Building High-Quality Data for Personalised AI Fitness Recommendations
1,589 responses collected (158% of target), 1,037 validated datasets ready for ML training
A leading Korean fitness recommendation platform worked with Tictag through Korea's Data Voucher Program to build a high-quality consumer behaviour dataset — sharpening recommendation precision and unlocking actionable marketing insights.
Problem
Despite 4.2 million active fitness enthusiasts in Korea, personalised product recommendations remained underdeveloped. Consumers wrestled with fragmented information sources while trying to match products to their goals, and brands faced rising marketing costs with limited market visibility.
Solution
Through Korea's Data Voucher Program, Tictag managed the complete data lifecycle. We ran structured consumer research capturing fitness product usage and preferences — standardising variables across workout goals, supplement habits, brand preferences, and purchase intent. Raw data was then cleansed, validated, and structured for correlation analysis between fitness goals, lifestyle factors, and spending patterns, meeting machine-learning compliance standards.
Results
- 1,589 responses collected — 158% over initial target
- 1,037 valid datasets secured after processing (≈65% retention rate)
- Sharper recommendation precision through behaviour and brand-affinity signals
- Actionable marketing insights enabling optimised consumer targeting