Cases
Migros lowers CPA 30% during Ramadan with IBQML predicted users.
Challenge:Migros, one of Turkey’s largest supermarket chains, aimed to capitalize on the increased demand during Ramadan and the approaching Eid. The goal was to maximize this seasonal trend by predicting customer behavior more accurately.
95%
Higher ROI (Demand Gen)
111%
Higher CVR (Search DSA)
Higher AOV (App Engagement)
14%
Approach:Migros aimed to enhance customer demand prediction by using first-party data (1PD) and privacy-safe user data through Instant BigQuery Machine Learning (IBQML). The strategy involved analyzing insights from the previous three Ramadan periods, focusing on variables like city and time of day to identify high-potential conversions for Ramadan 2024.
Partnering with SEM:In partnership with their data partner SEM, Migros used historical data from Google Analytics 4 (GA4) and BigQuery to model customer segments using IBQML. Supported by Google Cloud’s Vertex AI, they utilized the full range of IBQML functionality to create automated customer match lists and targeted marketing strategies.
Results:Integrating Ramadan-modeled first-party data into full-funnel campaigns led to the following results compared to campaigns without 1PD:Cost: 30% lower cost-per-acquisition (CPA).Profitability: 95% higher ROI and 60% higher return on ad spend (ROAS).Conversion & Basket: 111% higher conversion rate (CVR) and 14% higher average order value (AOV).