The primary purpose of this analysis is to demonstrate how granular creative performance data can inform user acquisition (UA) strategy and production cycles. By examining 360 unique creatives over a 30-day period for the mobile game Robot Breaker, the analysis highlights the critical distinction between high-conversion metrics, such as Install Per Mille (IPM), and actual business profitability, measured by Day 1 Return on Ad Spend (ROAS). The central thesis posits that relying solely on algorithmic engagement metrics is insufficient; instead, marketers must synthesize creative-level revenue data with visual performance tags to identify which concepts warrant further investment.
The study covers a €41k spend across four major networks, resulting in 6,342 installs and a blended IPM of 3.80. Key findings reveal a significant mismatch between engagement and ROI: while playables achieved an IPM of 21.75—vastly outperforming video at 2.75—they did not always correlate with the highest financial returns. For instance, on Facebook, plane-led creative concepts yielded a 46% D1 ROI compared to 36% for character-led ads, despite the latter having a higher IPM. Furthermore, the analysis identifies that animated-real footage blends consistently underperformed, suggesting that production resources should be reallocated toward proven, high-ROI formats like character-led playables for AppLovin and plane-led, 40-60 second animated videos for Facebook.
Methodologically, the analysis utilizes an automated creative intelligence dashboard to tag visual elements, such as dominant subjects and video hooks, without manual input. This approach allows for the identification of specific creative patterns that drive both algorithmic favorability and actual revenue. The conclusion emphasizes that effective UA management requires a disciplined, data-driven approach to production, where creative brainstorming is limited to iterating on proven winners rather than testing random concepts.