Case Study: 44pixels achieves a 23% ROAS lift with Voyantis predictive acquisition signals

44pixels lifts ROAS 23% with Voyantis

44pixels, an AI app studio behind products like Cue and Pixi, faced a challenge with its Google Search acquisition campaigns. Their tCPA strategy was acquiring subscribers efficiently but could not distinguish between high-value annual subscribers and lower-value weekly subscribers, leading to an inefficient mix and varied retention profiles. The company partnered with Voyantis to solve this by turning long-term customer value into an actionable acquisition signal.

Voyantis implemented a solution using predictive modeling to forecast each subscriber's value from their first session. They built separate models for weekly and annual subscribers and engineered these predictions into calibrated signals for Google's bidding system. This allowed 44pixels to bid based on a user's predicted long-term worth. The results from Voyantis included a 23% increase in ROAS, a 33% decrease in refund rate, a 12% decrease in cancel rate, and a 6% reduction in cost per subscriber.

44pixels