How Southeast Asia's leading super app achieved +170% CTR improvement with personalized restaurant recommendations across 6 countries.
68M+
Users
200K+
Restaurants
+170%
CTR Improvement
6
Countries
With over 2 million drivers, 68 million users, 3.5 million daily rides, and 1 billion rides booked to date, this food delivery platform processes 2TB of data per day. Operating across 6 countries with 200,000+ restaurants, the company struggled with restaurant discovery. Users in Singapore, Indonesia, Malaysia, Thailand, Vietnam, and the Philippines had vastly different cuisine preferences. The existing recommendation system couldn't handle the cold-start problem in new markets or the diversity of local food cultures.
We designed a three-band recommendation system tailored to user maturity. New users in each market received location-aware popularity recommendations reflecting local cuisine trends. Users with order history got personalized rankings based on their cuisine preferences, price sensitivity, and ordering patterns. The most engaged users received deep personalized suggestions that factored in time-of-day preferences, cuisine exploration patterns, and cross-market taste profiles. Each country's model adapted independently to local food culture while sharing learnings across regions.
Measurable impact across key metrics
Click-Through Rate
Before
Baseline
After
+170%
Value Generated
Before
—
After
$256M
Data Processed
Before
—
After
2TB/day
“The revenue generated by recommendations alone paid for the entire cloud infrastructure across all business units — turning AI from a cost center into the company's most profitable investment.”
Let's discuss how RecoMind can help your business achieve similar results.
Contact Sales