted whether a user would become valuable in the long term. The proxy event we designed and the developers created was only triggered when a user performed a full set of actions - allowing Google's algorithms to learn quickly and accurately who was worth acquiring in ad campaigns. - Data transfer and preparation of the analytical environment
At the beginning of the cooperation, we built an analytical foundation for further activities - including data analysis and the design of predictive models. We integrated Google Ads and set up automatic data transfer to BigQuery, where further processing and analysis takes place. - Product dashboard (parallel project)
In parallel to the project related to prediction and targeting of valuable users, we developed a comprehensive product report on Google Cloud Platform. This report is used for daily monitoring of the app's KPIs and is used independently of the UA campaign. It allows the team to better understand player behavior, analyze retention and monetization, and supports decision-making in product development.
Results
Key results
- Increased ROAS from 8% to 41% - a fivefold increase in campaign effectiveness.
- Campaigns based on the new targeting model delivered many times better returns than campaigns focused on installations.
- Scalability of the model - ability to test markets and segment by user behavior from different countries.
- Transparent reports with data processing automation, updated daily in BigQuery.
- Reduced costs by replacing full use of AppsFlyer with simpler integration with Firebase and dedicated features for iOS.
Business benefits
- Significantly increase the profitability of advertising campaigns.
- Reduce costs by simplifying the measurement environment.
- Automation and standardization of UA and LTV reporting.
- Increased competence of the client's team and inspiration for further analytical activities.
- New data quality - fast access, better precision, lower costs.
Summary
The CC Games case study is an example of how advanced data analytics and an informed marketing strategy can achieve significant advantages in the competitive mobile gaming market. Together we designed a solution that not only increased ROAS, but also built a solid foundation for further development of analytics in the company.
If you want to hunt whales, not just catch random fish - contact us!
