Pinduoduo (PDD) launched in 2015, which is absurdly late by Chinese e-commerce standards. Alibaba dates to 1999, and the other big competitor, JD, is even a year older than that. So just like Sea Limited, a company we covered a couple of weeks ago, PDD doesn’t fit the standard narrative that new entrants have essentially no chance at real success once one or two incumbents exist and reach logistics scale.

That thesis might still hold for Europe and the US, where most consumers are okay with paying up for quality and comfort, which means there is no wide gap beneath Amazon for someone cheaper to fill. In China and many emerging markets in the world, that’s different, though.

Colin Huang, PDD’s founder, saw a market that Alibaba and JD weren’t serving. They basically spent fifteen years building for Tier 1 and Tier 2 cities and largely ignored Tier 3 and below. And a “lower-tier city” in China in 2026 is not a village.

As I told Shawn on the show, I planned on visiting one of those Tier 3 cities with a friend of mine whose parents grew up there and still have a house there. He said it was one of the “smaller” cities. Turns out, it has about 12 million citizens. For context, so-called lower-tier Chinese cities reflect something like 70% of the population and well over half of GDP.

The other big tailwind was mobile adoption in the early 2010s. Cheap Android phones with Alipay and WeChat Pay built in flooded the lower-tier cities, so hundreds of millions of people who’d never shopped online suddenly could.

But Alibaba and JD’s marketplaces didn’t offer what most of them wanted. These new customers were intensely value-conscious, willing to put in the effort to save money, and they treated shopping as a way to pass time rather than a quick transaction – essentially a mall, just online.

This is where I’ll once again bore you with my e-commerce ladder framework. As a quick refresher, I tend to look at e-commerce companies through the lens of three different levels of development. Level 3 marketplaces cater to users with clear purchase intent who prioritize speed, quality, and service over price. Amazon, Coupang, and Meli are such marketplaces. They also tend to be vertically integrated and run massive logistics ecosystems, as well as payment and other value-adding operations.

Level 2 is the middle rung where the shopper has a general intent to browse but doesn’t yet know what exactly they want, and cares more about price than speed or quality. And then there’s level 1: more or less pure impulse, with no initial purchase intent when the app is opened, and discovery mostly through content. Think of TikTok Shop. Most people don’t go on TikTok to shop, but they stop on a video, like one of the products, click, and boom, you’ve bought a new shirt.

Part of why this is easy to do is because it’s so cheap! Price is arguably the most important factor.

In the beginning, PDD was unambiguously Level 1. It was defined by value-conscious buyers, gained traction through gaming and discounts, and had no ecosystem beyond the marketplace (i.e., payments or logistics). But that’s how Level 1 players can optimize for their target audience.

PDD reaches those buyers through a supply-chain model called C2M – consumer-to-manufacturer – built on top of something called team-buying. I’ll explain both properly in a moment. For now, the important part is that the model gave PDD a structural price advantage over Alibaba and JD, and it worked immediately. Just three years after its founding, in 2018, PDD went public on Nasdaq and surpassed Alibaba in annual active buyers.

The most difficult thing for me, looking at PDD, was how little insight we have into the business and numbers. PDD is probably the most secretive company I have ever looked at. PDD doesn’t give guidance; they didn’t have a CEO for years; they have a ~$75 billion cash pile (including short-term investments) that sits on the balance sheet unused; and when things go very well (as in 2024), there’s a good chance you’ll tune into an earnings call to hear them sounding off as quite bearish.

This culture was set by its founder, Colin Huang (or Huang Zheng). He was an early Google China engineer and had already built an electronics site and a gaming company before Pinduoduo. And in 2006, he had lunch with Warren Buffett, who was a huge inspiration for him. He was only 26 at the time and was brought by his mentor Duan Yongping – the man behind BBK Electronics and a legend in Chinese business.

But in 2020, Huang stepped down as CEO of PDD. While precise details are not known, the timing lined up with China’s broader crackdown on tech founders. This was right around when Jack Ma vanished from public view. So this might have been a de-risking move to lower the company’s political profile and get the famous founder out of the crosshairs. He’s still by far the largest shareholder, though, owning north of 30% of the company, so I’m fairly confident the long-term vision is still his.

Day to day, the business now runs under two co-CEOs: Lei Chen, who runs Temu and the whole international, technical, and regulatory side that comes with it, and Jiazhen Zhao, who built Duoduo Grocery and runs the agriculture and supply-chain side of the company.

While we lack many details, we should still dive into the business units and analyze what we have. On that note, roughly half of revenue is from the core China marketplace, a third is Temu (the international e-commerce arm), and the remaining ~15% is Duoduo Grocery.

The China marketplace was built on the C2M model I teased earlier. Team-buying lets you unlock a deeper discount by forming a temporary group to buy in bulk. That group can just be you and your friends or family, but you can also team up with a group of strangers via a link shared on WeChat or other socials, with a 24-hour window to fill the group.

Sellers on PDD’s marketplace usually list two prices for the same item – a higher one for individual shoppers and a lower one for the team buy. And the pitch to the manufacturer is that a bulk order gets a better price, with the savings flowing back to the group. The whole idea is to capture scale economics on the smallest possible scale. There’s also a network effect here that competitors don’t have. I don’t directly benefit when someone else buys anything on Amazon, but through team-buying, I very much do benefit on PDD when someone else buys something. So I want to tell friends and pull them into these groups.

The advantage for the factory is that it knows exactly how many units to make, so there’s no unsold inventory. And because the products are white-label, there’s no brand marketing, no distributor, no middlemen. And logistics aren’t PDD’s job either – the merchant ships straight to the buyer with a third-party courier.

The grocery business is the smallest of the three, so I won’t spend too much time on it, but it’s vital to PDD and was one of the earliest ventures. It runs on the same aggregation trick as team-buying, just applied to a neighborhood. Someone local becomes the “community leader” – often a local shop owner – and runs a group chat where members pool their orders. Those orders go to the merchant, and the next day everything is trucked to a single pickup point, usually the leader’s shop, where people collect.

The genius of it is that it’s pickup, not delivery. PDD’s customers don’t demand the luxury of home delivery. For a lower price, they’ll happily collect it themselves. And that matters, because fresh groceries are the hardest thing in all of e-commerce to make money on. Home delivery of a five-dollar bag of vegetables is a guaranteed money-loser, but pickup group-buying is one of the few models that actually solves it.

Temu was PDD’s attempt – a very successful one, at least at first – to take the factory model international. It launched in the US in 2022 and spread from there into the entire world: Europe, South America, South Africa, and so on. The same white-label factories that shipped to Chinese consumers now sell directly to consumers in the US or Europe. I still remember when it “hit” Germany. It came out of nowhere and was suddenly everywhere, mostly because you could buy an enormous amount of cheap clothing.

Since those changes, Europe accounts for around 40% of Temu’s GMV, its biggest market, but the EU is moving forward with its own de minimis removal, with the Digital Services Act bearing down on product safety while the FTC is poking at it too. Despite all that, many estimates still expect an inflection to profitability, around 2028 to 2029. Watching this will be an important part of the thesis.

China is the most competitive e-commerce market in the world, so it’s astonishing what margins PDD has put up. Part of the truth is that the government stepped in and told players to stop the subsidy war, so it took some intervention. But there’s another reason so many players coexist without destroying one another – the market’s scenario-based nature.

PDD, for example, optimizes for conversion, while Alibaba and JD.com optimize for the expected value of a purchase. Say JD figures there’s a 10% chance a customer buys a $1,000 iPhone if it’s recommended, versus a near-certain chance of a $5 pack of toilet paper. JD recommends the iPhone, because the expected value is higher. PDD does the opposite – it goes for the toilet paper. It prioritizes making the sale and building the habit of shopping on the platform. Especially when it comes to essential goods. You buy toilet paper every month, so that’s the conversion PDD wants – everyday goods, high frequency. And it’s working. Order frequency has gone from 15–20 times a year early on to more than 70 times today. Alibaba averages around 90, so there’s still room to grow.

Now, onto how PDD actually books revenue. It reports two lines – “online marketing services and others” and “transaction services” – split roughly 50/50. Online marketing services is the classic third-party-marketplace business: transaction fees plus merchant advertising, with most of the revenue and profit coming from advertising. The advertising take rate alone is estimated at around 4% to 4.5%. And I’ll keep saying “estimated,” because PDD gives us so little to work with.

The trend is what matters, though, and that one looks promising. In 2019, the advertising take rate was closer to 2.5%, so it’s roughly doubled, and I expect it to keep climbing.

For years, PDD acquired users so efficiently that the buyer side grew much faster than the merchant side. So you had a wall of demand chasing a thin set of merchants, and in that world, a merchant doesn’t need to advertise, because demand finds them anyway. But as the merchant base expanded, merchants started competing for the same users, bidding against each other for limited ad space. So the rising take rate isn’t PDD jacking up prices for ad space – it’s merchant-versus-merchant competition for a finite amount of attention. That’s the fairer, more durable way to monetize.

For the longest time, the Chinese fight was among the big three – Alibaba, JD, and PDD – who together control something like 70% of Chinese e-commerce GMV, with much of PDD’s gains coming straight out of the incumbents’ share.