coly Achieves 180%+ D90 ROAS With InMobi Advertising’s User-Level Targeting

InMobi DSP applied advanced machine learning to power coly’s user acquisition strategy. The platform built predictive models using behavioral and engagement signals from coly’s existing high-value players, enabling the identification of new users with similar traits and high LTV potential. These audiences were then activated across premium app categories, such as gaming, social media, and Manga, ensuring that spending was directed toward the most relevant and conversion-prone users. This approach allowed coly to efficiently scale user acquisition while maintaining strong alignment with its target demographic.

The campaign utilized InMobi Advertising’s machine learning–driven Auto Bidder, designed to optimize user acquisition performance in real time. The system ingested granular signals, such as install probabilities, in-app engagement patterns, to train predictive models. These models continuously adjusted bids across exchanges and app bundles, prioritizing high-intent cohorts while suppressing spend on low-value traffic sources. Leveraging reinforcement learning, the Auto Bidder dynamically reallocated budgets toward channels with the highest expected LTV, enabling coly to scale a steady flow of quality users while driving sustained improvements in ROAS.