CLV-driven acquisition & budget allocation

A fast-growing consumer fintech startup operating a mobile app with three subscription tiers was heavily reliant on paid acquisition, particularly Meta, to fuel growth. While acquisition volume was strong, profitability and unit economics were increasingly sensitive to user quality, not raw installs.

At acquisition time, users look identical — install, signup. True value (premium conversion, retention, monetization) materializes weeks or months later. Media platforms optimize for short-term events, not long-term value.

This led to three issues: CAC was optimized on volume, not value. Marketing teams lacked visibility on which acquisition sources produced premium users. Budget allocation decisions were reactive rather than data-driven.

The core business question: How do we acquire fewer users, but more of the right ones — those who convert to premium and generate long-term value — while keeping CAC under control?

The solution was an acquisition decision system that predicts Customer Lifetime Value early in the user lifecycle, focuses acquisition on future premium subscribers, feeds high-quality signals back to Meta to improve Lookalike targeting, and enables explicit CAC caps aligned with long-term value.

A predictive CLV model was built using early in-app behavioral signals, transactional patterns, subscription funnel progression, and engagement intensity. The objective was ranking, not perfect prediction — identify users likely to become premium and generate high LTV. This allowed segmentation of new users into value tiers shortly after acquisition.

The key enabler was closing the loop with Meta: matching high-value users to their Meta identifiers, sending these back as custom audiences, and building Lookalike audiences based specifically on high-CLV users — not all users. Lookalikes were trained on value, not volume.