This episode breaks down Coupang, the Korean e-commerce giant that shifted from a Groupon-like model to building its own logistics, offering 'order by 11pm, deliver by 7am.' The guest, Drew Cohen, is bullish, saying Coupang wins by optimizing for speed, trust, and convenience—hard for rivals to copy. Key holdings: Coupang (profitable, 14M members, $1.5B free cash flow), Naver (biggest threat, low fees but struggles to coordinate delivery partners), and Market Curly (pioneered 8-hour fresh delivery, now expanding into beauty).
Coupang is a South Korean e-commerce giant that adopts a founder-led Amazon model, achieving success through logistics innovation (such as Rocket Delivery). The core argument is that, after undergoing a major strategic transformation, Coupang has replicated Amazon's operational model and secured a l
Guest Drew Cohen (Speedwell Research) provides an in-depth analysis of South Korean e-commerce giant Coupang. The core narrative: how Coupang transformed from a Groupon-style social commerce platform into an Amazon-like full-stack logistics e-commerce company under the leadership of founder Bom Kim, ultimately becoming the top e-commerce player in South Korea. The most impactful insight of the episode: Drew Cohen argues that Coupang's core advantage lies not in being a "first mover," but in its ability to identify consumers' unmet preferences (delivery speed, reliability, trust, and ease of ordering) and restructure the entire business model around these six preferences. This "full-stack optimization" creates a moat that later entrants cannot replicate, even in an ideal market like Seoul.
Drew Cohen believes that Coupang's most critical strategic decision was to voluntarily withdraw its IPO at the last minute, abandoning the proven Groupon model in favor of the asset-heavy Amazon model.
Bom Kim dropped out of Harvard Business School during his first year of the MBA program and introduced the Groupon model to South Korea, founding Coupang. By 2014, the company had generated hundreds of millions of dollars in revenue and was preparing for an IPO. However, Kim realized that the Groupon model was essentially a discovery platform for "impulse purchases and non-essential goods," lacking sustainability. "He said they weren't delivering a 100x experience for consumers" (meaning: he believed the existing model fell far short of the service level it should achieve).
So Kim voluntarily withdrew the IPO just weeks before it was scheduled, pivoting to a 1P direct sales model with in-house logistics. Drew emphasized: "He could have made tens of millions or even hundreds of millions of dollars, but he gave it up." This decision later attracted investments from Sequoia and SoftBank (SoftBank invested $1 billion), and the company has since largely relied on self-sustaining operations.
Key data: The transformation began in 2014, and the company achieved profitability in 2023, despite still posting a $1.5 billion EBIT loss in 2021.
Drew Cohen argues that Coupang's logistics system is the product of combining South Korea's population density with meticulous operations, achieving efficiency far beyond that of its U.S. counterparts.
South Korea has a population of 50 million across 22 million households, making it extremely dense. A single delivery driver can sometimes complete an entire day's deliveries within one high-rise building. Since 2014, Coupang has built its own logistics network and introduced the "Rocketman" delivery system—drivers write handwritten notes and remember household details (e.g., whether a family has a baby).
Key Milestones:
Mechanism Breakdown: Coupang's trucks have side-opening doors (instead of rear doors), saving seconds per delivery; plastic packaging replaces cardboard boxes, reducing cardboard usage by 85%—not for ESG reasons, but to pack more parcels into each truck. "All these process innovations stack on top of each other, continuously driving down the cost per delivery."
Key Difference from Amazon: Amazon initially relied on FedEx for delivery and only later built its own logistics network; Coupang built its entire logistics network from the start, as South Korea lacked a mature small-parcel delivery infrastructure.
Drew Cohen believes that Coupang's biggest competition comes from Naver (the Korean version of Google), but Coupang occupies the "most valuable transaction behavior" in consumers' minds—high frequency, essential needs, and instant decision-making.
| Competition Type | Representative Players | Characteristics |
|---|---|---|
| Department stores/Hypermarkets | Lotte, Shinsegae, Hyundai, Costco | Offline-focused, some experimenting with omnichannel |
| Pure e-commerce platforms | Lotte On, eBay Korea, SSG.com, Market Curly, 11th Street | Mostly third-party marketplaces, inconsistent delivery experience |
| Internet giants | Naver, Kakao | Naver is the biggest threat, with search gateway + payments + smart stores |
Naver's threat: Naver is South Korea's search gateway, with 20% of consumers using it as their preferred e-commerce platform. It offers Naver Pay (similar to Apple Pay), Smart Stores (unified product catalog format), price comparison features, and even aggregates Coupang's products for price comparison. Its blended commission rate is below 5%, far lower than Coupang's.
Coupang's defense: 60% of consumers list Coupang as their preferred e-commerce platform. Drew argues that Naver attempts to coordinate multiple logistics partners (Naver Shipping Alliance, in collaboration with CJ Logistics) to deliver a unified experience, but "incentives among different participants are misaligned"—Coupang can optimize for just a few variables, while Naver must balance the interests of multiple parties.
Consumer behavior insights: Drew's consumer survey reveals—for low-priced items (e.g., daily necessities), consumers go directly to Coupang and do not care about minor price differences; for high-priced items, consumers are willing to pay more for Coupang's return guarantee; mid-range items (apparel, cosmetics) are Coupang's weak spot and serve as entry points for competitors.
Drew Cohen argues that due to the accounting differences between 1P and 3P, investors should understand Coupang's economic model from a GMV perspective rather than revenue.
| Metric | Value |
|---|---|
| 2023 Revenue | ~$25 billion |
| 2018-2023 Revenue CAGR | 45% |
| Active Buyers | 21 million (22 million households in South Korea) |
| Rocket WOW Members | 14 million |
| Membership Fee | ~$5/month (raised to $7-7.5) |
| Annual Spending per User | ~$1,100 (3x increase from 2018) |
| Gross Margin | ~25% |
| 1P vs 3P Mix | 55% : 45% (estimated) |
| Total GMV | ~$38 billion (estimated range $35-42 billion) |
| Blended Take Rate | ~11% |
| 2023 Free Cash Flow | ~$1.5 billion |
Commission Structure Breakdown:
Drew believes Coupang's advertising business is still in its early stages — "It was only a few years ago that merchants began complaining about increased ad placements, indicating that ad monetization has just started." In the long term, advertising revenue could reach 4%-5% of GMV.
Mature Profit Margin Projection:
Operating Leverage: EBIT loss of $1.5 billion in 2021 → profit of approximately $500 million in 2023. Capital expenditure is more than double depreciation and amortization, but a significant portion is growth capex (47 million square feet of logistics facilities, 100+ fulfillment centers). "When large-scale construction is no longer needed, margins will continue to move higher."
Cash Flow Characteristics: For every $1 of inventory held, there is $3 in accounts payable (similar to AutoZone's negative working capital model); inventory turns nearly 10 times per year.
Drew Cohen believes that Coupang's growth primarily stems from category expansion and deeper consumer penetration, rather than entering new markets—though Taiwan is a noteworthy exception.
1. Category Expansion: Apparel, beauty, and cosmetics are weak areas, currently being supplemented by 3P sellers; the acquisition of Farfetch ($500 million, $4 billion GMV) may be used to strengthen the apparel category.
2. Deepening Consumer Behavior: Annual spending per consumer cohort joining Coupang is growing, and best-customer behaviors can be replicated across more users.
3. Advertising Monetization: Still in early stages, with room for multiple-fold improvement.
4. Taiwan Market: Entered in 2023, replicating the full-stack model from Korea, and has already achieved the status of "Taiwan's No. 1 Shopping App."
In the Japanese market, Coupang attempted "instant delivery" (delivering 300 SKUs, such as lunch boxes, within 30 minutes), but consumers did not respond. Bom Kim decisively shut it down. "He didn't make excuses; he just cut it off—exactly what you want to see from a founder."
| Position | Analyst Stance | Key Data |
|---|---|---|
| Coupang | Bullish | $25 billion revenue, 14 million members, $38 billion GMV, $1.5 billion free cash flow in 2023 |
| Naver | Risk Warning (largest competitor) | 20% consumer-preferred e-commerce platform, comprehensive commission rate <5%, owns Naver Pay + Smart Stores |
| Market Curly | Neutral (competitor) | Pioneered 8-hour fresh food delivery, now expanding into beauty products |
| SSG.com (Shinsegae) | Risk Warning (attempted omnichannel but failed) | Tried rapid delivery in Seoul, unable to succeed due to reliance on store inventory |
| Farfetch | Neutral (recent acquisition) | $500 million acquisition, $4 billion GMV, potentially used to strengthen apparel category |
| CJ Logistics | Background Mention | Naver's logistics partner (Naver Shipping Alliance) |
| SoftBank | Background Mention | Invested $1 billion |
| Sequoia | Background Mention | Early-stage investor |
1. Drew Cohen: Coupang's "Meta-Optimization" Framework — The company restructured its entire business model around six consumer preferences (delivery speed, reliability, trust, ease of ordering, selection, price), rather than optimizing just one or two dimensions. "You can't do this half-heartedly; you have to get all of them right."
2. Drew Cohen: Latecomers Can Win by Identifying Unmet Preferences and Full-Stack Optimization — Coupang entered the e-commerce market nearly 20 years later than many competitors, but redefined the rules of competition by identifying the gap in "trust and reliability."
3. Drew Cohen: Coupang's Flywheel Starts with Warehouses, Not Customers — "If you want to compete, you first need to build warehouses, stock inventory, take a 1P approach, secure favorable procurement terms, attract customers, build habits, generate data, optimize logistics... each step is interdependent."
4. Drew Cohen: The Threat from Naver Is Overestimated Because Coordinating Multiple Stakeholders Is Harder Than Single-Entity Optimization — "Coupang can optimize for just a few variables, while Naver needs to balance the different incentives of all its logistics partners."
5. Drew Cohen: Advertising Is Coupang's Underestimated Profit Engine — Merchants only began complaining about increased ad placements in recent years, indicating that ad monetization is still in its early stages; long-term ad revenue could reach 4%-5% of GMV.
6. Drew Cohen: Coupang's "Bees and Flowers" Analogy — "Merchants are bees, customers are flowers. Bees always go where there are more flowers." The platform that controls traffic holds real bargaining power, and merchant protests have never affected sales.
7. Drew Cohen: Coupang's Cash Flow Characteristics Are Superior to Amazon's at the Same Stage — Every $1 of inventory corresponds to $3 of accounts payable, inventory turns nearly 10 times per year, and it achieved $1.5 billion in free cash flow just 10 years after the transformation — much faster than Amazon.
8. Drew Cohen: The Contrast Between Failure in Japan and Success in Taiwan Validates the "Full-Stack vs. Lightweight" Difference — The lightweight instant delivery attempt in Japan failed, while replicating Korea's full-chain model in Taiwan succeeded, indicating that Coupang's core capabilities are indivisible.