Coupang Eats Prepares AI-Powered Hyper-Personalized Menu Recommendations

Coupang Eats Filed Related Patent Last Month · Reflects User Priorities Including Ratings and Reviews · Rising Prices and Information Overload Fuel Consumer Fatigue · Aims to Expand Market Share Through Technical Convenience · Coupang Eats Says "Service Launch Timing Undetermined"

Markets|
|
By Lee Yong-sung
||
null - Seoul Economic Daily Finance News from South Korea

It has been determined that Coupang Eats is preparing a 'hyper-personalized' menu recommendation service utilizing artificial intelligence (AI). This is interpreted as a strategy to respond to changing consumption patterns in an era of high prices by presenting user-customized options centered on 'total expenditure' including food prices and delivery fees.

According to comprehensive reporting by Seoul Economic Daily on the 9th, Coupang Eats is pursuing the introduction of hyper-personalized ordering content and filed a related search and recommendation algorithm patent last month.

The core of this service is a paradigm shift in menu recommendation methods from the existing 'food type' focus to a 'price' focus. When users set budget ranges such as 'under 10,000 won' or 'under 20,000 won,' the system presents restaurants and menus matching that price range by reflecting current location, distance to restaurants, and real-time delivery fees.

Additionally, it narrows down restaurant and food options that meet conditions by reflecting elements users consider important as priorities, including △delivery speed △ratings △number of reviews △price △availability of benefits. It was reported that options like 'anything' may also be included for users who haven't decided on a menu.

Data utilization to improve recommendation accuracy is also expected to be implemented alongside. Methods being discussed include analyzing groups with similar tastes to the user and reflecting 'representative user' ordering patterns, or continuously updating preferences based on recent order history.

This attempt is analyzed as a result of the combination of market conditions where consumers have come to consider 'total expenditure' before menu items due to soaring dining-out prices, along with structural changes in the delivery market. As dining-out prices continue to rise, consumers have increasingly tended to order based on 'how much they're spending' rather than the menu, and in addition, selection fatigue due to information overload has grown as the number of restaurants on delivery apps has increased.

In this regard, Coupang explained in its patent filing documents that "the rate at which users purchase products can increase when personalized content is included on e-commerce pages." It is known that other delivery platforms besides Coupang Eats have not been reviewing or introducing this type of recommendation technology to date.

Forecasts suggest that if Coupang Eats actually implements such a service, the landscape of the delivery app market will change. The observation is that Coupang Eats, which has risen to second place in the industry surpassing Yogiyo, can narrow the gap with first-place Baedal Minjok through 'technological convenience' going forward.

According to analysis based on Mobile Index data from IGAWorks, Baedal Minjok maintains first place with a 57.6% share based on monthly active users in March. Coupang Eats has recently broken through 30% market share, widening the gap with third-place Yogiyo (10%) while chasing first-place Baedal Minjok. Coupang Eats' average monthly per-person credit and debit card payment amount as of March is 122,129 won, ahead of Baedal Minjok (105,687 won) and Yogiyo (67,582 won).

However, Coupang Eats maintains a cautious stance regarding specific service launch schedules. A Coupang Eats representative said, "This is one of various patents for features to provide a better customer experience, and there are no confirmed plans for implementation."

AI-translated from Korean. Quotes from foreign sources are based on Korean-language reports and may not reflect exact original wording.