111Percent reached new high-value users and increased ROAS through Moloco

About 111Percent

111Percent is a South Korean game company that brings unparalleled joy to users through its unique games. With a diverse range of mobile game titles such as Random Dice: Defense, Random Dice: Go, and Roll Roll: Dice Heroes 111Percent has garnered significant love from global users.

‍

Challenge 

111Percent aimed to increase return on ad spend (ROAS) by acquiring high-value users at a lower cost for its title 'Random Dice: Defense.' Due to the nature of the game, where leveling up requires more time and effort than casual games, securing users with high retention was crucial for increasing ROAS.

Prolonged marketing efforts over four years had already engaged a vast user base, decreasing new users and making it challenging to secure high-value users who could ensure high retention and conversions. To improve ROAS efficiency, 111Percent needed a new user acquisition (UA) strategy.

‍

Solution

Before diving into a new UA strategy, 111Percent conducted internal data analysis to understand the characteristics of their high-value users for the 'Random Dice: Defense' title. They hypothesized that users likely to purchase shortly after app installation significantly contribute to high, long-term retention and revenue. To validate this hypothesis, 111Percent partnered with Moloco to develop a customized UA strategy leveraging advanced machine learning capabilities.

Moloco identified that the D1 action model was the best solution aligned with 111Percent's internal analysis. Moloco's D1 action model shortens the learning cycle post-installation, focusing on actions occurring within a day. Leveraging 111Percent's first-party data, the D1 model trained on user actions within the first day post-installation and identified high-value users demonstrating a high likelihood of purchasing shortly after app installation. Additionally, 111Percent proceeded to execute A/B tests with Moloco to ensure the effectiveness of the D1 action model by utilizing Moloco's traffic control feature to divide traffic within advertising groups for precise targeting. 

‍

Results 

The A/B test results showed better ROAS and cost-per-install (CPI) in the D1 action model campaigns, confirming the validity of 111Percent's hypothesis and enabling them to effectively reach and acquire new high-value users. 111Percent improved its UA strategy through this collaboration, increasing long-term retention and sustaining revenue growth for 'Random Dice: Defense.'

• Increased ROAS efficiency by 2.5X with D1 action model
• Decreased CPI by 30% with D1 action model
• Exceeded target D1 ROAS goal by 7%

‍

“Through the D1 ROAS model, we were able to attract high-value users with purchasing power and maximize the efficiency of our ROAS campaigns. Particularly, the transparent provision of detailed campaign data instilled high confidence in our marketing performance compared to other platforms. Being able to analyze campaigns from various perspectives provided significant advantages in gaining insights. Additionally, thanks to the proactive support from Moloco's expert team, we were able to easily set up campaigns and monitor performance.”  — Seungwoo Yang, UA Team Lead, 111Percent