Coupang has begun building a system that uses artificial intelligence (AI) to forecast product demand and link it to suppliers' production and delivery plans. The move is seen as an effort to improve logistics efficiency, reduce the risk of excess inventory and stockouts, and strengthen the competitiveness of its Rocket Delivery service.
According to reporting by Seoul Economic Daily on Tuesday, Coupang (CPNG) has started developing a system to predict and manage suppliers' delivery volumes in advance. The related technology was filed as a divisional patent on May 6 and disclosed the previous day.
The system centers on Coupang forecasting future order demand and providing it to suppliers, who then establish delivery plans based on the data. In particular, order forecasting draws on machine learning technology based on Coupang users' past order histories. Various other data, including season, weather, region, and special events such as the World Cup, will also be reflected.
For example, if Coupang analyzes the past summer sales history of bottled water products along with the Korea Meteorological Administration's heat wave forecasts and predicts that 1,000 orders will occur over the next month, this data is shared with the bottled water manufacturer. Based on this, the manufacturer enters its production and delivery plans into the system, which Coupang then reviews. If demand and the delivery plan do not match, Coupang flexibly adjusts the ordering schedule and order volume to prevent stockouts.
In addition, Coupang plans to incorporate suppliers' past delivery records into the order forecasting system. If supplier A has repeatedly missed delivery deadlines in the past, the company can calculate the expected future revenue loss based on this record and take measures such as switching to another supplier B.
Coupang designed the system to repeatedly verify and check the AI-based information exchange process and to precisely forecast the next order volume based on feedback from users and suppliers. As user purchase data and suppliers' production and delivery feedback accumulate in the system in real time, forecasting accuracy is expected to improve further.
The reason Coupang has undertaken this technology development is interpreted as an effort to reduce the cost burden of excess inventory. At Coupang Inc.'s first-quarter earnings conference call held last month, Coupang Inc. Chairman Bom Kim cited "inventory cost burden" as one of the factors behind the 350 billion won loss. "When external factors such as a personal information incident occur, actual demand falls short of plans, and idle facilities and inventory costs arise in this process," Kim said.
The industry views this technology development as a way for Coupang to further strengthen the competitiveness of its Rocket Delivery service. Indeed, in its patent explanation materials, Coupang noted that "in preparation for changing customer demand, it is necessary to receive an appropriate quantity of products from suppliers and store them in fulfillment centers," adding that "inventory shortages or delivery delays can lead to revenue losses and a deterioration in customer experience."
It continued: "Suppliers find it difficult to predict future demand and order volumes, and they cannot know e-commerce companies' future order volumes in advance, so there is a risk of excess inventory. By confirming in advance the quantity suppliers can deliver in the future and adjusting order volumes, revenue losses from inventory shortages can be prevented and the inventory storage efficiency of fulfillment centers can be improved."
In response, Coupang stated that the technology is at the patent application stage, and whether it will actually be commercialized and a specific implementation date have not yet been determined.
