Based on public information as of November 5, 2026, South Korea's November shopping season has opened with three of the country's largest retail groups running overlapping discount programmes in the same ten-day window. Gmarket tripled its deal item count to 3,000, 11st bundled 180 brands into its Grand 11st event, and E-Mart put 160 tonnes of Korean beef on sale at half price while Coupang advertised discounts of up to 70 percent. The immediate read is a price war. The operational read is a store and delivery capacity contest, because every one of those offers has to be fulfilled from a defined trade area within a defined promise window.
What the November campaigns actually changed
The programmes differ in ways that matter more than the headline discount numbers. Gmarket is expanding assortment breadth, which stresses supplier readiness and warehouse pick capacity. 11st is pairing price with AI-assisted discovery, which shifts traffic toward whatever the model surfaces. E-Mart is using a single high-recognition fresh category as the traffic anchor, which puts the load on store-level stock rather than on a central warehouse.
Assortment breadth versus fulfilment depth
Tripling the number of deal items is a supply-chain decision before it is a marketing decision. Three thousand promoted items means three thousand chances of a stockout inside the campaign window, and a stockout on a promoted item is more damaging than an ordinary one because the customer arrived with a price expectation already formed. Retailers that publish availability per store, rather than a national in-stock flag, are the ones that keep the promise they advertised.
The fresh category as a traffic anchor
Half-price Korean beef is a category-level traffic instrument, and it behaves differently from a general discount. Fresh demand is perishable in both directions, so the campaign has to be matched to store-level throughput: how much the counter can cut and pack per hour, how far the cold chain reaches, and how quickly unsold stock converts to markdown. A retailer that allocates volume by national average rather than by store potential will over-supply low-traffic stores and lose sales in high-traffic ones.
Three retailers, three different bets
The table below compares the public positioning of the main participants in this season. Figures are as announced or reported, and the last column states what each bet depends on operationally. Reading it row by row shows that no two competitors are competing on the same variable, which is why a single discount index will mislead anyone tracking the market.
| Retailer | Public move | What it stresses | Operational dependency |
|---|---|---|---|
| Gmarket | Deal item count tripled to 3,000 | Assortment breadth | Supplier readiness and pick capacity per fulfilment node |
| 11st | Grand 11st with 180 brands plus AI shopping and service tools | Discovery and conversion | Quality of product data feeding the recommendation layer |
| E-Mart | 160 tonnes of Korean beef at half price | Fresh traffic anchor | Store-level allocation and same-day picking throughput |
| Coupang | Discounts advertised up to 70 percent | Price leadership | Delivery density and promise accuracy in each trade area |
Read across the row rather than down the column. Each retailer has picked one variable to lead with, and each of those variables fails in a different place: breadth fails on availability, discovery fails on data quality, fresh fails on allocation, and price leadership fails on delivery promise accuracy. A retail operations team monitoring the season should therefore watch a different indicator for each competitor instead of a single discount index.
Why store coverage decides the winner
South Korean grocery and general merchandise demand is served through an unusually dense store network, and density changes the economics of a promotion. Where a customer has three stores within a short distance, a stockout at one store is partly absorbed by the others, and the campaign survives. Where the network is thin, the same stockout becomes a lost basket and a support ticket. The consumer trend behind this season is not deeper discounting but wider availability, which is why instant retail and same-day delivery promises are only credible where store coverage and stock data are managed at trade-area level rather than nationally.
Promise accuracy as a competitive metric
The delivery promise is now a published commitment, and customers measure it. A retailer that advertises same-day delivery and misses it on a promoted item loses more trust than one that never offered the promise. The practical metric is not average delivery time but the share of orders that arrive inside the stated window during the peak days of a campaign, broken down by store and by trade area.
Where AI enters the loop
AI shopping assistants and AI customer service are being layered onto the campaign rather than replacing it. Their measurable contribution is narrower than the marketing suggests: they shorten the path from intent to a specific product listing, and they absorb repetitive service questions during peak load. Whether they lift the basket depends on whether the product data behind them is complete enough for the model to answer with a specific item instead of a generic category suggestion.
What operators should do this month
The sequence matters because each step produces the input for the next. Store-level availability without a promise metric tells you what you had, not what the customer experienced. A promise metric without competitor tracking tells you how you performed against your own standard while the market moved underneath you.
FAQ
Does a longer campaign window really reduce fulfilment risk?
It spreads volume but not necessarily evenly. Peak days still concentrate, so the useful control is per-store capacity planning rather than the length of the window.
Why does store coverage matter more than discount depth?
Because the discount only converts if the item is reachable. In a dense store network a stockout is partly absorbed by neighbouring stores; in a thin one it becomes a lost basket.
How should a retailer measure AI shopping features during a sale?
By the share of sessions where the assistant returned a specific product listing rather than a category, and by the conversion rate of those sessions against non-assisted ones.
What is the first data problem to fix before a peak season?
Store-reachable stock. National availability figures hide the allocation errors that cause most promoted-item stockouts.
