*Please see our disclaimers at the bottom of the page and our full disclaimer here. By using our website, you acknowledge you have read our disclaimers and agree to our terms of use.

**Thank you to all of our subscribers for supporting DJY Research! We hope you enjoy our JD Deep Dive! Just incase you are curious, the founder of JD’s name, Liu Qiangdong, is pronounced (lee-oh chee-ong-dong). Despite this being a rather long report, JD is such a large company that there is still plenty more to talk about! Feel free to reach out and ask us questions in our Discord if there is something you want to learn more about!

Founding History.

Liu Qiangdong was born in 1974 in a rural village to an enterprising family that used to transport goods along the Yangzi River before they were forcibly relocated and lost everything in the 1949 Cultural Revolution. Their village had neither water nor electricity nor roads, and they only ate meat twice a year. His dad was an accountant in the “production brigade”, tabulating farm produce, but in 1979 a local government program allowed limited borrowing and they purchased a freight boat to resume their family trade of transporting goods. While this brought in some more income compared to the other rural peasants, it also meant that his parents were often out working, leaving Liu to be raised by his grandmother. Despite their improved income, Liu speaks of eating nothing but permutations of potatoes for half of the year and corn for the other half. In 1992, Liu was one of the top scorers of the Gaokao in the Jiangsu Province, which gained him acceptance into Beijing’s Renmin University. However, his parents were too poor to spare the RMB 500 required to send him there, leading to the entire town pitching in to help him. With RMB 500 and 76 eggs, he was off to Beijing.

Liu, who adopted the English name Richard, majored in sociology. However, his real education came in the form of self-taught programming. As one of the few people capable of coding right at the birth of the internet, his part-time moonlighting earned him substantial wages, and he bought his family a new home, along with a Motorola cell phone for himself, which cost a magnitude more than the Beijing travel fare he had to borrow just a few years ago. Feeling flush, when a popular restaurant he liked to frequent was receptive to selling itself, Richard kicked in all of his savings and borrowed the remaining RMB 200k required to buy it. Aware that he had no idea how to manage a restaurant, he thought the best method would to be as hands-off as possible. In turn, the staff colluded to steal from him, inflating purchase orders for kickbacks and misreporting receipts until he had to shut it down less than a year later. To pay off his debts and reset, he worked at a health products company called Japan Life for two years, and ultimately learned that his restaurant failure was the result of not having a management structure, clear procedures, or financial oversight, a lesson he made sure to never forget. After paying off his debts and with RMB 12k in his bank, he was ready to start his next business venture.

Background History.

On June 18th, 1998 (a date that today is best known for JD’s massive 618 sales), at just 24 years old, Richard rented a small 4sqm stall in Haikai Market in Beijing’s technology hub of Zhongguancun and started selling magneto-optical products (hard drives and CDs). He called it JingDong Multimedia, or JD Multimedia, by combining his name (Dong) with his then-girlfriend’s (Jing), who he awkwardly would not ultimately marry. Whereas most of his peers sold products that were often fake and haggled over for each yuan, Richard set JD Multimedia apart by putting price labels on everything and only selling authentic products with official receipts (a practice counterfeit sellers couldn’t mimic for fear of reprisal). Price labels allowed customers to know that they were getting the same deal as everyone else, and lowered their fears of being ripped off. While his prices weren’t the lowest a customer could find in the Haikai Market, the authenticity guarantee carried weight. Richard would often tell his customers, “If you are sure those other products are authentic, you should go ahead and buy those”. Not wanting to gamble, the customers would usually end up paying the slight premium for JD’s authentic products, especially since data storage devices were not products that they wanted or could risk failing on them. While it started as a one-man operation selling a limited array of magneto-optical products, within 5 years JD Multimedia grew to 12 stores that sold a wide assortment of electronic products and was earning RMB 10mn. His goals at the time were to have 200 stores selling IT products within a decade.

A JD Multimedia counter that sold mostly magneto-optical products.

But this all changed in 2003, when the SARS epidemic hit China, and JD Multimedia had to close down all of their stores. An employee suggested they try selling their inventory online, and they began posting merchandise on various online forums. Showing how building consumer trust can repay itself in unintended ways, customers who were otherwise reluctant to transact online were comforted by seeing the seller was a known trustworthy retailer and unlikely to swindle them. Seeing some success with internet sales, they launched their own website in 2004, called JDlaser.com. When the SARS epidemic ameliorated, they ran their online store in parallel with their physical retail stores until Richard decided that the better inventory control, quicker price change responses, and wider distribution made the internet a better channel. In 2005, he made the decision to close all of their stores and focus entirely on the online opportunity, despite it meaning JD’s 2005 sales would be half of what they were the year prior. Innovator’s Dilemma evaded.

JD’s first website was launched in 2004 and called JDlaser.com

JD continued to focus on the 3C category (computers, communication, and consumer electronics) initially, which had the benefit of having a more limited SKU set with fewer relevant brands (relative to FMCG) and also higher ASPs that would allow the consumer to rationalize the shipping costs easier. In the early days, Richard would push his procurement team to get the best deals possible on products, even if it meant that they would have to put in large purchase orders. When the purchasing team wanted to buy 200 units of something, Richard would tell them to buy 10k so the cost would drop and they could spur more sales by lowering the price. In a telling example, a computer display that sold for RMB 649 and cost RMB 600 could be purchased at a greater scale for RMB 550 and sold for RMB 599, which gave customers a better deal while giving JD the same unit gross profit. The only way this model would work, though, was if they could grow sales velocity, so inventory costs were manageable. While ordering products in quantities above their current customers’ needs in hopes the lower price would stimulate new demand was a risky proposition–as they could be stuck with more inventory than they could sell, which would incur warehouse storage costs, tie up working capital, and risk inventory obsolescense–it worked time and time again, and became JD’s formula. In 2007, they changed the website name to 360buy.com, and the company renamed itself to JD Mall, foreshadowing a future desire to move beyond just electronics.

JD changed their website name to 360buy.com in 2007.

During these years, they faced cutthroat competition–not from Alibaba’s leading C2C site Taobao (whose reputation for being flooded with fake goods turned off many potential customers)–but from online retailers like the China-native Dangdang and US-based Newegg and Amazon, as well as offline retail chains like Suning and GOME. Employees would wade through competitors’ product listings around the clock, adjusting JD’s listings to always be in line or cheaper than their competitors’. Everyone knew that the strongest competitive position would ultimately come from being the most scaled player, since that would allow them to place the largest orders with suppliers, thus reducing their costs of goods sold, which in turn would allow them to increase demand by lowering their prices. So to build that scale, each retailer was willing to give up margin for the volume that would ensure longterm viability. In October of 2006, with RMB 60mn in revenue (growing 10% m/m with no advertising spend) and 50 employees, Richard looked to raise $2mn to further scale their operations as customer demand inflected. He met Capital Today’s Kathy Xu, who insisted he take $10mn, introduced him to other investors, and with whom he mutually agreed to only focus on growth for the next four years and not worry about making profits until the fifth year. Sales grew rapidly– JD increased revenues more than 3.5x to RMB 360mn in 2007.

Richard Liu (similar to Coupang founder Bom Kim), noticed that over half of all customer complaints related to shipping delays and damaged packages, so he started to set his sights on not just handling the warehousing and fulfillment of online deliveries, but also the last mile logistics capabilities (which was years before Amazon would start to do the same, as China had no reliable national logistics providers like FedEx or UPS in the US). In controlling the full logistics chain, JD hoped to ensure the best and quickest delivery experience, which would also allow them to accept COD (cash on delivery), which was still popular at the time (most 3rd party logistics providers wouldn’t provide this service–or if they did, they commonly stole the money or at best remitted payment back very slowly). This capability would be a differentiating factor vs. the competition who were admant that logistics was too capital intensive and would be too risky. They weren’t wrong–at the outset, Richard estimated that it would cost at least $1bn to build last mile capabilities and would require 2k orders/day in most cities to break even–when they were averaging around just 20 orders/day. Around this time, in order to help spur the orders required to support their logistics network, they launched general merchandise categories with the aim of eventually selling everything. Sales increased over +250% y/y to RMB 1.3bn in 2008.

However, despite the strong sales growth, the financial crisis hit JD hard, and Richard had to scramble for more funds. Many suppliers began to demand more stringent payment terms, which drew on their cash flows. JD’s valuation dropped from $150mn to a low of $30mn. Capital Today made several bridge loans during this time, but JD was still often operating with only a few days of runway and under constant pressure of missing payroll. However, the stress this period inflicted on JD also neutered some of the competition. After successfully pulling through the crisis, JD earned a $265mn investment from Zhang Lei’s Hillhouse Capital at a $1bn valuation in 2010. In a telling show of Richard’s character, he insisted they get audited statements at the company’s cost so the investors knew there was nothing dishonest about their figures. This investment gave them a war chest to take share from other leading e-retailers right at the time competitor Dangdang IPOed on the Nasdaq, raising ~$272mn.

In 2010, JD also launched a “book war” against Dangdang and Amazon, who had close to a duopoly in the online book market (Taobao had a sizeable presence with used books). They launched aggressive promotions, giving free books out to their most loyal customers (Diamond users who spent more than RMB 30k) and steep discounts to everyone else. Within a year, Dangdang’s gross margins dropped 600bps to 14% and their stock dropped over 70%. Gaining share in the book category was also helpful to JD since most of the oldest ecommerce users were book buyers who first used Dangdang or Amazon who learned about JD for the first time through these promotions. JD, which was run very leanly and efficiently (for instance, they figured out six box sizes was the optimized number between increased logistics complexity and wasteful costs from mix-matching item size to box), was able to stomach the losses better and Dangdang was effectively defanged. Amazon would also soon decide its resources were better utilized in other markets.

A similar war would play out a few years later in appliances (yes, they shipped huge items like air conditioners, refrigerators, and washing machines) against the big box retailers Suning and GOME. Ultimately, JD, with fewer distribution layers and lower inventory cost, coupled with the advent of JD Service Stores in remote locations (local authorized dealers that provided delivery, installation, and later maintenance, but where JD would still be the product seller), gained material market share. Winning this category also meant building out robust logistics networks with a “bulky supply chain system” capable of handling large and heavy goods like stoves and ovens.

A picture from JD’s promotional material showcasing their bulky logistics chain.

A few years prior, back in 2010, JD started opening up their platform to 3rd party sellers with their online marketplace. Similar to Amazon’s FBA, JD started providing sellers logistics services, which increased the volume running through their network and helped it scale more effectively. They also started developing services for sellers like advertising and financing, breaking out their finance unit as an independent business group in 2013. Also in 2013, JD Mall made another name change, dropping the “Mall” and moving their online website to the doman JD.com from 360buy.com. They also announced RMB 100bn in GMV and prepared for an IPO. A few months before their IPO, JD announced a strategic investment from Tencent, which was a game changer for them that greatly increased their reach to the Chinese consumer. In March of 2014, Tencent agreed to invest ~215mn and give JD their ecommerce assets (Paipai among them) in exchange for a 15% stake as well as a promise to subscribe at the IPO price to an additional 5%. JD IPOed on the Nasdaq in May of 2014, raising ~$1.8bn at a share price of $19, giving them ample liquidity to continue to invest in logistics and take advantage of their newfound partnership with Tencent.

In the years that followed, JD would add over half a billion new customers, going from 38mn in 2Q14 to 552mn in 3Q21. On the growth of their ecommerce platform, they were able to create new businesses that would later become their standalone companies with significant outside ownership or having gone public. JD Health, JD Technology, JD Logistics, JD Auto, JD Properties, and JD Industrial Technology (MRO), are just a handful of them–with others currently being incubated. Richard Liu continues to run JD today with a ~14% stake in the company (reduced after a recent charitable donation).


As it is probably clear by now, JD is an online retailer that facilitates the sale of their own first party goods and 3rd party merchants’ products through their online platform that is vertically integrated with their own logistics solutions, giving consumers the fastest delivery and strong customer service on a large selection of quality goods (over 9mn 1p SKUs).

87% of their revenues come from their direct product sales segment (1P) with the remainder attributed to Marketplace, Marketing, and Logistics Service revenue. Given the accounting of reporting 1P revenues at the full cost of the item vs 3P where only the commission is reported, this understates the importance of their marketplace business (see our SE piece for an accounting walkthrough of this). JD has not broken out 1P vs 3P GMV since 2016, where it was about 57% 1P and 43% 3P. We figure 1P is still higher than 3P, but the gap is likely less. Also last disclosed in 2016 was their GMV mix of Electronics & Home Appliances vs General Merchandise, which was split exactly 50/50. However, from 2013 to 2016, General Merchandise as a % of GMV grew 1400bps and it is likely it today represents a larger portion of GMV than Electronics & Home Appliances. However, this difference does not show up in the segment revenue mix since JD overindexes in Electronics & Home Appliances 1P (>60% of 1P). Below, you can see that 1P revenues from Electronics & Home Appliances and General Merchandise are 54% and 34%, respectively. Marketplace & Marketing revenues are at 7%, slightly higher than logistics sedrvices at 5%.

Revenue Segment Definitions:

Product Revenues. These are all 1P transactions (whereby JD is acting as the principal and responsible for fulfilling the promise to provide the sold goods). They further split up product revenues between Electronics & Home Appliance and General Merchandise (a byproduct of having started as an Electronics retailer). It is worth noting that for all 1P transactions, revenues are recorded net of value-added taxes (VAT), but the VAT would be counted in GMV (which is industry standard accounting).

Electronic and Home Appliances. This includes sale of 3C (computers, communication, and consumer electronics) as well as Home Appliances (including large items like refrigerators, air conditioners, ovens).

General Merchandise. This includes everything not listed above, but they specifically call out the following categories in their annual report: food & beverages, fresh produce, baby & maternity products, furniture & household goods, cosmetics & personal care items, pharmaceuticals & healthcare products, books, automobile accessories, apparel & footware, bags & jewelry. This is also all 1P.

Net Services Revenues. This is further broken down into two segments.

Marketplace and Marketing. These are revenues from charging 3rd party merchants commission fees that are either negotiated or a fixed rate, but 2-8% is the common seller fee range. Their 3P marketplace business is also referred to as POP, which stands for “Platform Open Plan”. Additionally, JD provides advertising services that allow sellers to place ads across their website and apps, as well as affiliated internet properties through their ad network. Their two categories of ad products are performance (pay for clicks or on the basis of a successful transaction) and display ads (more for brand advertising).

Logistics and Other Services. JD allows their 3rd party sellers to utilize their fulfillment and logistics infrastructure, as well as other sellers not on JD’s platform. The revenues generated here are from providing warehousing, distribution, freight, and express delivery services. Included in the “Other Services” is JD Plus, their membership program that provides users various benefits (more on this later).

It’s worth noting their revenue segments shown below are different than their operating segments (JD Retail, JD Logistics, and New Businesses) which we will discuss later.

As seen above, LTM revenues are RMB 900bn ($140bn) and have grown ~8x since 2014 for a 36 CAGR%. Net Product Revenues have grown similarly at a 34% CAGR since 2014, whereas Services revenues have grown at a 55% CAGR. Today, Service revenues are growing at a >50% clip (9M21 y/y) whereas product revenues are +26% (9M21 y/y). Within Service revenues, we see that Logistics & Other Services has been growing faster than Marketplace & Marketing at 72% vs 25% for 2020, respectively (they do not disclose this breakout quarterly).

Above, we show that GMV has reached RMB ~2.6tn (~$410bn) for 2020 and as of 3Q21, active customers (at least one purchase in the prior 12 months) have reached 552mn. We will momentarily say a lot more about the GMV, how it is calculated, and frankly how it is very misleading (which isn’t a problem secluded to JD, but given their prior disclosures we can actually get a sense of the magnitude of GMV inflation. Still, this should have been emphasized more in out BABA piece). At a high level though, GMV/customer has been increasing to almost ~double over the past 7 years (the 2013 and 2014 GMV figures are prior to the definition change as we detail below). Note that GMV/customer growth can look lumpy when customer growth is very strong, as new buyers are not contributing a full year of purchase activity. Additionally, new buyer spend is less than seasoned buyers, but over time, returning buyers tend to spend more on the platform, especially as JD has continued to add more categories.

As you can see above, it looks like Covid-19 has been a boon to JD’s business, adding 190mn new customers, or almost 2/3rds of their total customer base at the end of 2018 in under 3 years. Recall Covid-19 started appearing in December 2019 for China, and in 4Q19 they added 27mn new customers (vs 11mn and 13mn for the two quarters prior). However, it turns out that a lot of this buyer increase can actually be attributed to them launching a money-for-value app under a separate brand, called Jingxi.

From 1Q18 to 3Q19, customer growth was very lackluster, with only 30mn users added in that whole period. However, it has been reinvigorated by Jingxi, Covid tailwinds, and more recently, their supermarket initiatives. Since both happened right around the same time (4Q19 Jingxi launched and first lockdown in 1Q20), it is hard to parse out what growth to attribute to each. Management’s call commentary stated over 80% of new users coming from lower tier cities might make you think it was mostly Jingxi, but Covid could have all the same been the impetus to make them another ecommerce platform, especially since JD’s logistics operation allowed them to deliver orders with far fewer disruptions than Alibaba and others.

JD delivers packages during the Covid outbreak. They experienced fewer disruptions than competitors.

There are 3 ways to buy something on JD: 1) their website, 2) their mobile app, 3) their Weixin mini program. The experience is very similar to other ecommerce sites, with a consumer able to complete a purchase in just 3 clicks after finding the item they want.

Swipe through to get a sense of JD’s mini program. The tabs are: 1) home page, 2) categories, 3) nearby, 4) shopping cart, 5) profile.

JD also offers free shipping on all orders over RMB 99 (~$16), but charges RMB 6 for orders under RMB 99 and RMB 8 for orders below RMB 49. Though with their loyalty program JD Plus, members can get free shipping on items under RMB 99 (but it is oddly limited to only 5 vouchers/month, which seems like an unnecessary stipulation for your best customers). JD Plus members also get discounts and access to monthly Plus Day sales. They also have other partnerships with a variety of providers that give members discounts and benefits at 15k+ hotels, free bundles with Tencent’s QQ Music and Bilibili’s premium video subscription, as well as limited promos like access to Baidu’s IQiyi video at no extra cost. Lastly, they have loyalty points, called “JD Beans” that users accrue with platform activity and purchases, which can be redeemed for various discounts. While every user can accumulate these JD Beans, Plus members get ~10x the points, which is another reason for ardent users to join the program. Similar to Prime, JD Plus members spend >2x non-Plus members. (They last disclosed 20mn Plus members in 2020, but it is likely now closer to 30mn as they noted it was growing 30% y/y.)

There are a couple things that are unique about JD, the first of which is somewhat a novelty, which is JD’s Luxury Express delivery service. Whereas orders placed are usually set to their default shipping option, which is JD’s inhouse logistics operation that delivers >90% of orders in 1 day or less (faster and more consistent than alternatives), there is an option for the “Luxury White Glove” service. As the name suggests, JD will send a delivery personnel in a suit with white gloves to delivery your package in a fancy box. While the vast majority of users will never use it, this service is popular with ultra-high-end brands like Audemars Piguet, who partnered with JD to not only provide the delivery, but also to set up their pop-up store on Weixin.

JD’s Luxury White Glove Delivery service, which is usually utilized for ultra-high-end goods like expensive handbags and watches.

Audemars Piguet utilized JD’s Kepler Mini Program Solution to create a mini program on Tencent’s Weixin, which allows them to have a presence on Weixin to access their >1bn users, while still reaping the benefits of JD’s retail infrastructure. The Kepler program allows a seller to copy their entire JD store, including product SKUs and descriptions, into a Weixin mini program. The merchant not only benefits from easier access to Weixin’s audience and a more sharable format between their users, but is now also searchable within Weixin (and if you recall from our Tencent piece, more search is moving within the Weixin app with general search player Baidu becoming increasingly irrelevant). Tencent benefits because JD already had relationships with many brands, and this program helps move more activity onto their platform, which increases the functionality of Weixin and also creates more utility for Weixin Pay (the exclusive payment mechanism on Weixin at the time). JD benefits because they take a seller fee on the merchants’ sales and presumably, they could sell more when accessing Weixin’s audience than otherwise. This is part of the shift to “Retailing as a Service” that JD is pushing, whereby they enable the merchant however they want to sell, irrespective of whether the consumer enters directly through a JD storefront. The Retailing as a Service strategy allows a brand to control more of the consumer experience, while JD operates the infrastructure including warehousing & fulfillment, last mile delivery, returns, customer service, as well as financing and billing (more on this strategy later). This seems like a win-win-win for Tencent, JD, and the merchant, and it is this sort of beneficial arrangement that Tencent originally conceived when they invested in JD.

Audemars Piguet’s CEO Francois-Henry Bennahmias and JD.com founder Richard Liu take a photo for promotional material announcing the partnership.

When Tencent invested in JD in 2014, they not only provided capital and their C2C marketplace Paipai, but they also critically helped JD with preferred placement in Weixin. There are “Level 1 Entry Points” and “Level 2 Entry Points” that Tencent can grant a company or service, and JD received both. Starting in 2014, a user who opened the “Discovery” section of Weixin or QQ Mobile (see our Tencent writeup to see these) would have an option to directly click on JD to open their mini program (this is also why this is referred to as “direct access”). This priority placement in the app meant that it was unavoidable for the hundreds of millions of people who opened Weixin dozens of times a day to not see JD. Level 2 access is in the Weixin Pay page, and shows up any time a user goes to their QR code to pay for something or check out financial services. This was invaluable promotion for JD. The year before they were promoted in Weixin, they had 47mn customers, which was up ~18mn or +61% y/y. The year following the introduction of the partnership, customer growth accelerated to +104% y/y with them adding ~49mn new customers—more than their entire customer base that they grew over the preceding decade. Following the 2nd year on Weixin/QQ Mobile, they grew customers +71%, still faster than prior to the Tencent relationship, even with a much larger base of users. Simply placing JD’s store link in Tencent’s app created enormous value for JD (and Tencent’s JD investment) by giving them a ton of traffic for free (technically valued in their deal).

QQ Mobile Level 1 Entry Point or “direct access” gives JD priority placement that virtually ensures every app user will eventually stumple upon their app.

JD has entered other partnerships since then, including one with Walmart that started in 2016. Under the initial partnership, JD acquired Yihaodian, Walmart’s online marketplace, but Walmart retained the Yihaodian direct sales business and started operating it as a store on the JD.com platform. This gave JD customers a marked increase in selection of high-quality goods from a trusted seller including a critical category—grocery (recall their General Merchandise push was hardly 2 years old at this point). Walmart would also become a preferred seller on JD’s O2O platform (quick local delivery) and receive a 5% stake in JD. A year later, they announced an inventory integration whereby JD customers’ orders could be fulfilled from a JD warehouse or Walmart store based on what is most logistically efficient. They also set up JD Home stores with Walmarts that primarily sold appliances and electronics (during their appliance war with Suning, GOME, and Alibaba). This “store within a store” allowed them to expand a physical retail footprint that could showcase products before customers purchased them and could accommodate returns and service requests.

A Walmart with a JD-branded section within the store.
JD-branded promo plastered on the outside of a Walmart. Walmart participates in JD’s 618 sale, which is named for June 18th, the founding date of JD.

It isn’t clear how many stores JD ultimately built in Walmarts across the country, but it is clear that they have a significant physical footprint today, and the vast majority of the stores are stand-alone with them significantly expanding product selection. Over the next several years, they would open up over 15k JD Home stores, virtually all of which are franchised out. The stores utilize digital price tags, customer tracking systems, facial recognition technology, among many other technologies, to feed an algorithm that helps inform what product SKUs to stock.

In addition to the JD Home format, they have started experimenting with many other formats.

JD Convenience Stores. This is a smaller store format that sells basic CPG items and food. They have intended to open (or rather, franchise) 1mn of these stores. They also have technological integrations with the rest of JD’s selling ecosystem.

JD Mini-Convenience Stores. These are even smaller than the above stores and are unmanned, relying on technology (facial recognition, video tracking, sensors) to charge a customer.

JD Mall. The JD Mall is a huge, 40k sqm physical space that houses over 150 brands and carries 200k+ products. All inventory integrates with their online channels so a consumer could have anything delivered to them rapidly. More unique though, is the space they allow brands to display their products.

JD E-Space (Super Experience Stores). These are technology enabled spaces with various “interactive zones” that include beauty services, wine tastings, coffee tastings, and gaming zones. These shopping centers are consumer-electronics-themed, but their products vary across most categories.

7Fresh. These are JD’s supermarkets that they are opening to address the grocery opportunity and tie in with their omnichannel efforts. Alibaba’s Hema supermarkets would be their most direct competitor here.

Other Physical Retail Store Formats. While JD now has ~1mn brick and mortar stores, management has signalled their desire to have 1mn brick and mortar convenience stores alone, and 5mn total brick and mortar stores by the end of 2023. In addition to the store formats listed above, they also have store chains for Tobacco (JD Tobacco), Alcohol (JD Alcohol World), Automobiles (JD Auto – more on this later), and Appliances (Jiangsu 5 Star), among many others.

JD acquired a 46% stake in appliance store chain Jiangsu 5 Star in April 2019.

Connecting the physical stores to their online channels is JD Daojia, their O2O (quick delivery service), which later merged with Dada (but is still a standalone public company that JD now owns 51% of after a $800mn investment). While it is hard to say how much consumers care about getting many goods in <2 hours (usually <1 hour), grocery is the clearest use case and thus JD and Alibaba are focusing here. However, we don’t want to give the illusion that all of this physical retail is simply to support their online businesses—the high penetration of ecommerce vs the rest of the world is largely a byproduct of the fact that the offline retailers never fully developed. Some estimates put China’s total physical retail square footage per capital at 1/10th the amount in the US. The whole array of specialty chain retailers like Home Depot and Lowes for home improvement or AutoZone, Advance Auto Parts, Pep Boys, and O’Reilly’s just for aftermarket auto parts & accessories never developed in China. While it seems likely the US has too much retail, China clearly has the scope to grow their retail footprint and JD and Alibaba are increasingly moving to fill that need. They are both adding a tech layer to their retail, revisioning retail without the legacy footprint and aged systems most retailers operate with today (Amazon is doing this too, with their Go Store).

JD Daojia Dada quick delivery service.

JD should, in theory, have much higher inventory turnover with lower obsolescence risk given they can coordinate purchases between their huge volume of online orders and their warehouses and stores. However, the store format does not lend itself to quick fulfillment, which is why Alibaba and JD both felt the need to add quick delivery service capabilities (Ele.me / Consumer Local Services for BABA and Dada for JD). In grocery, this could be an important strategic advantage as grocery is high frequency, high total order value, and provides many opportunities to cross-sell, while simultaneously habituating the user to their platform. The grocery TAM in China is estimated to be in excess of $2tn and can grow as more Chinese acquire an affinity for higher end foods. Today, it is highly fragmented with the largest chain, Sun Art (now controlled by Alibaba with a 72% stake), estimated to have a single digit market share (in a 3Q20 earnings call, management said the top 10 supermarket brands only take a 5% share). With a huge, fragmented TAM, JD is pushing to take their share with JD Super, 7Fresh, Dada, and their Walmart partnership.

Slide from a 2020 Dada investor presentation detailing the synergies between JD, Dada, and Walmart.

Grocery stores typically have very thin margins with multiple layers of distribution that mark up the product at each stage and are subject to large losses from food spoilage. Both are problems that JD hopes to be able to solve: 1) utilizing their logistics network to bypass distributors by going directly to the farm to get produce for their supermarkets, 2) using their AI technology to better estimate demand for each product to decrease food spoilage. Additionally, the consumer would receive a higher quality and more consistent product vs. the array of mom-and-pop markets that order from a panoply of different distributors. The strategy shows some signs of working—in 2Q18, they noted that their 7Fresh stores were producing 3-4x the sales/sqft of a traditional supermarket. However, there is very limited other info on their stores and essentially nothing on how the economics of these stores and franchises work.

Other Businesses.

Jingxi. The Jingxi business is a mix of three different businesses. The first is the money for value and social commerce platform, Jingxi (and when we refer to Jingxi throughout the piece, this is what we are referencing. The second is their community group buying app Jingxi Pinpin (referred to as Pinpin hereafter). The third is there convenience store business (referenced above), which is referred to as Jingxitong. We will mostly focus on Jingxi and not Pinpin, since Jingxi is far larger. Just so you know some differences though, Pinpin is more gamified with countdown timers, essentially everything on the platform requiring a group (sometimes as low as 2), and with sharing as a more focal feature. There is also more produce and food on Pinpin. Jingxi is a little less gamified with a more eclectic selection of products. Jingxi has offerings that do not require groups to buy, features more deals, and has livestreaming.

Photos 1 and 2 showcase the Jingxi homepage and a sample selection of goods. Photos 3 and 4 showcase the Pinpin homepage and a sample selection of goods.

This business originally started as JD Pinguo, and was designed to acquire users through group buying with the ultimate aim of bringing them into the greater JD ecosystem (originally designated as a user acquisition department and not an independent business). However, in 2019, when JD and Tencent renewed their 5-year deal, they opted to put the newly-named JIngxi at the Level 1 Access Point in Weixin. It was at this point that Jingxi helped JD reaccelerate customer growth, which had dramatically slowed.

The slowdown in buyers was because their service primarily targeted Tier 1 and Tier 2 cities, with lower tier cities being harder for them to penetrate. Jingxi was designed as a new channel, under a new brand, to expand their user base into the lower tier cities that JD’s value proposition was not a proper match for. Jingxi was very different than the JD.com app with users spending more time exploring, being more spontaneous in purchases, but also, critically, being highly price sensitive. This is the market that Pinduoduo brilliantly attacked early on, and at one point last year, had more buyers than, but since BABA re-entered this market with Taobao Tejia and revamped their Juhuasuan (flash sales), reclaiming their spot as the ecommerce company with the most buyers (today at 867mn vs BABA’s 953mn).

JD was late to focus on this demographic as well, but with Tencent’s Level 1 Entry Point, they were able to quickly acquire a lot of users without the same level of promotion and discounts that PDD and BABA needed (estimated ~250mn users in 2020). Some estimates put Jingxi’s GMV at ~10% of JD’s total, despite users spending less with lower AOVs (average order values).

While it may seem like they are all similar offerings, JD’s positioning is slightly unique as they rely on their preexisting supplier and manufacturer relationships to procure goods, which in theory translates to higher quality. It is also worth considering that they are the only player with a direct sales business (PDD exited theirs), which could help them procure goods at prices that a manufacturer may not be willing to offer otherwise (also recall how JD go started in the early days). Additionally, their logistics network allows them to deliver packages cheaper and more efficiently to customers than BABA and PDD, who rely on 3rd parties. While Jingxi clearly is starting out with good traction, their competitive advantages seem more theoretical with the Chinese internet riddled with complaints of shoddy ordered products that were canceled and disappointing products. There likely is some bias whereas you aren’t going to write a review if you are happy, but it’s worth keeping in mind that Jingxi isn’t just a poor man’s JD, it is through and through a lower quality service (poor customer service is another common complaint). This is why JD bifurcated their brand in the first place when they moved down market. This isn’t problematic per say as the alternatives are all similar and these customers value a low price more than anything else. Thus, they are likely to come back even if there is a poor experience to get another big deal, but the issue is if they are burning money to support this business that may never be unit economic profitable.

Their lack of profitability could come from three main reasons: 1) their customer acquisition costs are too high, 2) their retention rate is too low, 3) each transaction isn’t contribution margin positive. The first issue and second issue are inter-related, with the concern being that they are paying for a user that JD ultimately does not make a positive return on. For instance, if they pay RMB 40 to acquire a user, and that user only makes three purchases with AOVs of RMB 30; with 20% contribution margins (margin after direct variable costs), they lose RMB 22. They can make money by either keeping their acquisition costs under the user’s contribution margin (under RMB 18) or by increasing the user’s retention rate (getting them to order more than three times). With poor consumer experiences, they create churn events where the customer never comes back (like the several reviews we read who swore they would never again shop on Jingxi). However, for this business model, high churn is not totally consequential, as the cost to get that user back in most cases tends to be rather low—a small discount on already cheap goods will probably do it given how price sensitive these users are. While it never seems sane to say that high churn is okay, we just don’t think it is necessarily problematic for this sect of the market. As far as customer acquisition costs (CACs) go, we actually think JD’s CACs are rather rational. One source put it RMB <10, with a popular promotion being small cash rebates of a few RMB after you share a link with 4-5 friends. Management also noted in their call commentary that they were skeptical of some of the subsidies that were going on in the sector. A low CAC means that you do not need to sell many items to recover your cost and make a profit.

This leads us to our 3rd factor, the contribution margin of each order. With low AOVs, that means you need a higher # of orders to make back your marketing cost. The high churn, though, is an obstacle to high frequency. But putting that aside, the low absolute value of each order will make it very hard to make a profit. If customers have an AOV of RMB 30-50 ($4.70-7.80) and if the item was marked up 15-30% (recall Costco marks up items 15%), then that would be a 13-23% gross margin, or RMB 3.9-11.5 ($0.60-1.80) to cover packaging, fulfillment, and last mile delivery, not to mention all of the personnel involved in procuring merchandising. When we checked the platform, most items, despite being only a few RMB, had free shipping. (The uneconomic shipping was likely not included in our sources’ CAC calculations, but it should be.) We don’t have any hard evidence to inform out unit estimates here, but we wanted to illustrate how difficult it is to make money delivering small value packages of discounted goods to fickle customers. At a high level, we can say that it is likely burning money though. Below, we estimate that Jingxi is losing at least ~$100mn/quarter, but could be much larger (total new businesses are burning an average of ~$385mn/quarter in 2021). This run rate loss isn’t terrible in the scope of things, given their Retail business cash flow can sustain it, but we would hope that they do not continue to burn that unless there is a clear path to unit economic profitability with a sustainable consumer base.

This is all to say that this business is unlikely to ever generate meaningful profits relative to JD’s core retail business, but there could be an ecosystem advantage to having it. Suppliers like having this channel option that allows them to quickly sell goods in bulk should they need to unload some obsolete inventory, and the higher package volume helps them continue to refine their network’s efficiency. Lastly, “group leaders” (the ones who pick up and disseminate the goods in group buying), which would not only allow the shop owners to make incremental money, but it also drives traffic to their stores. JD management seems to have their focus on the right factors here, saying on a recent call that they “believe cost efficiency and customer experience are always the key to the longterm success of the retail industry”, which translates into the supply chain and logistics capabilities rather than merely competing through subsidies to expand scale at whatever costs. Assuming they can make the unit economics work, we could actually see JD having an advantage vs. PDD and BABA with their direct supplier relations, more scaled delivery network, as well as their retail foodprint which could act as pickup points. However, make no mistake—this is a much lower quality business than their core.

International. JD has started to operate international operations, mostly in Indonesia, where JD.ID is valued above $1bn. They started a trial operation as early as 2015. With an initial focus on 3C and appliances, now maternal products, food, and FMCG are among their most shopped items. Their value proposition is similar to what it originally was in China a decade prior: fast delivery and authentic products. While they have a sizeable 1P business, they have since added over 30k merchants to sell alongside them. With features like “Nearby Shops” that allow a user to see what stores are nearby and what items they have in stock, they offer something unique for merchants and consumers alike. Likewise, their logistics operation is a big competitive advantage that allows >85% of orders to have same or one day delivery on a wide range of their 1P products (competitors typically take 5-7 days). They have also partnered with Gojek (and received an investment from them), for quick, on-demand delivery for groceries. They already have 18 warehouses and 142 distribution stations in Indonesia, which cover 90% of provinces. They have an estimated 20mn+ buyers today with 350k+ SKUs.

This is one of JD’s stores that primarily focuses on 3C goods.

JD is also expanding with a physical retail footprint and cashier-less stores. This will help them build a brand and draw attention to their online offerings while allowing integration with omnichannel initiatives. They also have started to move into Thailand (JV with Thai conglomerate Central Group), Vietnam (via an investment in Tiki), and The Netherlands (via 2 physical stores). While Shopee, Lazada, and Tokopedia, all currently have a lead on JD, they have a distinct value proposition, and we could see them as a dark horse, gaining a loyal consumer base over time as they address the two weakest points of these other marketplace platforms.

JD cashierless stores in Indonesia.

In additional to going international, they also help cross-border trade with JD World and their Shopify partnership, whereby they let global sellers access the Chinese domestic market through moving their listings onto JD’s platform. Lastly, they have a Google partnership where Google invested $550mn in JD, and will help provide them listings on their Shopping tab, with a focus on developing markets like Southeast Asia. However, they will also make a limited selection available globally, including the US and Europe. We are positive on their efforts to help facilitate cross-border trade, especially into China, and think their efforts in Indonesia and Southeast Asia could be a great longterm growth opportunity. That said, we wouldn’t hold our breath for any inroads in the more mature US and European markets.

JD Logistics. JD Logistics is operated as a separate busines sunit and recently IPOed on the Hong Kong Stock Exchange (HK: 2618), raising $4.5bn to continue to invest in their network and infrastructure. JD retains an ~70% stake in them and consolidates their operations on their financial statements. Their current market value is ~$20bn.

Slide on JD’s logistics capabilities.

JD Logistics is JD’s inhouse logistics platform that they not only opened up to 3rd party sellers to sell on their platform, but also to 3rd parties who sell elsewhere. JD’s logistics network is one of their strongest competitive advantages that enables >90% of orders to be delivered in one day or less, so it may seem odd that they are giving this capability away to everyone. The logic is likely that more package volume increases their network density and thus decreases their cost per package. This allows them to reap cost savings, while they also can earn a return on delivery 3rd party packages. Given their scale, they can still offer low prices and make more profit vs. altnerative options. Furthermore, as Alibaba’s Cainiao stitches together many different 3PLs, over time they may be able to provide a competitive, if not better service than JD. This could mean JD has less of a lock on their sellers, who utilize their fulfillment service, as Cainiao increasingly becomes a better option. To be clear, this is far off from happening, but nevertheless, to get ahead of this, JD is taking on volumes unrelated to their own retail business in order to continue to drive down their cost structure. As they continue to entrench their position as the leading scaled logistics provider, they can offer better delivery to JD customers as a lower cost to themselves, while profiting off of 3rd party volumes.

As JD Logistics is the only global stand-alone logistics operation that grew from an ecommerce player, there are a ton of different insights we can draw that not only are interesting to JD, but other global ecommerce players. However, in order to best capture all of those, we have decided it would be best to do a separate deep dive into JD Logistics as our next report, so stay tuned!

JD Property. Established in 2018, JD Property owns, develops, and manages the logistics facilities that support JD Logistics. JD Property also helps facilitate the sales and leaseback transactions so they can recycle the capital back into other development projects. They have raised two Core Funds so far, whereby outside LP investors provide 80% of the capital in return for a steady return, and JD Property puts in 20% as the GP. Logistics properties are in high demand since they do not have the same tenant risk that commercial properties may have, and investor can diversify into a somewhat different asset class (when retailers and commercial lenders were wrecked by Covid, logistics warehouses were untouched). This strategy allows them to reduce their capital needs while allowing them to build properties to their needs and keeping them in control. They will likely IPO this at some point in the future.

JD Technology. JD Technology was born as JD Finance in 2013, and their first product was Baitiao, which is a credit product that allows users to pay for a good 30 days late, or over a 3-12 month period (basically BNPL). However, Alipay was much more popular, and their Huabei product, which was actually launched after Baitiao, blew them out of the water. JD also created their own payment system, but their market share essentially rounds to 0. They had some success with merchant financing (including Baitiao) to their 3P sellers, but generally speaking (compared to Alibaba and Tencent) they missed the finance opportunity entirely. Recognizing this, they tried to shift their value proposition, attaching AI and Cloud services to their offerings in a confusing array of services. By 2017, they were talking about the “Financial Cloud”, which confuses us as much as it seemed to confuse management, since they never found anyone who needed it. In 2018, it was restructured and renamed to JD Digits to help their “de-financialization” of the unit to broader technology services in part to avoid greater regulatory scrutiny. They had raised $5bn over several years, but struggled to find clear use cases that customers wanted. However, their motley crew of products that served intelligent cities, agribusinesses, financial institutions, and AI services generated enough revenue to put together an IPO slated for April 2021. At the time, they only disclosed 2019 revenues, which were $2.8bn, growing 37% y/y (a 1,600bps deceleration from the year prior, despite a much lower revenue base vs. peers). However, with increased regulatory scrutiny in the financial sector (Alipay IPO being pulled), they pulled their IPO. They restructured again, laid off staff, and purchased JD’s AI and Cloud assets for $2.4bn before changing their name, again, to JD Technology, in an attempt to explicate themselves completely from their original financial services ambitions (however, the vast majority of their revenue reportedly still comes from financial services, BNPL in particular). Lastly, owing to a stipulation in their 2017 fundraising round, if they do not IPO by 2022, they will have to buy the investors stakes back at an 8% annualized rate (something they could be rather ecstatic about if they were successful, as an 8% cost of capital is a rather low hurdle). Thus, they are planning a second attempt at an IPO in 2022, not even a year after restructuring their business. While JD only owns a minority ~42% stake, Richard Liu has majority control.

JD Health. Capitalizing on their consumer base and ecommerce capabilities, JD started connecting with healthcare and pharmacy providers to open up an online “Pharmacy”. While the capability to deliver products to consumers was solidly within their prior competence, there was a clear opportunity to expand into healthcare services like online consultation, prescription renewal, and telemedicine. In similar themes with other business, they saw the desire for omnichannel, so they launched partnerships in over 200 cities to help get a patient to see a doctor in person or a hospital if need be. They have over 109mn active users and are run-rating over >$4bn, but are still operating at a loss. JD IPOed this business in December 2020, raising ~$450mn in the process to continue to penetrate the healthcare opportunity. JD has retained a ~70% stake and today it has a $25bn market cap. We may do a separate deep dive on them if there is interest.

JD Industrial Technologies (MRO). MRO is an acronym for maintenance, repair, and operations, and JD MRO is a B2B service that caters to manufacturers. In a press release, they note that JD MRO is “connecting big companies with industry-leading service providers for manufacture-level omnichannel service solutions”, which sounds like a whole lot of jargon to us. But we essentially think they want to help connect manufactures to service providers so JD can sell them more specialized equipment that would require expertise to install. They note that at first, they will focus on 1) sending teams to a client’s factory to help consult with purchases, 2) offer professional installation and maintenance, 3) provide forwarding warehousing services. While there isn’t a ton of information on this segment, we do know they raised ~$230mn at a >$2bn valuation last year, and are looking to IPO soon.

JD Auto. This is a commerce segment that was carved out to focus exclusively on Automobile services. While they had auto supplies on their JD.com platform for a long time, there was a consumer desire for more auto specific expertise, as well as installation and repair services. To meet this need, JD rolled out the JD Car Club, which allows repair shops to join their network after being screened for quality. This allowed JD to extend their trusted reputation to a field that is usually fraught with exploitation and inconsistent service quality. Those in their Car Club network also get access to JD’s B2B auto parts platform, whereby they can aggregate their demand to get better terms with suppliers, passing off savings in the forms of lower costs to the consumers. Additionally, they link into JD’s CRM systems that allow JD to help control their inventory levels, including auto ordering. JD’s ability to have full visibility into the supply chain will help reduce redundant inventory, saving costs, while keep stock at adequate levels. (This is capability is called “Cloud Match” and something similar is used in other businesses). They go even further by training technicians on car part ordering policy as well as maintenance skills, which helps standardize services across their network. In total, they have a network of over 1,200 maintenance shops (and another 300+ core members who utilize the Cloud Match service). Lending their brand to 3rd parties is a double-edged sword though: it can either be an easy way to create differentiation and provide value to the merchant at zero costs to JD while simultaneously creating more merchant lock, or it can lead to a degradation and dilution of the brand value should their 3rd parties represent them poorly. As long as they keep screening new sellers and closely monitor any potential customer complaints, franchising is a great business.

Note the JD logo on his uniform. JD is benefitting from their reputation as a trusted brand.

Here we are seeing JD’s strategy of incubating these niche selling opportunities in various sectors into full standalone businesses (see what the standalone app looks like below). As we mentioned previously, Chinese retail environment is very underdeveloped despite “being ahead” on ecommerce. These high penetration rates are largely a byproduct of there being no good local alternatives. In the US for instance, as we noted, there are no fewer than 4 auto specialty stores that serve consumers as well as mechanics with a staff that are experts in that domain. This creates a very interesting opportunity for JD to go from general online retailers to one that caters to various specializations via omnichannel. Of course, this is no easy feat, and these new businesses will require very different focuses than their legacy consumer ecommerce business, so it makes sense that JD wants to make these independent businesses with different leadership, outside ownership, and capital. However, JD still remains a majority stakeholder in most of these for now, even after some have IPOed (ie JD Health, JD Logistics). We think this strategy allows a lot of optionality with minimal risk and helps embed JD’s commerce operations very deeply in a plethora of more merchants.

Operating Segments.

Earlier, we went over JD’s revenue segments. However, their operating segments are slightly different than their revenue segments and the way we presented them above. They operate through three segments: 1) JD Retail, 2) JD Logistics, and 3) New Businesses. JD Retail includes their core 1P and 3P business, as well as their seller services including advertising. JD Logistics is their logistics operations including warehousing, line haul, fulfillment, and last mile delivery services. New Businesses include JD Property, Jingxi, JD Health, and their convenience stores (Jingxitong). It is not entirely clear where they place all of their other retail initiatives as they are too immaterial for them to separately break them out, but we believe they are housed in the New Businesses segment.

JD Retail has a 3% operating margin (before unallocated overhead including stock-based comp) or ~$4bn in operating income, which still has room to improve. Management guides longterm retail margins to being similar to other large retail chains like Walmart, who is currently around ~4%. However, they also call out seller services (advertising is especially important) as raising that 1-3 points for a high-single digit steady-state margin. We will talk more about this later in our revenue build. Also note that 1) JD Logistics has operated profitably before, but as they reinvest in the business, they are now burning ~$650mn annually and 2) their New Businesses are run rate losing ~$1.25bn. A quick note on these figures: they are before including RMB 3.9bn (~$615mn) of SBC and amortization LTM. We are okay ignoring the amortization as it is the result of acquiring intangibles from acquisitions and not a real economic cost (although their mixed acquisition history may make you want to account for it), but either way, the SBC represents ~90% of the unallocated amount and applying a majority of that to JD Retail (given it is their biggest operation) would clip their margins ~40bps.

Before we go into China market dynamics and JD’s value proposition vs. the competition, we wanted to touch on some of the accounting behind the metrics we reference. (You may skip to the next section if in-depth accounting discussions aren’t your cup of tea, however it does allow us to back into some interesting insights on JD that we will utilize later in our revenue build.)

GMV and Revenue Accounting.

To start, JD used to report Core GMV and GMV. The difference being that Core excluded Paipai, the C2C platform acquired from Tencent in connection with Tencent’s JD investment. Paipai was relatively small and the difference between the two was around ~5%. Paipai was eventually closed down (largely due to a large presence of counterfeit/inauthentic products), but we will ignore the impact of Paipai for the rest of the GMV discussion, as relative to the other factors that affect GMV, it is trivial.

The confusion with GMV comes from the fact that the “industry standard” is a very flawed metric, which JD tried to improve on, but only exacerbated the confusion in the process. At one point, they had 3 different GMV metrics: Core GMV, GMV, and then a GMV figure in the footnotes which were in-line with the industry’s definition. The variance between “industry standard” GMV and the GMV figure JD highlighted was large, almost 40% greater (RMB 939bn vs. RMB 658bn for 2016). Compounding this problem is the fact that since JD reports 1P revenues (Net Product Revenues), when investors wanted to draw a comparison between their 1P revenues and their 1P GMV (which was temporarily disclosed), there was a further >50% difference (RMB 372bn vs. RMB 237bn for 2016). All of this added to the impression that there was some accounting malfeasance and JD was targeted with several short reports. While we will readily cede that all of this is unnecessarily confusing and could definitely be presented in a clearer way that is indicative of the actual underlying platform economic reality, we are able to explain all of it and largely piece it together. The outcome is that GMV is greatly overstated from the actual transactions that take place (remember: this is an industry problem, not specific to JD). However, the different pieces we go through are not just accounting chicanery, but actually informative to JD’s business.  

As shown above, there are two GMV metrics JD reported for 2015 and 2016 (also reported for a partial period of 2014 and 2017). To help simplify, we will only focus on annual 2016, which is the last full year they reported Method 1 GMV. As you can see, there is a huge variance between the original method (Method 1) and the second method they moved to report exclusively in mid-2017. The difference between the two methods is explained in the footnote below.

As the boxes in the picture explain, Method 1 basically excludes any orders over RMB 2k that are not ultimately sold or delivered, and Method 2 raises this exclusion to RMB 100k and also excludes buyers who spend more than RMB 1mn/day. This exclusion is ostensibly done in order to reduce fraudulent transactions, or what is known as “brushing”. Brushing is the practice of creating fake purchase orders in the aim of moving your product up the search result page as well as having the opportunity to create fake positive reviews in order to entice other buyers to purchase your product. JD notes this in their risk disclosure below. Why RMB 2k is picked as a threshold is not clear, as it is still a sizeable enough sum that you could conduct brushing, but as it will be made apparent by the end of this section, it still does eliminate a ton of uneconomic activity.

Given the nature of 1P, whereby JD is the seller of the goods, there isn’t an incentive at the company level to create “phantom” transactions as JD would want the customer to be able to best delineate between mediocre and quality products. However, within the group level, there are some stories of product managers being bribed to make a certain product look more favorable by creating fake sales. That said, this behavior is rare, and JD deals with it harshly, immediately firing them and sometimes pressing charges (some have even served jail time). This very harsh crackdown partly harkens back to Richard Liu’s childhood where he would notice that the village head always had pork available to himself, but the village seldom ever had it (Richard has said that he would only get to eat it twice a year at most). He thought there was something immoral about someone using their power to accrue gains to themselves at the cost of everyone else and has since created a culture at JD that is extremely intolerant to graft and even the appearance of malfeasance. Nevertheless, there are still occasional stories of suppliers trying to bribe the purchasing managers in one way or another, but we believe that this is very rare within their 1P business. However, their marketplace business is much harder to police, and we are sure brushing was and still is widespread, which inflates the reported GMV figure. Interestingly enough (and bizarrely), they fully acknowledged this on a 3Q17 earnings call saying:

It is odd management acknowledges the futility of a metric they will continue to provide, while making clear they opted for the worse metric (that fits the industry definition), rather than their metric that they believe was more indicative of the underlying economic reality of the platform. While we can give them credit for flatly saying “it should not be relied on for financial analysis” purposes, it is very odd that they chose to provide a metric they think is nonsense to conform to industry standards over helping their investors understand the business better. We have heard they did this because the press would constantly misreport their market share, using the smaller figure that has a more onerous exclusion whereas peers had a more generous definition, but this seems like a shallow reason to us. Alibaba’s definition of GMV doesn’t actually make clear if it is the same definition that JD now uses. In BABA’s definition of GMV (shared below), we see that they use similar language as JD and note they exclude certain product categories over certain amount and buyers who spend a certain amount per day. But we don’t actually know what these limits are. Alibaba does say they try to eliminate the impact of fraudulent transactions, which seems somewhat comforting, but is sufficiently ambiguous that you can’t actually draw any conclusion from it. The more limited information compared to JD’s disclosure does make it seem a little misleading, but it is also possible they have a more nuanced system than a blanket RMB threshold. (We should have emphasized this point more in our Alibaba write-up, but in contrast to JD, BABA stands by their GMV figures, even using them to calculate their monetization rates. We have a more mixed opinion on the veracity of this having dug into JD more.)

However, we couldn’t argue if someone claimed that whatever level of brushing JD’s platform has, Taobao (and Tmall to a lesser extent) is as bad, or worse, since JD has a more stringent 3rd party seller screening process and is harsher on seller misbehavior. (We would caution, though, that it is hard to make any claim about the magnitude of brushing on Alibaba as there isn’t any disclosure that can inform us, and even when there is brushing, we don’t know whether it is included in GMV. While this isn’t an Alibaba piece, we would say that the implications would be their monetization rate could look higher than it really is and spend per buyer would be less. The former would be a negative and the latter a positive.) Another factor that could mitigate this behavior on JD is the fact that they charge seller commissions, whereas Taobao doesn’t, which makes brushing near costless.

While we have noted the differences in GMV methods for JD, before concluding with the ramifications, we first look at 1P GMV vs 1P Revenues below. As you can see, 2016 1P GMV is RMB 372bn whereas reported 1P product revenue is RMB 238bn. Given this huge disparity of RMB 134bn, or 36% of GMV, we wanted to walk through the delta below.

We first start with Net Product Revenues of RMB 238bn (2016), and we gross it up for the VAT (value-added tax). This variable is relatively easily accounted for, as they note it ranges from 13-17% based on product category. Most merchandise appears to fall into the 17% category, so we assume a figure towards the high end at 16%. We take the reported net product revenues and gross this up by the VAT to get the product amount including taxes. The difference between this figure and their 1P GMV is a grab bag of different items. Remember we are comparing 1P revenues to 1P GMV, and we are assuming that JD does not engage in brushing on their own platform for reasons mentioned prior, so that would not be a factor here. This difference would be (1) discounts that JD offers consumers, (2) orders that are returned, (3) placed orders that are later cancelled, and (4) other business taxes or fees, which we think is the least consequential factor. As far as items 1-3 are concerned, it is hard to discern how to weight them, but it is not uncommon to have a ~20% return rate. While this rate isn’t problematic in and of itself for an ecommerce company, it is interesting to see it calculated as it is seldom broken out directly. Below, we use it to inform our analysis on the 3P marketplace.

The analysis we ran in the prior exhibit allows us to estimate what 3rd party sellers net revenues are, which when layering in Marketplace and Marketing Service Revenues, gives us an estimated take-rate of 9.2%. This will be relevant later on when we think about take-rates, even if it is a figure that is 5 years old. Additionally, we note that every $1 of “Method 1” GMV turns into an estimated 65 cents of revenue (a ratio that would hold for 1P as well). We call this ratio out because it will help make apparent how much the quality of GMV has deteriorated. It should be noted that the Method 1 definition only excludes transactions that are not completed above an arbitrary limit of RMB 2k (or ~$315). So to the extent there is brushing in JD’s marketplace of transactions under RMB 2k, then seller revenues would be less and the take-rate would be higher.

Next, we run this analysis on 2016 figures, because under the latest “Method 2” GMV figures, all of this is greatly obscured as we show below. 

In the exhibit above, we make up two terms to help clarify our analysis: 1) Platform Revenue which we define as total revenues generated from the underlying GMV. This is calculated by adding JD’s Net Product Revenues to the 3rd party revenue estimate we calculated in above.  2) Revenue Monetization Rate which we define as Platform Revenues / GMV. This metric shows how much revenue is generated for each dollar of reported GMV and is designed so we can get a sense of the quality of GMV. A high rate suggests that most of the GMV has a true economic nature vs. just being brushing, returns, cancels, taxes, or prices prior to discounts. As you can see, the rate deteriorated dramatically under the Method 2 accounting treatment to the point that not even half of each dollar of GMV translates to revenue for JD or a 3P merchant. In 2016, GMV increased an astonishing RMB 281bn under Method 2 accounting, but of course this was not accompanied by any increase in real economic activity. We partially want to give kudos to JD for originally excluding all of this activity (transactions over RMB 2k that were not completed) from their GMV calculation, but then are disappointed they reverted to inferior metrics for ostensible vanity reasons and have to take stock in management’s judgment for that decision. (If the issue really was the press kept taking the GMV figure at the top of the report and not in the footnotes, then why not just switch their positions…)

Market Dynamics and Value Proposition.

With a strong sense of JD’s accounting, we will now move onto JD’s value proposition vs. their competition, and general market dynamics. As we have noted in prior pieces, China’s offline retail environment never had the chance to mature before the advent of the smartphone and widespread adoption of the internet. As a result, ecommerce subverted a lot of physical retail development. Ecommerce penetration (portion of retail sales conducted online) in China remains the highest globally at an estimated ~35%, but physical retail square footage per capita is about 1/10th of the United States. The rapid rise of the Chinese middle class coupled with inadequate retail infrastructure set the perfect stage for ecommerce to thrive. While Alibaba’s Taobao (2003) was the first Chinese-born company to capitalize on this opportunity in a big way, JD was close behind, moving fully online just a year later. However, the differences in their models meant that Alibaba could grow much more rapidly with minimal capital expenditure, whereas JD’s growth would be limited by their financial credibility which constrained their ability to procure a wide variety of goods at cheap prices. Thus, JD expanded slowly category by category. Alibaba enabled merchants to conduct sales online and could disregard the merchants’ listing prices and their COGS as the marketplaces’ competition would ensure the lowest cost producer would win out overtime. When Alibaba moved into logistics, they took a similar approach in creating Cainiao, a logistics platform that 3PLs (3rd party logistics providers) plug into. This essentially empowers logistics companies to effectively compete against each other, driving down the cost of providing delivery services. Here again, JD took the more expensive route, opting for a fully integrated in-house logistics system, which had to be built out painstakingly and expensively, district by district. By the time JD opened up their platform to 3rd party sellers, they already had a strong reputation that allowed sellers to comfortably associate their brands with JD’s retail channel whereas Alibaba had to start over fresh with Tmall to bifurcate their seller base and signal higher quality to the consumer. This all means that the average consumer would associate JD with quality followed by Tmall and then lastly (and much further behind) Taobao. Below, we will touch on the different ecommerce players in more depth (taking much inspiration from our Alibaba Piece’s Competition section), but first we wanted to think about different factors that a consumer considers when weighing different shopping options. We will do this by talking about a concept we will refer to as the “Consumer’s Hierarchy of Preferences”, a term we do not believe has been used elsewhere but is very apt for the concept we want to convey.

The Consumer’s Hierarchy of Preferences.

There are ten factors an ecommerce provider can focus on to provide a distinct value proposition to the consumer. We will spit this into two groupings: the first group is 6 factors that a consumer would always want to be the best possible, and the second group is 4 trade-offs. In the first group are 1) trust, 2) selection, 3) price, 4) delivery speed, 5) consistency and 6) order ease. These are all things a customer wants to be as good as possible. The second group is 1) discovery commerce vs. intent-driven commerce, 2) product standardization vs. uniqueness, 3) gamification vs. utilitarian UI, and 4) group buying vs. independent purchasing. These are trade-offs even though a platform may try to cater to both ends: in reality, they only tend to do one well.

Similar to Maslow’s Hierarchy of Needs, once a consumer gets one of their preferences filled on some level, filling their next order preference becomes more important than improving on that first one. For instance, many consumers may want the lowest price, but once it is competitively priced, they would care less if it’s 5% cheaper than if they can receive it this week, and if that condition is met then they care more about free returns than one day shipping and so on. We will refer to these different variables that a customer weighs and shifts as lower order preferences are met as the Consumer’s Hierarchy of Preferences. (This is a concept we have alluded to, but never defined in prior pieces.) The first group of 6 preferences need to all be done well at some level for an ecommerce company to be successful, but different players may emphasize different variables more. Coupang, for instance, has competitive pricing, but they are not always the lowest, and we could find some items that were priced 10-15% higher than other online options. However, buyers in Coupang’s “Consumer’s Hierarchy of Preferences” value other variables higher, so Coupang moved to focused on those variables, accepting that a few items may be listed somewhere else for more.

How you create consumer value is finding out the Consumer’s Hierarchy of Preferences and meeting as many of those preferences as possible. Once you address a minimum level of preferences (and these preferences can be quite low– like a convenience store having some drink that can quench your thirst), you have successfully earned a customer. However, operating your business to provide customers the worst service they would still transact at (essentially their indifference point) is a dangerous proposition as consumers’ preferences will change the moment someone offers a superior value proposition and that will become their new standard. For instance, a decade ago, guaranteed 4-day shipping on any good would sound delightful, but today we would scoff at anything longer than 2 days (and 1 in most cases). So, in order to have a strong hold on your customers, a company’s aim should always be to meet as many of a consumer’s preferences as possible.

When a company addresses a Customer’s Hierarchy of Preferences beyond their “indifference point” of transacting, they are creating surplus consumer value. While there is a tendency with many businesses to want to move as much “value” to the shareholders, the best businesses always leave some surplus consumer value. The reason for this is simple: the most valuable customers are those who stay. We don’t mean this flippantly, but rather if you were to run an LTV (lifetime value) analysis on a customer you would find that the biggest factor that influences a consumers lifetime value is churn. Of course, there is no mathematical equation that tells you when you reach a consumer’s indifference point (ignore the economics lessons that say otherwise), so being obsessed with capturing value rather than creating it usually results in pushing too far and thus eroding the customer value proposition, which is very hard to recover from (customer re-acquisition cost is going to be higher than that incremental dollar you grabbed). This is clear when Bezos endlessly talks about “customer obsession” and “always delighting the customer”: he is not concerned about pricing to maximize dollars today. Similarly, Costco steadfastly refuses to mark-up products beyond 14-15% above cost regardless of whether the market would bear a much higher price. This comes from the same ethos of creating as much consumer surplus value as possible in order to keep customers “delighted” so they stay long-time Costco shoppers.  Creating surplus consumer value is like digging a deep ditch where consumers just roll into it. The larger the consumer value, the deeper the ditch and the faster the consumers “roll in”, and the steeper it is for them to get out.

We can see JD exhibits some of these qualities when Richard Liu rebuked an employee for raising prices on an item that was selling out so quickly that they were almost out of inventory. Liu noted that this was short-sighted and chided him in a meeting, “You just think of earning money. Have you ever thought of the feelings of the customers that would hurt? You’ve made a few bucks in a short time, but you’ve lost heavily on the customer front”. Such sentiment demonstrates the concept of creating excess consumer value and there are many other similar examples from JD.

10 Ecommerce Variables of Consumer Value and Competition.

Below, we will define and exhibit the main different variables we see in creating a distinct consumer value proposition and differentiation vs. competition. In the first group is 1) trust, 2) selection, 3) price, 4) delivery speed, 5) consistency and 6) order ease. (We will only include a few snippets from our Coupang piece below, which has fuller explanations for why these variables matter and is worth revisiting).

1) Trust. For eCommerce, this was originally simply whether the buyer believed they were going to be scammed when transacting online with cheap, fake goods abound. JD’s reputation for only selling authentic goods is a differentiator vs. competition, but Alibaba’s Tmall is close in reputation (for illustration, luxury online platform, Farfetch, had a partnership with JD who even invested in them for a 3% stake, recently dumping JD for Tmall). Today though, trust entails much more than selling real goods— it encompasses much more customer service, including friendly return policies and attentive service reps for potential issues. If something is wrong, customers need to be able to trust that they can easily reach the company and find a solution they find acceptable. JD’s customer service was legendary in the early days for always going the extra mile for the customer. Some stories include employees tracking down missing packages to deliver them themselves when they were off the clock, as well as a rep staying on the phone for hours to help a new father keep his newborn from crying. Richard Liu himself recalls spending a week with a customer teaching him how to use a computer (this was back in 1998, but is representative of the ethos he tried to instill in JD). JD has a generally accommodative return policy with all products (excluding perishables) allowed to be returned within a 7-day period and the shipping fee covered by JD if it was originally free. None of the consumers we spoke with ever had an issue returning an item on JD. As we said in our Coupang piece:

“The ease at which they allow returns and readily capitulate that almost any issue is their fault has instilled significant trust in consumers. In some sense, you can think of their strategy as instead of reacquiring customers through marketing, they obviate the need to reacquire customers in the first place by avoiding potential churn events.”

We think JD follows a similar playbook, although perhaps not to the extend of Coupang (but more so than any other Chinese-based player).

2) Selection. Historically, JD was always considered the strongest in electronics and appliances whereas Taobao/Tmall had a much larger selection with a long tail of millions of small merchants. While today JD’s selection is still more limited vs. Alibaba’s Taobao and Tmall (especially in apparel), JD has rapidly added many categories over the years and, broadly speaking, has everything a consumer would need in the typical month. Since they opened up their platform to 3rd party merchants, selection has increased a lot, although Alibaba’s “2 choose 1” policy of not permitting merchants to sell on both platforms greatly harmed JD’s merchant variety. (Given BABA’s larger consumer base, merchants mostly picked to stay there). However, with China’s anti-monopoly scrutiny last year, the government began cracking down on the practice (it was always technically illegal but never enforced) and specifically fined Alibaba a record amount for doing so. Now, more merchants are selling across both platforms, but effects of that policy still linger. Nevertheless, JD has over 9mn 1P SKUs and even more 3P options.

3) Price. JD’s prices are generally in-line or lower than other platforms for like-for-like products, but the big caveat to that is that they often do not carry all the cheapest brands or the exact same version of a product line. Since JD started as a 1P retailer, they had to have their purchasing team build relationships with each supplier. In order to receive good prices, the purchasing team needed to place big orders, meaning they would be somewhat limited in how many suppliers they could buy from at that scale. If you recall, their strategy was to place orders larger than their current sales rate so they could get larger discounts that they in turn passed off to their customers in the form of lower prices in hopes of stimulating more demand. In order to get the scale necessary to do this, you have to consolidate your SKUs. The byproduct of this is that you have many popular products that are the lowest prices, but then other SKUs that are priced the same as everywhere else if not higher. JD focused restlessly to make sure they were never more expensive than other sellers in the early days (today they have automatic web scrapers to monitor prices across platforms and ensure they are priced slightly cheaper) and today the prices are close enough that consumers focus on other factors when purchasing (incremental better prices are lower in their Consumer Hierarchy of Preferences now). Today, this means that JD has some products where they are able to procure for very good prices and draw consumers in (similar to Costco where you would occasionally find an especially great deal on a random product) with the rest of their 1P SKUs being priced in-line with competitors. Their 3P merchants set their own prices, but usually keep a similar price across platforms.

It is also worth mentioning that manufacturers and brands who don’t want to run into channel conflicts with their offline retailers would often create a slightly different product, so it was not comparable to the version they sold elsewhere. While they often did this so they could sell a cheaper product online than in the store (which had higher mark-up to sustain the costs of the physical store), it still meant the products were not totally comparable by design.

In conclusion, JD’s prices are low, but as we noted, that is only for like-for-like products. There are many smaller and cheap brands that JD simply does not carry, which leads to their reputation for being more expensive than Taobao/Tmall.

4) Delivery Speed. With over 90% of orders delivered same day or next day, JD has the fastest delivery network in China. Controlling their own warehouses and fulfillment center with almost a quarter of a million delivery and warehouse personnel has allowed them to not only achieve more rapid delivery speed than peers, but also have fewer errors. Around the time of big shopping festivals like Singles’ Day, their superior shipping capabilities really stand out, as Alibaba’s system often gets congested and leads to long shipping times in excess of a week, whereas JD’s system is able to handle the increased volume. In fact, JD actually poked fun at Alibaba in a viral ad campaign that showed a woman who ordered sunscreen but got sunburnt before it arrived, and a man who ordered a razor with his beard touching the floor by time it came. For goods under their “211 program”, any orders received before 11am will be delivered the same day and orders received by 11pm will be delivered by 3pm the next day. There is no extra charge for this, but consumers can pay a surcharge in major cities for expedited <2 hours shipping.

5) Consistency. This variable is essentially to what degree does everything “just work” flawlessly. When a consumer wants to order something, stockouts aren’t an issue, the delivery is on time, the package isn’t damaged, the goods that arrived are the correct ones, and if there is any issue JD consistently makes things right without a fuss. It takes only a few bad experiences to turn a customer off forever. If you have a beloved restaurant that you frequent often, you will probably let one bad meal slide. But after the second, you would worry that they have slid down in quality. Mess up a third time, and the restaurant is probably off your rotation forever. This is especially true if you complain, and they deny there is an issue. This is why consistency is critical, which requires very robust systems to catch and prevent issues from happening in the first place.

The other piece of this is their “Haodian” (Good Store) seller ranking system that gives 3rd party sellers a marker from JD so they are reasonably guaranteed high quality. Factors like customer ratings, amount of time items are in stock, and speeds to ship are used for a merchant to earn this seal of approval. This seal pushes merchants up in the search results, which is very valuable as it generates demand, so sellers are highly incentivized to do the best they can on these metrics.

6) Order Ease. This relates to how quickly a consumer can find what they want and purchase it. JD is similar to Amazon and Coupang here with a fairly clean website that is designed to get users in and out quickly. After you search for your product, you can purchase it in just 3 clicks with your address and payment info already preloaded. The exception to this is that JD is experimenting with more group buy functionality (in addition to the Jingxi standalone app) and if an item is available as a group buy than you need at minimum 2 buyers to get the discount on the order. Frankly, this seems discordant with the JD value prop and it is hard to understand what this does for the consumer or for the merchant. For example, if you look up a frying pan, you can find the one that is 10% cheaper on group buy. So, in order to get the discount, you have to start or join a group. The order will not be placed until a group of 2 is complete. We will talk more about group buying below and we understand the rationale, but in this situation, it just seems like a nuisance to have to start a group a wait few hours in order to get a ~10% discount. Maybe there is an excitement factor once the order is completed but again that is not what customers go to JD for: 90% of orders shipped in one day or faster and waiting several hours to complete an order does not seem like a coherent consumer message.  

The other big factor here is that there is no option to pick different logistics providers, as every order uses JD logistics vs. Tmall where the user will pick from several different logistics providers. This decision is annoying for consumers and creates paralysis which is partly why SE’s Shopee is starting to eliminate multiple shipping options—it’s okay to have shipping options based on “when” the item arrives as the consumer will easily know if they want it tomorrow or if they can wait 4-5 days for a credit, but making them choose between ZTO, STO, Yunda, SF Express, ect. can build decision paralysis and cart abandonment.

The one “friction” we see in JD’s ordering process is that orders below RMB 99 are not free and even for JD Plus members they only get 5 coupons a month that waive the shipping fee (for RMB <99 orders). While RMB 99 ($15) is a fairly low minimum, it seems unnecessary to create this friction for members since it seems unlikely that they are regularly ordering items under 99 RMB more than 5 times a month, or 60 times a year, anyway. Coupang and Amazon want to eliminate every excuse a consumer could have to not order the second you realize you need something because this behavior is so much more valuable than the money saved by not serving a few more low value orders. (This is especially ironic in the context of them covering free shipping on RMB <99 orders for Jingxi users who have far lower LTVs).  No consumer we spoke with recalled wanting to buy something but then abandoning it in their cart since they didn’t hit the minimum, so it is not practically problematic, but it just seems at odds with the rest of their ethos.

The prior six factors are all things you want to be as good as possible. The next four variables are all trades-offs, but you will see a common theme between them all: they are all centered around the decision of giving a consumer a quicker or more immersive ecommerce experience.

1) Discovery Commerce vs. Intent-driven Commerce. Discovery commerce is when you buy something you didn’t know you wanted. The platforms that do this well are Douyin/Tiktok, Pinterest, Instagram, Pinduoduo, and Taobao. As you may have noticed, most of these platforms have social functionality that either directly relies on a friend/influence/KOL recommending something or helps tailor the algorithm for better recommendations. Pinduoduo’s “recommendations” come directly from friends who would send you a link through Weixin (WeChat) to join a group buy. Instagram’s ads and their ability to integrate commerce on an influencers page means users who are just scrolling through will often find something they didn’t even know they wanted. Taobao is the only player that transverses both, but leans more towards discovery with a huge amount of products by independent merchants that are searched through (remember Taobao means search for treasure) and their large livestreaming presence coupled with short videos bolsters their Discovery capability. Short videos and live streams are also a good way to discovery products you didn’t know you wanted while simultaneously getting an ”endorsement”. When commerce becomes entertainment, it is Discovery.

Intent-driven platforms are designed to get you in and out as quickly as possible. These include Amazon, Coupang, and JD. While they all have recommendations, they mostly serve the customer who has a strong idea of what they want and want to complete the purchase quickly. If you know what you want, you can easily buy something in well under a minute on these platforms: just type in your search and click buy. When buying something is considered a hassle that should be minimized as much as possible, then it is intent-driven commerce.

As a side note, it is interesting to consider that this distinction isn’t relegated to just ecommerce: stores like Marshalls or the 99-cent store have a notable discovery aspect to them whereby users hunt for deals on products that greatly vary every time they visit. Most physical retailers are firmly in the “intent-driven” category and will make sure to always have products housed in the same section of the same aisle because it helps the consumer get in and out quickly. The rule of thumb here would be that if a retailer is regularly changing what products are stocked on a shelf, then it is more discovery commerce, whereas if it is consistently the same products it is “intent-driven”. (That’s not to say physical retailers won’t play games putting popular items at the back of the store to make you walk through the whole store, but they would never change that placement every 2 weeks).

2) Product Standardization vs. Uniqueness. This is in reference to the platform’s product selection. Most ecommerce platforms only sell standardized products, but some like eBay, Depop, Carvana and Facebook Marketplace mostly list “unique” product SKUs owing to the fact that most items are used (or collectible in eBay’s case). Interestingly, there isn’t a very popular platform in China that specializes in used goods, likely owing to the fact that there are many options for cheap goods and most cheap goods will not last a second lifetime (Beijing Zhuanzhuan is $3bn used-goods platform backed by Tencent with Level 2 Entry, but very few consumers we surveyed had even heard of it). Selling non-standardized products creates many issues that most ecommerce players would prefer to avoid that all together. (Etsy, a marketplace that focuses on handmade items, is an interesting cross between since in some sense they are attempting to “standardized” homemade items which are definitionally non-homogenous products.)

3) Gamification vs. Utilitarian UI. This variable was critical to the rise of Pinduoduo and Shopee. We defined gamification in our BABA piece as the following:

The idea of gamification is simple: make the app look more like a game and reward the consumer behaviors that you want to incentivize within the game. In one of the original PDD games, users receive “water droplets” and “fertilizer” by sharing with friends, buying goods, visiting merchant pages, and checking in daily, which they can use to grow a fruit tree of their choice, receive a shipment of actual fruit at maturation. Features like this help increase user engagement, stickiness, and virality.

Gamification adds a fun layer to the shopping experience, but the trade-off is between making the platform more fun where you spend more time immersed in the app or if you make the app more “utilitarian” with the design aimed at getting people in and out as quick as possible. JD, Amazon, and Coupang are all in the group of minimizing the amount of time it takes for a user to find what they want and to buy it—it is almost comical imagining them have you water a virtual tree in order to get a discount. This is one of the biggest trade-offs we have recently seen in ecommerce and it is a critical one as it is one of the starkest decision an ecommerce company can make.

4) Group Buying vs. Independent Purchasing. Group buying makes commerce more social, requiring a user to get friends (or strangers) to join their purchase in order to get a deal. It may seem slightly gimmicky, but it does have strong business rationale as larger purchase orders allow a manufacturer to produce goods cheaper, with the social function acting as a free customer acquisition channel for both the ecommerce platform and the seller on that platform. Below is our explanation of how it works from our BABA piece:

While there are a few different ways to implement this practice, the general idea is that the platform helps aggregate purchase orders and sends them directly to the manufacturer, allowing them to circumvent a distributor who traditionally played this role. The value prop to the manufacturer is clarity in demand forecasting coupled with scale, allowing them to offer customers the same very cheap prices that a distributor would usually buy at. The consumer gets cheaper goods, but they often arrive slower as they are not on demand. PDD also has group leaders that can start a group purchase order and enlist friends often via WeChat. The group leader (depending on the model) can receive a commission from the order and is also responsible for collecting and distributing the items, which is a very elegant solution to the last-mile delivery problem.

Group buying is more popular with cheap goods that are not a “high priority” need. While you can still group buy essential goods likes paper towels or milk, it is done spontaneously because you see a great price on it and know you will eventually use it, rather than because you are restocking after you ran out.  Group buying tends to merge well with discovery commerce and gamification.

These are all the main variables of competition for ecommerce. Other variables could include being omnichannel, so consumers have a pickup option as well as being more integrated with entertainment whether that be live-streaming or short-form videos (we mentioned in passing in the discovery section, but this is an important draw for many users. JD does have live streaming, but it has a more professional feel to it to not degrade their high-end image). Having a large library of reviews can also be a distinguishing feature. Additionally, new features like AR/VR experiences and the ability to “scan” an item with your camera and buy it with a few clicks could also become important. We will apply these different ideas now to the big ecommerce competitors below. (For BABA readers, a good amount of this will seem familiar.)

Bytedance. Bytedance may be thought of more as a social media company, but they are parlaying their captive user base to move more into commerce. The large amount of user time spent with full screen videos makes Douyin a natural venue to serve ads and facilitate in-app commerce. They are estimated to support around 600bn RMB in GMV this year. Their ads tend to be highly immersive, and most users do not quickly realize it is an ad (if they do at all). KOLs and influencers can also create videos of products they like and send users to the product’s store via a link and receive a commission on sales. While the nature of these purchases tend to be discovery-driven and more impulsive, they are building out their search capabilities that could serve more intention-driven purchases in the future. Their large base of videos could serve as a great way for users to learn more about specific products and see them in use. There is a very clear use case for verticals like makeup where a user could search “dark lipstick” and get a wide range of high-quality videos that showcase different products. In the future, Douyin may allow users to directly purchase it in app. However, most users do not think “Douyin” when they want to buy something—and changing how a user interacts with a product is hard. Also, today many merchants drive traffic back to Taobao, Tmall, or JD stores, but Bytedance is clamping down on that as they are trying to get merchants to set up stores on-platform (similar to Instagram building out Shops). But even when a merchant does open a Douyin store, it is often for a limited amount of their SKUs. Generally, we see Douyin as a totally different value prop to JD. Douyin might hurt some growth opportunities for JD in verticals like cosmetics and apparel, but we see it being relatively incremental, especially since JD is weaker there anyway (the threat to Taobao and Tmall is larger). To draw a parallel to the West, Tiktok may be scaring Facebook investors, but it has not had a material impact on Amazon.

Kuaishou. They are in a very similar boat as Bytedance in value prop, but much further behind in execution. Their higher portion of lower tier city users also means that disposable income will be lower. The videos within Kuaishou have a somewhat more “lower class” reputation, which some brands will probably not want to be associated with. Either way, similar to Bytedance, we think JD’s value prop is totally different and they are not direct competition. We can also put Bilibili and Xiaohongshu’s ecommerce ventures in a similar bucket.

Pinduoduo. Pinduoduo has become a phenomenon in China and popularized gamification and group buying. However, compare PDD to JD and you’ll see that they are at opposite ends of the market on almost every factor with JD focusing on mid to high income consumers with high purchase intentionality who want to buy things quickly versus PDD on the money for value side with a high portion of rural consumers who mostly spontaneously buy goods solely for their price. JD’s shipping is rapid, whereas PDD users usually don’t care that it’ll take several days since they’re not goods they even realized they needed before they entered the app or were sent a link by a friend. PDD’s reputation for being lower end will also preclude many sellers and brands from wanting to sell on that platform for fear of being associated with a “cheap” or “discounted” distribution channel. Pinduoduo clearly has their use case and has grown rapidly to over 867mn buyers (much more than JD’s 552mn), but we see the consumer value prop as very different, and it is hard to see many JD customers moving purchases over to PDD.

Meituan. Generally speaking, Meituan is not a JD competitor, but they are taking share in “quick commerce”, which is purchases for items you need in <2 hours. JD’s general approach to fend off this threat and still be considered the fastest delivery option is to partner (and later take a majority investment in) with Dada, a crowd-sourced, local delivery platform. JD embeds quick delivery options within their JD app, website, and mini program (when you are at checkout you can pick quick delivery for a fee or can search through quick commerce options) for delivery within all major cities. The difference is that Meituan helps enable local businesses to deliver whereas JD deliveries products stocked in their warehouses. In the Meituan app, people are more likely to search by store whereas in JD consumers would search by product. The desire for quick commerce isn’t so large that this is a critical growth area, but JD still wants to be able to defend here. Having said that, we don’t consider Meituan leading quick commerce as an existential risk to JD by any means, given the limitations of parlaying that advantage into the level of selection JD has available at same day shipping, especially considering it costs more (and going back to our “Consumer Hierarchy of Preferences” framework, most users do not care if an item comes in 2 hours or 24). JD’s partnership with Dada is also partially strategic, so JD has a better local delivery network for grocery, where delivery speed does matter due to perishability and fact that most produce is fragile and bagged, not boxed (making it more delicate to enter into their traditional logistics network). Grocery is the most important quick commerce category owing to its high frequency and stickiness (more on JD’s efforts here later), but it is still very early innings.

Elem.me. Ele.me is a similar competitor to Meituan in quick commerce, but given their Alibaba ownership, they could potentially have an advantage in this area if they plug their Ele.me delivery network into Cainiao to deliver from local warehouses in addition to local stores. Again though, 2 hour delivery is lower on the Consumer’s Hierarchy of Preferences with most consumers preferring free shipping to paying a surplus fee for rapid delivery.

Weixin mini stores. These are stores that merchants set up to sell directly to consumers within the Weixin app. They usually are only one of several channels a merchant will use, but very small merchants may only sell through Weixin since it is easy to set up a store and they can use a content marketing strategy to build a customer following without having to invest heavily in marketing. Brands are also increasingly using their Weixin presence to build their customer relationship and selling to them in-app. This is partly a risk, but also a positive for JD who partners with sellers to help them make Weixin stores (Kepler program). JD offers other services that Tencent never will (like logistics and inventory storage), so we do not view this as big competitor to JD, especially since it shifts the field of competition from something JD was just okay at (capturing consumer attention) to something they are the great at (logistics and seller services).

JD has fewer buyers than both BABA and PDD, but their GMV/Buyer is more than twice PDD’s.

Alibaba. JD’s most direct competitor is Alibaba with their Taobao and Tmall stores. In conversations with consumers, most used both with Alibaba slightly preferred more than JD (and this is among higher income users, which was somewhat surprising). With all the talk about Alibaba losing market share, people forget that virtually everyone who shops online in China still uses an Alibaba property. In fact, since we last wrote our BABA piece, Alibaba has regained their position as the most-used ecommerce company in China with 953mn buyers vs PDD’s 867mn as they leaned into acquiring more rural area customers with their money for value apps (ironically, Pinduoduo was probably Taobao’s biggest customer acquisition channel). Alibaba has more selection and more goods that are sold at cheaper prices versus JD, and most BABA consumers consider their shipping options to be good enough. While customer service quality varies merchant to merchant, it doesn’t seem to hurt Alibaba too much. The Taobao app also has short videos and livestreaming as well as more discovery functionality that makes it more fun of a shopping experience vs JD which looks comparatively utilitarian. While for the more high-intent commerce, you can simply search for a product and buy from an approved brand store on Tmall with more quality assurances.

Alibaba has earned consumer mindshare as having everything, and should in some sense be thought of as a well-executed “Google Shopping”, but with millions of merchants that are exclusive to their property, powered by the largest 3PL logistics integration software (Cainiao) with the most popular payment method (Alipay) embedded. JD is similar to Amazon as a fully integrated logistics operator, with the important exception that JD has an appreciably smaller selection (and is usually the smaller of the two channels that 3rd party merchants utilize). From this reframing, the competitive dynamics might seem more natural to grasp: the question is more akin to whether you start your product search on a lionized version of Google Shopping or a somewhat defanged Amazon. The outcome is clear though, more users still prefer to conduct most of their purchases on Alibaba with about a 60/40 split in Alibaba’s favor (per the GMV per buyer stats above which also comported with our consumer checks).

While this conclusion today is indisputable that Alibaba is much bigger than JD, the concern is really what does the future look like. Here, we are fairly confident that JD will not lose consumer spend to any other platform (with one caveat below). The consumer who values quick shipping, frictionless returns, and quality assurances is highly unlikely to ever decide that they don’t mind getting worse shipping, potentially hassle-laden returns, and the risk of shoddy goods. What a JD consumer values is the hardest to replicate and that gives JD a moat versus Alibaba. If consumer preferences shift over time to value JD’s unique value prop over Alibaba, then Alibaba will be in trouble since they will not be able to fully replicate it by patching together 3rd party services. A common thesis on JD is that as the Chinese consumer gets richer, they will undergo a “consumption upgrade” that will be in JD’s benefit and to Alibaba’s detriment. However, as we mentioned the high-income consumers, we spoke with were still regular users of Alibaba, so we are not convinced of this thesis playing out in mass. Rather, our base assumption is that JD will continue to win consumers who value their service, but most consumers will continue to use both. However, there is a factor in JD’s favor: Since virtually every customer that would be a JD customer has already used Alibaba, that implies that Alibaba is unlikely to gain share as any current JD spend was spend that Alibaba already lost. This goes back to the Consumer Hierarchy of Preferences, and as we mentioned, the better service JD offers is a “one way decision” since no one is going to opt to go back to worse service. You may want to argue that Alibaba can improve shipping times and their Xiaoers (dispute mediators) can become more consumer friendly, but even still the consumer that JD won wouldn’t care since they value the assurance of knowing everything on JD will be okay and a few bad experiences on Alibaba likely turns them off forever. For these factors we think overtime JD will grow slightly faster than the ecommerce market.

Competition Conclusion: Many of these platforms take up some portion of disposable income that theoretically could have been spent on JD, but the point we want to emphasize is that the flow of causation is very different. By and large, a JD consumer who wants to buy something will still go to JD (maybe price-check against Tmall for larger purchases), but PDD, Bytedance, Kuaishou, and Taobao will not have a chance to earn that sale. However, it is possible that a consumer gets sent links to PDD or sees other stuff they want when they are watching Douyin and since they spent too much elsewhere, they delay a purchase on JD. Everyone is fighting for the same disposable income dollars, but the vectors that they fight on are very different. This is analogous to media where there is a limited amount of screen time that everyone is fighting over: Netflix and Tiktok are clearly a very different product, but a better NFLX show can eat into a user’s Tiktok time and vice versa. This sort of competition is important to keep an eye on, but a company generally has a limited ability to respond to it as it has more to do with changing general consumer preferences. (For example, Coca Cola’s decline in popularity in the US has nothing to do with a different soda taking their sales, but is simply the byproduct of preferences shifting away from soda to healthier alternatives.) Understanding the specific value prop and use case (or “what job you are hiring this product for?” in Clay Christensen speak) is important. As Bezos notes though, consumers are never going to want a more expensive product that gets to them slower, so focusing on aspects that are timeless is critical to building a sustainable business. The aspects that JD does well, they do better than anyone else (namely delivery, customer service, competitive prices) so even if new ecommerce players chip away at total disposable income, they will not be able to compete against JD on their core value prop, which is always going to be valued by consumers.


JD executed superbly several times in its lifetime. Starting as a stand in an offbeat market, Richard was able to extend that to a budding electronics retail chain, but instead permanently closed every one of his stores and moved entirely online at a time when most Chinese neither had computers nor internet access. Whereas the “safe” thing would have been to keep both businesses, he thought not going fully into ecommerce was an optimization of mediocrity. This was him burning the boats. When their volumes couldn’t allow them to procure goods at the competitive rates their larger competitors could, he pushed the buying teams to put in orders magnitudes beyond the current sales volumes in hopes that the larger discounts would draw customers in. Whether intuition or a gamble, it worked, and the lower prices were an effective customer acquisition strategy that also helped them build the inventory turns they needed to make ends means as a very low margin retailer. While a miserly “Walmart-esque” operating margin of a few percent was the goal, for years they would sustain heavy losses in a struggle to build the scale that allowed this model to work, operating in a market that is notorious for cutthroat competition with almost every player willing to lose money to gain share. The goal was to build the volumes that allowed them to offer the lowest prices to the most people and take a tiny profit on each transaction, which when multiplied across billions of orders would be a cash machine that had an ample moat.

Emphasizing customer satisfaction, they rolled out their own logistics network to address the largest customer complaint: slow delivery. Not shirking from the task of building out a whole logistics network to cover an area larger than the U.S., they soon built out over 21mn sqft of warehouse space with a quarter of a million-delivery personnel to get billions of packages delivered in under a day. To beef up their selection, they opened the platform to 3rd party merchants, directly confronting Alibaba; today they support over RMB 1tn of 3P GMV on their platform. Monetizing their users’ high intent, their seller services quickly included advertising, which would bolster their profitability: now JD management thinks that their mature margin will be 1-3 points higher than leading offline retailers at maturity. Today, JD, next to only Amazon’s North America business, has the distinction of being a profitable online 1P retailer, showing a positive net margin of >3%. With 552mn customers, they have finally achieved the volume they need to run through their enormous logistics network at an efficient enough scale to achieve profit; inventory days were last reported at 30 days. Wal-Mart is at 39. Having accomplished the enormous feat of transitioning an outdoor one-man-operated stall to one of the world’s largest retailers that can deliver millions of different products in under a day—and do so profitably—you’d think the future is already history. But competition isn’t standing still, and consumer preferences are a fickle thing. However, not to lose the main point: we think JD has more opportunity to win than risk of loss.

JD has always had a high-end reputation with authentic goods aided by not mixing them with the millions of cheap products that purvey Taobao’s marketplace. First a differentiator, now might be an offering gap. Alibaba moved up-market with a separately branded platform, Tmall, but now JD wants to go down market with Jingxi, similarly taking a dual brand strategy. An investment thesis on JD for years was predicated on the belief that as consumers got wealthier, they would upgrade their consumption habits and JD would be a clear beneficiary, saving them from the “cheap crap” that Taobao is littered with. But before the launch of their money-for-value platform, Jingxi, buyer growth was lackluster, increasing just 14% from 4Q17 to 3Q19. With Jingxi, a platform that seems at odds with everything JD staked their reputation on, it raises the question of whether the “consumption upgrade” won’t happen en masse as people still like buying inexpensive stuff and not just when they don’t have any other options. JD’s value prop was clear for years as a high-end retailer, but increasingly every ecommerce player is moving into each other’s domains. Alibaba’s Cainiao logistics network can now enable one day delivery in most main cities and brands are selling not only through Tmall, an old threat, but also increasingly directly to the consumer through their Weixin mini stores, utilizing various channels like Douyin to source customers. JD has gotten ahead of this with their Kepler program that helps retailers get onto Weixin and they have opened their logistics network to everyone, even restructuring it as separate business unit and IPOing it, but the question is what does this all mean about the future of JD’s core retail business?

Clearly, platform usage never became mutually exclusive with 953mn users on Alibaba and 867mn on Pinduoduo to JD’s 552mn. Part of Amazon’s genius is the Prime program—far from a triviality to just subsidize free shipping, Prime customers spend >2x nonprime customers and tend to buy anything that can be bought on Amazon, on Amazon. While of course there are things Prime customers buy elsewhere online, by and large this is not a promiscuous customer base. JD’s Plus program has a whopping 20mn members at last disclosure, a <5% penetration of their user base. Maybe it’s just that JD Plus has more limited benefits, but it is also a tell that the JD consumer is not the Amazon consumer, even if JD is the “Amazon of China”. Spend per buyer on JD lags 40% behind Alibaba’s, despite Alibaba catering to far more lower income users and over indexing on “cheap crap”. In all likelihood, if you find a JD consumer, you’ve also found an Alibaba one.  

With all this talk of competition, don’t forget that JD has still grown GMV, revenues, and buyers every year with their profitability steadily improving. As long as the pie keeps growing, there will be multiple winners. There clearly are consumers who prefer JD’s value prop to alternatives and over the longer term the number of consumers who fall in that category can grow, but to what extent is unclear. JD’s strong value prop is met by a strong response from BABA. While Alibaba continues to rebar their weak spots in logistics, authenticity, and customer service, JD is taking a swing at Taobao’s (and Pinduoduo’s) core consumer in the money-for-value space. It seems unlikely that BABA will be able to steal loyal JD customers since their preferences for a hassle-free order would preclude them from ever being willing to switch back to BABA, but those who shop in the money-for-value space are highly price sensitive which gives JD an opening. Clearly, Jingxi is growing and could have well over 250mn buyers on the platform, but these users are worth far less to JD and their fickle platform loyalties means it is an outstanding question if they yield a positive LTV. However, with a ton of merchants already on JD, perhaps having the option to sell their products heavily discounted in a different channel is valued by them. That said, we still don’t like this part of the story.

However, this can be read as part of the broader JD theme, which is to enable the customer to buy stuff however they want, or said another way, “retailing as a service”. From this framework it makes sense that they, after preaching quality and authenticity, are not only rolling out an app that is the anthesis of that, but they are also arming off-platform merchants with their biggest weapon—their top tier logistics network. This is while they are lending their inventory expertise with “Intelligent Supply Chain” and “Shangling Saas Mall”, as well as their brand name to what they hope will be millions of physical retail stores. Most recently, they penned a partnership with Shopify, the posterchild for enabling the long tail of merchants, to support SMBs to sell in China. If the purpose was to just create the best “JD.com” experience versus the competition, they would be keeping all of these advantages they developed to themselves. But seeing that the consumer has a variety of ways they want to shop, JD modularized their offerings and lets anyone access them, which ironically made them also more embedded in many sellers’ operations. If you set up your Weixin mini app through JD’s Kepler, use their warehouse for storage, and deliver through their logistics network, it would be an inconvenience of epic proportions to rip that all out and move to an alternative provider. There probably won’t be some critical point of wealth someone reaches where they ditch their prior consumption habits and moves to be an exclusive JD customer as was hypothesized in the original “consumption upgrade” thesis, but it is very likely that consumer habits are changing in a myriad of different ways and whatever ultimately captures them on their Hierarchy of Preferences is in some way powered by JD.

But like all investments, it is a question of what you are paying for and what you are receiving in turn. Putting aside their loss-making logistics operations, and ~$1.2bn of New Business losses, JD generated ~$3.3bn of LTM NOPAT in their core retail operations. Against a market cap of ~$100bn excluding net cash, that is a 30x earnings multiple (excluding mentioned losses), but on an EBIT margin that has ample room to expand overtime. Even if there is no consumer “upgrade” with buyers shifting where they spend, as the market grows JD will grow, and over time they will still be able to convert some consumers to be ardent JD buyers. But growth of this “core consumer” is going to be slower than their +25% y/y buyer growth with Jingxi, especially in the context of Covid accelerating customer adoption. We would think JD grows core buyers at a low to mid-single digit figure, slightly slower than their buyer growth rate in the 2 years prior with spend per buyer increasing slightly faster than overall disposable income as ecommerce penetration increases a bit with JD taking a slightly higher share of incremental ecommerce sales (Alibaba share is shrinking, but that is because they were first to market and used to have an unsustainably dominant 60% position. Over the last few years, as the market grew Alibaba still grew, just slower than the market). With a slower growing topline, we project a ~10% CAGR over the next decade, we estimate core retail earnings can still grow 20% compounded as their EBIT margins expand on mix-shift to more 3P sales supported by higher margin advertising. We will go through our estimate build in detail below, but at an undemanding 17.5-20x multiple we can see a low double digit return with a high teens annual return should they be able to hit the top end of their mature margin guidance.

However, as should be clear, that is not without risks in the fast-moving Chinese ecommerce environment. Their mature margins assume that competition ameliorates enough that JD can start to bear the full fruits of monetizing their platform. One of the biggest factors to growing margin is contingent on getting that high margin advertising revenue, but while merchants are happy with their return on ad spend on JD today, that return could deteriorate fairly abruptly. It also means that JD is in competition with every other seller of advertising, who may be able to connect sellers to their customers more efficiently or at the very least drive returns down for JD by flooding the market with cheap impressions as Douyin did. To be sure, advertising on JD’s platform will always be a unique value prop as they have high conversion rates given consumer intentionality and preloaded shipping and payment info coupled with JD’s high customer service reputation should anything go wrong, but whether we can count on this being enough to buoy a decade of 20%+ advertising growth is another question (more on this in the build discussion).   

Separately, capital allocation could destroy value as they direct cash flow from their maturing retail business into new initiatives that are outside of their core competence. While this is always a concern, it is more prevalent here as they have shown a willingness to experiment but have not had success outside of their core logistics and ecommerce operations. This isn’t like Tencent who has a track record of building businesses in disparate industries that were totally unrelated to their messaging/ social media roots with success in cloud, financials, and gaming. JD’s foray into financial services was a moderate success as it related to financing on-platform merchants (tangential to their competence), but then they tried to branch out with more financial services, cloud, AI and other technology services resulting in a restructuring, a rebrand to JD Digits that was spun-off with them selling a majority stake, followed by a second restructuring and rebranding to JD Technology. This whole ordeal is not confidence instilling in JD moving outside of retail, but for now it seems they have learned their lesson and new initiatives are at the intersect of technology and retail, historically a strong suite for them. However, aside from JD Finance/ Digits /Technology, their investment record is rather poor with tens of billions invested in companies like Bitauto (invested at least $400mn cash and $750mn of resources for a 25% stake that was valued at $300mn when taken private 5 years later), Tunio (invested ~$500mn to later sell it for $65mn), and Yongui ($700mn for a 10% stake with a ~$550mn market value 5 years later) and most recently Dada, which while strategic, is still a 3rd tier O2O player next to Meituan and Ele.me that is burning ~$100mn a quarter. Their CROIC since they went public in 2014 is an average ~8%, which isn’t terrible considering that embeds “growth” expenses and they have been materially improving it to >50% over the last two years. However, if you layer in their investments in their equity investees, that average drops to -25% and that >50% drops to 20-25%. Not included is the sale of their logistics properties that make JD a more capital light business, but the problem isn’t the capital intensity—they have shown they can get good returns in their core retail business—the issue is the free cash flow they destroy with poor investments. When thinking about the JD Retail business valuation, it is critical to remember that the difference between positive and negative free cash flow over the past 6 years is whether you count their investments in equity investees.

As long as China continues to grow and its middle class is prosperous, they will have more income to spend across all of these different businesses and that is important because as JD increasingly becomes a player across many different retail channels with their own JD.com platform, but also brand supported Weixin mini stores, franchised physical retail, malls, supermarkets, group buying, and quick delivery, not to mention entreenched backend inventory purchasing integrations with a network of stores, they are a bet on the Chinese consumer. One of the bigger risks is that the market stops growing and competition is no longer about fighting for a growing pie, but rather stealing from competitors, which could quickly breed competitive irrationality. Having said all that, it should not be lost that JD is a special company with a unique value prop and, again, one of the only online retailers to ever reach profitability. Their opportunity in retail really could have a second growth leg with omnichannel and as they fill out the appetite for physical retail space that still lingers with their trusted brand and quality products. JD customers love them and while they don’t have total lock on a consumer like Amazon with Prime, they do have a better grip than Alibaba, and over time we think JD has more opportunity to win consumer spend than they have risk of losing it.

Revenue Build.

We will now go through our revenue and EBIT build below. As we remarked above, JD is moving away from just serving the customer to helping the merchant serve the customer, so any build that is just based off of buyers may be wrong in the future if more revenues come from services. However, this isn’t the case today and it is unclear how they monetize a lot of these services, so we will stick with building our model based off on buyers, with the understanding that it is far from perfect. This model would be an estimate of what the future of JD would look like assuming their business is similar to today, however as we have noted, that is unlikely to be the case, with more channels of online selling, more seller services, and more omnichannel in the works. Nevertheless, we think this is a fair way to think about JD’s potential earnings over the long term and welcome anyone to adjust our build and assumptions as they see fit.

We focus purely on JD Retail since that is their main business and their other businesses have almost zero disclosures to draw any insights from (we will think about JD Logistics valuation in our separate JD Logistics write-up, but for now you may simply take their publicly traded market value of ~$20bn). We will conclude with more commentary on this though.

Our revenue build starts with 2020 buyers, so we can calculate GMV per buyer (Unfortunately, they haven’t disclosed any YTD or 2021 GMV figures yet—however, as this is a 10-year forecast, it shouldn’t matter much, and we took into account incremental YTD info). With 472mn buyers at the end of 2020, that equates to RMB ~5,500 per buyer. As seen below, GMV/buyer dropped about ~5% when they rolled out Jingxi, despite strong ecommerce tailwinds from Covid.

In thinking about GMV/buyer for JD, we assume they will continue to add more users (+80mn buyers YTD), but the GMV contribution from these new users will be much lower. As we had seen with BABA rolling out their money for value apps, they were able to quickly grow with lower tier users, after Pinduoduo first introduced them to ecommerce. Given these consumers are highly price sensitive and have little platform loyalty, we think there is a strong chance JD is also able to acquire a large portion of them. However, their low value will be a strong headwind to GMV/buyer.

Our back of the envelope math informing our 5-year out assumptions below is that they have ~440mn core users (~30mn are Jingxi-exclusive) who spend at the 2019 level of RMB 5,800. Growing this  core user-spend at a low double digit rate (12%) means this core group will be spending RMB >10k in 5 years, which is about where BABA’s spend/buyer is (and also the amount of time they think it takes for a buyer to come close to fully ramping up). This implies >85% of GMV comes from these core buyers with the reminder from new users who are going to be disproportionately lower value Jingxi users (and from lower tier cities, as management noted recently that is where 80% of new buyers are coming from). Pinduoduo’s GMV per buyer is around RMB ~2k, which is about what the implied GMV/buyer is for the remainder of their “non-core” buyers. Of course, these are very high-level assumptions with a large distribution of outcomes, but we are showing what we think is reasonable with the highest confidence level we can have (which isn’t that high in absolute). Nevertheless, this rolls up to a 14% 5-year GMV CAGR, which seems very plausible.

Over the next 5-year period for our “T+10” estimates, we assume much lower customer growth, but as the cohorts season, spend/buyer picks up. However, we do not believe these “non-core” buyers will season as well as prior cohorts (in 2020, management noted the 2015 cohort increased spend 5x) as they shop across multiple platforms in search of the best deals. (Also note that calling these buyers “non-core” is just for nomenclature simplicity to make a distinction between their traditional, higher-spending users and not used to imply they aren’t strategically important or somewhat valuable to JD). Rolling this up, we see a ~12% GMV CAGR for the next decade as very plausible (+25% y/y in 2020).

We then apply our Revenue Monetization rate we calculated above in the accounting section to get our Platform Revenue figure (defined in the accounting section as total revenue generated from underlying GMV for both JD and 3rd party sellers—in other words, adding JD’s revenues and the revenues the sellers make). After applying a mix-shift assumption in-line with what we last observed in 2016 and assuming it tilts more towards 3P over time as Amazon has, we get revenue figures for both their 1P and 3P businesses.  The 3P take-rate is informed by looking to Amazon, which disclosed ad revenues that equate to 8-10% of 3P GMV. If we assume JD’s average seller commission is 5% (2-8% typical range but varies for larger brands who can negotiate), then that implies that they have the scope to reach ~15% over time by reaching parity with Amazon. However, Amazon is still growing ad revenues rapidly (+57% y/y), so there could be further upside there. Nevertheless, since there are multiple scaled selling channels with consumer info, we assume competitive intensity keeps JD ad revenues below where Amazon is today, reaching just 13% in a decade (generally speaking, take-rates have also always been much lower in China). This could be a source of upside to our estimates.

Below, we translate JD’s two revenues streams (1P and Marketplace & Marketing) into operating profits. We back into gross margins by segment based on their commentary for JD Retail which we know to be at a 14.7% gross margin. However, the split of gross margin between the 1P and Marketplace & Marketing segment is less clear. We assume JD 1P was an 11% margin with the other segment higher at 65%. However, this may not be the cleanest way of looking at it as we unearthed in our diligence that there can often be gross margin agreements with 1P suppliers that require them to guarantee a gross margin through advertising should there be a shortfall. These agreements are more common with big brands where they want to maintain control over their selling prices, with the ability to lower it to increase volume should they decide that’s best, but JD who is placing a huge order wants to make sure they will make an adequate margin. In these circumstances, marketing becomes the “plug” where brands make up the shortfall of the agreement by spending on JD’s advertising. The increased advertising also helps sell more units that rationalizes the brands decision to lower the price, so it is a bit of a feedback loop. Nevertheless, despite the split between the two, we know that 14.7% 2020 is a correct figure for the two segments.

They give their JD Retail segment EBIT as we noted it prior, but we adjust it by applying 75% of unallocated costs which reduces their margin by about ~18% to 2.3% (we acknowledge 2021 margins are already trending higher and took that into account in our T+5 estimates).

This all comes out to an estimated ~5.5% JD Retail margin in 5 years, which expands to ~8.5% in 10 years. Informing these estimates is management’s earnings call commentary, particularly from 1Q19 and 2Q19. They reiterate that their direct selling business has a lower cost structure than best-in-class physical retailers and thus should have a 1-3 points higher margin. We look at Walmart which has ~4% EBIT margins and Costco which is at 3% as a base, which means margins from the 1P business should be between 4-7%. LTM JD Retail margins fully loaded for unallocated costs is at 2.5% today with some quarters reaching near 4%. We think a mid-single-digit margin is very plausible in the 1P business as they continue to scale their logistics network and gain some operating leverage from other expenses. On those calls, management also noted that the 3P service is much higher margin, and layering that in should add up to a high single digit margin.

We take our revenue estimates and apply a steady-state margin range below, taxed at 15%, to get a TSR (total stock return) range of 9-19% per annum based on a 15-25x multiple. We would emphasize again that this is only valuing their JD Retail Business cash flows and not their JD Logistics business or other new businesses. We were tempted to do a SOTP backing out their JD Logistics and JD Health stake, but decided that wouldn’t make much sense since they are still using JD Retail cash flows to support them both. Instead, we held their investment value flat at today’s market value, a conservative assumption no doubt, but we didn’t want to take an opinion on their ultimate value as JD has had a very mixed history with investments and breaking out their businesses (JD Finance/ Digits/ Technology is the only one with a long history, but that has been very lackluster with 3 reorganizations). Generally speaking, we are more positive on their other businesses, especially JD Logistics. Separately though, we are not sure if you back out JD Logistics if you are double counting as JD Retail likely receives favorable better financial terms (internally) than if it was an arms-length relationship (we would also have to account for the intersegment eliminations). Lastly, it’s unlikely they fully monetize these stakes in the short term, which if backed out today, would inflate the IRR (by unrealistically reducing the cost basis).

The other thing to consider is if you believe they are investing their excess cashflows well or destroying shareholder value. While we value just JD Retail, you can’t buy just JD retail—you are also getting the slew of new business initiatives that are burning cash in the near term. As mentioned prior, they have a mixed history in creating value outside of their core high-end retail ecommerce operations. While we see the rationale in most of their new businesses and it seems unlikely they burn cash indefinitely in businesses that aren’t working, poor capital allocation is always a risk and one that can only be comfortably mitigated against by a long history of good capital allocation decisions, which JD is lacking. However, over the last two years, their CROIC has inflected positively >20% (including equity investee investments), so if they just maintain those levels, they can achieve a strong return on retained capital overtime.


1) Regulation. Regulation has actually been beneficial to JD’s business (but definitely not the stock) over the past year. Most notable was the crackdown on the illegal policy “er xuan yi”, which means “two choose one”. This was an Alibaba policy that restricted merchants to either selling exclusively on Alibaba or lose the right to sell there. Since Alibaba was most merchants’ most important sales channel, they chose Taobao/Tmall and were not able to diversify their distribution. However, once the government started cracking down on this practice, there was a flood of merchants who could now sell on JD, which increases their selection.

Generally speaking, we think JD has been out of the crosshairs of regulators despite their prominence since they avoid some of the more sensitive technology and financial areas that are getting increased scrutiny, but also because they are invested in the physical infrastructure of China. The internet industry in China is thought more of as the world of “bits” and it is that sector that saw the most scrutiny. However, despite being a native online retailer company, JD’s operations include building out a lot of physical infrastructure in the world of “atoms”. Additionally, JD employs over a quarter of a million Chinese in warehouse and delivery jobs, which are considered good jobs given local alternatives. One theory is that this is exactly the sort of development China wants to incentivize, so when Alibaba and Tencent announced 100bn RMB donations to the Common Prosperity Fund, there was no similar commitment from JD.

No matter, regulation is still a risk for JD and could run the gambit from future donations to more oversight which hamstrings operations. It is also possible (likely?) that Tencent will not be able to assist JD as much as they have in the past. A limited future value-add sounded like part of their rationale to divest their stake. Tencent opening their platform to Alibaba is also a negative competitive development for JD.

2) Capital Allocation. As we have mentioned, JD has a mixed reputation as an investor and poor capital allocation, regardless of how successful their core business is, will lead to poor investor returns.

3) Consumer Financial Health Degrades. JD’s success is almost entirely levered to the health of the Chinese consumer, and any adverse change, whether temporary or structural, would negatively impact their business. While there will always be economic swings, downturns can be brutal as competition intensifies and retailers cut their prices to stay afloat, which can have lasting impacts on the industry.

4) Franchising Dilutes Brand. JD is lending out their brand to a variety of 3rd parties who may not provide the same quality of service consumers expect from JD. Poor experiences under a 3rd party that is branded “JD” will make a consumer think less of JD, especially as they will probably not know it isn’t actually operated by JD.

5) Market Growth Slows. Part of our JD thesis is that there will be multiple ecommerce winners. But if the market shrinks, then competition can become much more brutal as everyone fights over a shrinking pie, driving already low profit margins down further.

6) Competition. We have talked at length about competition, but there is always the risk competitors out-execute JD or come up with something new that JD isn’t able to easily counter, ceding consumer and market share.

7) Losses never stop. They have several money-losing operations, and while we look at the “core retail” business to see underlying profitability, this could be illusory as they never stop plowing profits into money losing businesses.

8) Business Conflicts. With their various business units, many of which are becoming standalone public companies, it is likely that their interests stray from JD shareholders’. Services that were once contractually provided to JD at set values could be renegotiated in favor of the standalone company. It is worrisome to think the leverage JD Logistics could have over JD in a negotiation, given they could grind the operation to a standstill if they desired. This isn’t a risk today (JD still owns 80% of JDL) but could be in the future. Additionally, when spinning off a company, the value may not proportionately accrue to JD shareholders.

9) Tax Increases. They paid a 13% and a 3% tax rate for 2019 and 2020, respectively. While we assume a statutory rate of 15% going forward, but it could be higher as they currently enjoy both the Key Software Enterprise tax break as well as the preferred industry tax reduction. Removal of these breaks could mean taxes jump to 25%, or worse, if China feels the need to raise rates further. They also benefit from other tax breaks like the ability to deduct 150% of R&D expenses for tax purposes, removal of which would further increase their rate.

10) VAT. An increase in the VAT would increase prices and could cause a slowdown in consumer demand.

11) Land Rights. They purchase land and require land rights to operate their logistics facilities. In the past, this has been a competitive advantage for them as it is not simple to get land rights, especially when it is from government-owned land. If this becomes harder in the future or if they lose their existing land rights, their logistics operations could be impaired.

12) Trade restrictions or tariffs. Restrictions on international trade can hurt JD’s international businesses and domestic businesses that sell international products. Another trade war could result in higher prices and thus dent consumer demand.

13) Accounting Fraud. International companies have less stringent auditing processes than US ones, which when combined with very opaque financial statements and corporate structure, gives life to the opportunity to defraud investors.

14) VIE/Corporate Governance. As evidenced by the Ant Financial/Alipay debacle whereby shareholders lost their full interest in the business, VIEs have tenuous claims on the underlying assets. There is ample opportunity to move assets at will and shareholders have little or no recourse. Investors do not have the same voice in corporate matters either.

Summary Model.

Below is our JD summary model. We provided this so readers can get a sense of the different pieces of the P&L and how everything interacts (as well as to give our members who want the excels the ability to manipulate them as they wish). We do not spend much time getting these estimates inline with our other exhibits, but one take-away is that you have to be relatively aggressive with the revenue growth drop off to get the 1P business to match our 1P revenue build numbers above. It is up the investor to judge what they feel are reasonable assumptions, but we note that even though we model a ~12% 1P revenue CAGR versus their +26% 9M21 y/y, growth can fade much quicker than you can realize, as evidenced by BABA’s recent earnings. This can or cannot be problematic depending on how much growth was factored into what you paid for the business.


Thank you so much for your support, and we hope you enjoyed our JD write-up! Next month, we have decided to dive deeper into JD Logistics, which will help further your understanding of JD. We think a lot of what JD is doing is very interesting, and will continue to write about new developments and earnings, so stay tuned! We will also be opening up a new JD Discord channel for you to discuss JD or anything you liked or disagreed with in this piece! Despite this running well over 20,000 words, we have still not touched on many aspects, so please feel free to drop any questions you have in the Discord!

Thanks for reading this months deep-dive! The conversation isn’t over though, please join our Discord to post questions and comments in the new JD section! Thank you again for your support, and we hope you enjoyed it!

*Disclaimer: People who have helped craft this analysis may individually (or through funds they work at) have a position in the company. Absolutely nothing in this report is investment advice nor should it be construed as such. We make no claims to the veracity of all facts and figures presented in this report. Certain figures could be stated erroneously and certain analysis could be incorrect. Please see the full disclaimer linked above.

DJY Research, its author(s), any of its contributors, and any funds they may advise at may own securities in the companies, or related companies, written about on this website.

All rights reserved. No content on the DJY Research website can be redistributed without the express written consent of DJY Research.

%d bloggers like this: