South KoreaRetail and e-commerce2015 – ongoingKurly
Machine learning demand forecasting for fresh food
Kurly reported a 0.4% fresh-food waste rate in 2023 with in-house order forecasting
Problem
Kurly sold fresh groceries for next-morning delivery, with much of its stock lasting only days. Excess orders became food waste. Ordering too little led to stock-outs and lost customers.
Approach
Kurly began collecting and analyzing purchase data in 2015 and built a machine-learning system to forecast orders by region for the coming days. Buyers and logistics teams used the forecasts to plan purchases, warehouse capacity and staffing. In 2026, Kurly bought AI start-up Wonji Labs to strengthen its own forecasting and personalization technology.
Results
Takeaway
Kurly used order forecasts to plan purchasing and logistics as its reported waste rate fell from about 1% in 2021 to 0.4% in 2023 and inventory turnover improved. The waste figures were company-reported and not independently verified, and forecasting accuracy was not disclosed.
Categories
- Industry
- Retail and e-commerce
- Business function
- Logistics and supply chain
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and ML
- Adoption
- Built in-house
- Evidence
- Company's own claim
Sources
Source summaries
Dealsite linked Kurly’s smaller operating loss in 2023 to faster inventory turnover, which rose from 32.7 times in 2021 to 34.3 times on a consolidated basis. Kurly used algorithms and machine learning to forecast daily orders and plan warehouse operations and staffing. Dealsite reported a fresh-food waste rate of 0.4% in 2023, down from the 1% range.
Electronic Times reported that Kurly had used purchase data since 2015 and that its data science team forecast weekly orders by region with machine learning. Kurly reported keeping waste within 1% while handling about 120,000 orders a day. CEO Kim Seul-ah set a goal of 0.1%.
Newsway reported that Kurly bought Wonji Labs to bring forecasting and personalization technology in-house and use purchase patterns, seasonality and regional demand to manage inventory. Kurly recorded its first annual operating profit, 13.1 billion won, in 2025. Industry observers said the return on AI depended on how the company used data in logistics and inventory management.
Full details are in the original sources.