Case Study
Genshin Impact achieves 3.4X benchmark ROI with BrightLake's intent-based expansion
The Result
>60%
7-days ROI for cohort users
40%
Installs from non-branded keywords
30%
lower than average CPA

The challenge

As our client released the first open-world RPG games in the mobile platform and planned to expand into global markets, they encountered several key challenges when optimizing their Apple Ads campaigns. 1.Protecting branded keywords in multi-market at a comparatively low cost 2.High-performing non-branded keywords discovery when no direct competitors 3.Managing large-scale campaigns in multi time zones with efficiency

The solution

Optimizing Branded Keyword with Cost- Efficiency Bidding across markets We monitor hourly bidding fluctuations for branded keywords globally, enabling real-time, dynamic bid adjustments and effective keyword protection while optimizing for lower costs through impression share tracking.

High-performing Non-branded keywords discovery without similar products in the market Effective discovery of non-branded keywords is essential for driving incremental traffic and downloads, and despite initial challenges due to the lack of similar products at launch, our team leveraged insights into Apple's Black-Box Algorithm to create a high-quality keyword expansion database for testing and selection. By building predictive models that analyzed user engagement and payment behavior tied to specific keywords, such as retention rates, we efficiently identified high-performing non-branded keywords, significantly contributing to the client's long-term ROI goals.

Efficient Global Campaign Management Powered by Automation We implement automation rules and performance metrics for global campaign management, continuously monitoring keywords to optimize automated bidding and budget allocation, ensuring efficient achievement of the client's goals.

The results

The campaign achieved over 60% 7-day ROI—34% higher than the RPG category average—with 40% of installations from non-brand keywords (double the industry benchmark). Additionally, costs were 30% lower than brenchmark data, demonstrating both efficiency and scale.