The business problem
Bitso did not just need more app installs. In fintech, cheap acquisition can be fake growth if users never fund, trade, or become active. The real problem was user quality: attracting people likely to complete valuable post-install actions and stay active.
The system
1. Optimize beyond the install
The campaign architecture moved from surface-level acquisition toward post-install events. Registrations, funding probability, and meaningful app behavior became the signal layer.
2. UAC + Firebase learning loop
Universal App Campaigns and Firebase were used to train Google's machine learning toward users with higher expected value, not just lower acquisition cost.
3. Creative and reporting cadence
The system required constant creative iteration, event-level reporting, and disciplined reading of the gap between volume and quality.
The results
- +400% conversion rate versus campaigns directed to web.
- +5x monthly registrations.
- +240% improvement in user quality, defined by higher likelihood of funding and activity.
- Google Premier Partner Awards nomination for Best App Campaign in Latin America.
The operating lesson
Acquisition systems should never worship low CPI. Low-cost installs that do not activate are just expensive confetti. The useful KPI is not volume. It is qualified behavior after the click.