
The company possessed a massive database of 1.8 million dormant members representing significant untapped revenue potential, especially critical during the high-stakes CNY season. The challenge was to determine the most cost-effective channel (Meta vs. Google) and messaging strategy (Price-driven vs. Emotion-driven) to wake up these users without wasting ad spend on those who would have returned organically.
I designed and executed a comprehensive dual-axis A/B test (Meta vs. Google x Price vs. Emotion) across the dormant member base. To accurately measure true performance, I built Causal Impact models to isolate organic traffic from campaign-driven traffic, ensuring that the reported ROI reflected only the incremental value generated by the marketing effort.
The campaign successfully reactivated 50,000 dormant members. By isolating organic traffic, the Causal Impact models proved the campaign generated over NT$100 million in Incremental Sales Uplift, achieving an outstanding Return on Ad Spend (ROAS) of 190+.
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)