Building the Online Marketplace 2003 to 2005

As described in chapter 2 and appendix A, Taobao emerged in a relative vacuum of technologies and services that would enable online retailing. Taobao’s early task was very simple: get more people to buy online. Consequently, Taobao initially focused on putting the core constituents of the marketplace online: products, merchant-consumer interactions, and transactions. After this effort, Taobao became a functional online marketplace. It was a very simple version of what would later be understood as network coordination. Taobao got all the basic parts of the marketplace online, where they could interact through our platform. At this stage, Taobao lacked data intelligence, although the most basic online product database and bare-bones product categories laid the foundation for later datafication and for market mechanisms for matching buyers and sellers.

From the beginning, Taobao had a clear purpose: populate the marketplace, and establish mechanisms that facilitated direct connection and interaction among users. Initially, the Taobao website was essentially an online bulletin board. All users were created alike: everyone had come to do business in some way, whether that meant wholesaling or reselling, selling goods from their homes or presenting goods of their own creation. Slowly, sellers began to specialize and expand into multiple product categories.

Our internal KPIs for the first three years of the business were the buildup of three critical masses: products, merchants, and buyers. We also worked to forge connections between the constituents of the market. The Taobao forum allowed groups of consumers and merchants to connect and discuss common challenges. This community was especially useful in spreading know-how and helping newcomers to learn the ABCs of online selling. The Wangwang instant-messenger application connected consumers and merchants so they could discuss product offerings or transactions. And the assignment of credit ratings to sellers helped tremendously in overcoming the lack of trust in an online environment.

Building the Coordinated Network 2006 to 2008

With the basic ingredients for growth in place, the Taobao marketplace began to tackle the many challenges created due to increasingly sophisticated transactions on the website. It had to offer more functionality, most of which could not be handled by the platform alone. As needs surfaced, new or existing participants found ways to satisfy them. Individual actors began specializing or exploiting their own skills by becoming service providers.

In economic terms, as more and more actors joined the network and as product offerings and categories proliferated, growth of various functions and categories began to jump-start self-reinforcing network externalities. They sprouted from every surface of the marketplace, powering the market’s growth engine. This stage demonstrated Taobao’s potential for exponential expansion, as yearly GMV grew tenfold from 10 billion to 100 billion RMB (US$1.54 billion to US$15.4 billion).

The Emergence of Smart Business 2009 to 2012

In 2005, Alibaba Group acquired Yahoo! China as terms of its agreement with Yahoo!, one of the Group’s largest investors. Over time, teams and technology from Yahoo! China integrated into the platform. By 2009, the stage marking the true beginning of smart business in the Taobao marketplace began. The already complex network was reinforced by data-intelligence technology brought into the network by teams from Yahoo! China working on search. In these crucial years of the platform’s growth, the forces of network coordination and data intelligence began to build off each other.

The network expanded in breadth and depth. Taobao witnessed phenomenal category expansion, expanding into new markets. At the same time, the network incorporated even more actors and business functions. More professional merchants, including sellers of offline brands, began to operate on the platform, prompting the founding of the Taobao Mall, which would eventually become Tmall. Fueled by search and ad technology, the ecosystem incubated completely new organisms, such as the Tao brands. These brands were started by entrepreneurs on the platform without relying on any brick-and-mortar presence. Many grew to become recognizable national brands.

Taobao’s yearly GMV exploded from 200 billion to 1 trillion RMB (US$31 billion to US$154 billion). The double helix of smart business was complete, and it formed an extremely powerful engine for growth.

Exponential Growth through Smart Business 2013 to 2017

As reflected by the richness of product offerings on the Taobao app, the Alibaba ecosystem has grown stronger and stronger in the mobile era. The coordinated network keeps expanding, creating important new roles through social commerce, such as new content creators. At the same time, data intelligence has spread to every corner of the network, growing both more powerful and more effective. This growth is exemplified by the transition from search to recommendation throughout the platform, as well as new upgraded uses for AI technology, such as fraud detection, customer service, and automated design.

The combination of these two forces has made the Taobao app flexible, dynamic, intelligent, and powerful. The continual emergence of new actors and business models, chiefly the web celebs, testifies to the health and vitality of the Alibaba ecosystem and to Taobao’s continued strength in network coordination and data intelligence. As manifold networks of retail, finance, and logistics grow smarter and better coordinated, the entire ecosystem unfolds into new heights of prosperity.

The Birth of Taobao

In 2003, Jack Ma assembled a group of eight employees to begin work on a secret new project. Alibaba had been in business for four years. Its B2B platform was profitable and growing quickly. But Ma had bigger designs on the Chinese market: he wanted to reconstruct China’s antiquated and poorly developed retail industry.

In May of that year, Taobao was born. It began as a small, cute, and unassuming forum for selling goods. Ma and his fellow coworkers uploaded as many items from their own apartments as they could to rapidly populate the platform and give the appearance of activity. Little did they know at the time that their little forum would become the world’s largest online marketplace.

In the first few years of its life, Taobao successfully fought off the incursion of eBay, the world’s top e-commerce marketplace, into the Chinese market. Many other books have discussed Taobao’s tactics for beating eBay. Here, I want to discuss Taobao’s larger strategy for developing a more robust and flexible business network. Much of this strategy is only discernable in hindsight as the cumulative result of a multitude of small decisions, many of which were made to fit with Alibaba’s values and beliefs about the connected future of the e-commerce industry.

Taobao began by taking advantage of eBay’s greatest weakness.eBay’s markets are rigid by design. The company makes its money through listing fees and commissions: if buyers and sellers connect but do not eventually make transactions through the platform, eBay makes no money. Thus the core logic behind eBay’s marketplace is to mediate much of the interaction between its users to prevent buyers and sellers from making free connections. Transactions must be tightly controlled, and sellers primarily have to deal with eBay itself. The marketplace is a clearinghouse for products and mostly idle consumer goods waiting to be transacted. It is not a solution for enabling more-complex business activities.

In China, buyers and sellers had needs that the rigid structure of eBay could not satisfy. The country had given birth to an enormous number of small sellers without business and access to consumers. Buyers didn’t want to bid on only a few unique products; they wanted access to a vast universe of products from across China. And sellers wanted access to that multitude of sellers and the chance to compete on products and services. More importantly, China’s retail industry was extremely undeveloped. As the marketplace grew, sellers also needed solutions that would improve their businesses.

Taobao’s strategy was to build a bustling two-sided market, catering to both buyers and sellers. In 2003, employees did everything they could to fill the market with as many products as they could find. The next year, the goal was to bring in as many sellers as possible. Finally, with a critical mass of sellers, Taobao began 2005 by attracting buyers to a website that advertised itself as selling anything you could possibly imagine. In retrospect, this transformation was a classic exercise in building a platform through network externalities and using two sides of the market to build off each other. More sellers brought more buyers; better sellers brought better buyers.

Finding products to sell was often an exercise in tedium, as employees resorted to any tactics they could to catalog as many SKUs as existed in China. The platform attracted sellers through Ma’s famous policy of keeping Taobao free. Opening a storefront was free. Listing products was free. With no entry barriers, apart from the time and energy needed to learn how to use the website, sellers flocked into the marketplace. And as soon as buyers realized that Taobao housed every product on the face of the earth, they came in droves as well.

The constant flurry of activity that resulted knit a market out of a dense web of interactions. Bringing ever more players onto its platform and allowing them to connect and do business with one another engendered increasingly complicated coordination technology and systems substantially facilitated by Taobao. That coordination capability enabled progressively more advanced business models.

From Warmth to Trust

Despite its growing popularity, Taobao was not just a forum. It was a marketplace, and markets are not sustained by mere social interaction. They require means of exchange and other tools and structures to authenticate, safeguard, and enable healthy transactions and trust. The offline market economy took thousands of years to develop—from bartering beans for hides to credit card transactions for something produced halfway around the world. Taobao was faced with the same challenges, except in an online environment.

Taobao’s two technological keys that bootstrapped its market evolution were Alipay, Alibaba’s platform for secure escrow payments, and Wangwang, Taobao’s chat feature. By convincing consumers that online buying was safe, Alipay provided the foundation of trust that allowed Taobao’s transactions to proceed smoothly. Wangwang enabled sellers to build rich interactions with buyers and to offer differentiated services. Some sellers even became famous for the humorous exchanges they had with customers!

Any discussion of the early years of Taobao would be incomplete without mention of Alipay and escrow payment. In 2003, China lacked trust on a multitude of levels. There were no checking accounts, credit cards had just been introduced to a very small user base, and no trusted back-end settlement protocol existed. Early transactions on Taobao were small-scale and local, mostly restricted to offline trading in the same city. After connecting online, buyers and sellers would meet in person, inspect the goods in question, and physically exchange money. Transactions were slow and costly. More importantly, the setup was a recipe for fraud and cheating. The early days of Taobao abound with stories of sellers who scheduled a meeting with a buyer and arrived promptly with their products, only to find that before they could react, the buyer had driven past them on a bicycle, grabbed the product out of their hands, and pedaled swiftly away before anybody was the wiser. Clearly, the direct physical swapping of goods and money was not a recipe for safe, nationwide, frictionless trading that might lead to explosive growth.

The escrow solution solved these problems. Buyers would first entrust their money to Alipay, in the early days by wiring to Alipay’s bank account. Alipay would then hold the payment in a secure account until the buyer received the good he or she had purchased and confirmed its quality. Thereupon, Alipay would reward the seller with the earnings. For buyers unfamiliar with and suspicious of e-commerce, escrow provided the peace of mind needed to trust their hard-earned cash to a transaction with a stranger, no matter who the merchants were or where they were located. For sellers, escrow payment greatly broadened their client base. Now, instead of several thousand buyers from the same city, sellers could sell to hundreds of millions of consumers across the country. Trust had been built to undergird the marketplace, and the gates were suddenly thrown open to all. From this humble beginning, Alipay later built an online payment network that connected with all the online banking services in China and hence became China’s major online payment gateway.

Your Credit, My Credit

After escrow payment, Taobao’s next step was to establish credit ratings to digitalize reputation and thereby inspire good service and better performance. In an online setting, reputation is an asset like any other, albeit invisible and difficult to transfer. Taobao was by no means the first e-commerce platform to use such rating systems to quantify reputations; eBay pioneered credit ratings in online marketplaces, with a standardized reputation score system applied in the same way to every user. But Taobao’s employees quickly realized this reputation system pioneered by eBay was insufficient.

The central problem was that users engaged in both buying and selling behavior. In the early days of Taobao, all conversations and transactions occurred through a central forum. Users who wanted to buy goods found items and contacted other users. Essentially, any user could buy or sell as he or she wished. When a transaction was completed, the user’s credit score would improve. However, this design meant that a user who had bought a lot of items but had no selling experience seemed just as credible as, if not more credible than, an experienced seller. There was room for fraud: the signal of reputation within the market was ineffective.

For Taobao to become a major retailing platform, amateur sellers had to begin to operate like professional sellers. Taobao decided to separate reputation scores into buyer and seller scores. A user who had sold many items would not have a high buyer score unless the same user had similarly bought a lot of items, and vice versa. Furthermore, buyers and sellers would rate each other after the transaction was complete. In Chinese e-commerce, transactions are only completed when the buyer confirms delivery of goods.¹ The buyer’s confirmation not only alerts escrow payment systems to send the money to the seller, but also allows the buyer and seller to rate each other.

By splitting credit scores, Taobao also separated the buying and selling roles institutionally and technologically. For any selling activity, the credit score served as a concrete incentive to provide quality products and good service, as sellers with better scores attracted more buyers. Good buyer scores also supported the evolution of a healthy ecosystem, because the presence of good buyers encouraged the entry of more sellers. These credit scores were only the first step of deeper and broader differentiation in the market—a prelude to the inclusion of more and more roles in the platform and hence to a more and more complex ecosystem.

Settlers in a Virtual Marketplace

At first, the new inhabitants of this new market were all alike. Taobao’s early employees spent their days on the Taobao forum talking to sellers and facilitating transactions. Sellers were buyers, and buyers were sellers: in the first few years, when consumers had not yet grown accustomed to online buying, many sellers bought and sold from each other. The Taobao offices played host to a steady stream of local merchants discussing platform design and marketing tactics, suggesting rules and other features, and criticizing and complaining about shortfalls and imperfect user experience.

But soon, Taobao experienced the natural evolution of markets that in an offline setting might have taken decades. Roles began to separate, evolving new forms of cooperation and competition. Like cells dividing, the market grew more complex. I just described how credit scores established two roles characterized by very different behaviors. However, Taobao leadership quickly realized that buyers and sellers alone were not enough to sustain a functioning marketplace. Sellers needed more support to effectively engage in the complex practice of commerce.

The earliest of these new roles to appear was the customer-service representative. One of the biggest differences between Chinese and Western e-commerce is the role of in-house customer service. Whereas websites such as Amazon and eBay only have official customer hotlines, every storefront on Taobao has its own account on the official Taobao instant-messenger, Wangwang. The site puts a human face on each storefront: customers can ask about offerings or dicker over price. They can even chat about other things if they want. And as described earlier in the book, Wangwang customer-service representatives use an unusually informal tone, even calling customers “dear,” injecting warmth into an otherwise impersonal experience.

As explained in chapter 2, besides the traditional function of polite, helpful pre- and post-purchase service, the customer-service roles on Taobao differ in that customers expect service to be available 24/7. Taobao’s customer service has provided employment opportunities for tens of millions of Chinese people, including students, people with low incomes, people with little education, and people with physical disabilities that make working in the traditional economy difficult.

The Web We Weave Together

Most sellers in these early days were young, inexperienced in the retail industry, and often unable to adequately execute the difficult task of operating an online store. Early sellers were often individuals, at most a very small team, learning on their feet. They rarely manufactured (much less designed) their own products, often procuring goods from wholesale markets or factories around the country, in itself a challenging task in the still underdeveloped Chinese economy. While doing business on Taobao and processing transactions through Alipay was free, sellers had to learn to use the growing array of tools provided by Taobao for managing a storefront and connecting with customers. Thus, some sellers saw an opportunity to provide supporting services for quickly growing merchants.

With more sellers selling to more buyers, the ecosystem began to grow in complexity. But, in marked contrast to eBay, Taobao showed an impressive ability from the very beginning to engender connections and even organizations outside its official purview. Precisely because the world of e-commerce in China emerged without models or precedents, whereas eBay was born within the well-developed US retail environment, Taobao’s value as a marketplace quickly began to spill over. Sellers formed informal networks as more and more service providers flocked to the platform, driven by its ever-increasing potential business value.

Even in the first few years, sellers began to openly organize into local clubs and other associations, forming unofficial business forums and unions. These early spaces, offline accessories to the online marketplace, allowed sellers to connect, sharing knowledge and helping each other adjust to the new world of online selling. Some of these forums were localized, where sellers in one city or province would meet periodically. Sellers of all ages and backgrounds, from different product categories and industries, came together in search of business opportunities and improved performance.

Gradually, dedicated online forums for sellers emerged, among them the most famous being Weiya and Paidai. At the same time, a new breed of sellers began to distinguish themselves from the rest of their contemporaries. Whether through innate aptitude or pure happenstance, these sellers swiftly apprehended the mindset and strategies needed to succeed in those early, chaotic years. The group, however, did not conceal its knowledge or hoard the fruits of success. Instead, these sellers found meaning and even material reward from sharing their experience, for example, by organizing lectures online and offline; sharing techniques and tricks for better online marketing, including customer service and dispute resolution; and working with factories to improve supply-chain management. And as one might expect, many other sellers were willing to pay for the valuable, trustworthy information from their experienced peers.

After Taobao quickly realized that this group of lecturers performed an important service for sellers, it created the so-called Taobao University, a department within Taobao to certify and support the Tao U lecturers. Taobao’s goal was not to manage and standardize the rapidly growing market for education. By encouraging sellers to self-organize and cooperate, Taobao gave the most active instructors a platform for growth. Tao U lecturers quickly gained prestige and recognition and, with Taobao’s support, expanded the range of possibilities for earning money from their lecturing. Over the next decade, the lecturers exerted an often-overlooked but immensely important force throughout the market. Lecture after lecture spread waves of knowledge that helped sellers, especially the smallest and least experienced sellers, perform better in new and unfamiliar surroundings.

Category Fission

By 2005, more and more sellers had joined the platform, bringing with them ever-more items for sale and attracting an enormous number of buyers. The platform’s growth was exploding. With that explosive growth, Taobao had already begun its evolution from a simple online marketplace into a retail platform. Starting in 2006, Taobao began to build the core resources and mechanisms that would support its later growth in network coordination and data intelligence.

At first, Taobao’s breakneck growth brought practical design challenges. With an ever-growing array of buyers, sellers, and products, how should the platform allow these various groups to find and interact with each other? How should the network that connects buyers and sellers be designed to grow?

I have discussed how the website changed from a forum to a full-fledged market, but I have paid less attention so far to the design and infrastructure of the website, and hence, the market. The first incarnation of Taobao looked like a forum, but in actuality it was a very flexible database of product information. That database’s core functionality was to datafy products by encoding their characteristics into data, which allowed buyers to browse, find, and buy items on Taobao. The database thereby enabled Taobao to live up to its early reputation: you can find anything you can imagine.

The product database essentially created a series of “boxes” in which sellers could slot in their items, from apparel to food to electronics. This database acted like an index, almost like one you would find in a library, listing each product along with a predetermined set of features such as size, material, and brand. The platform could in theory create an infinite number of boxes with an infinite number of features attached to each box. Because growth in product categories necessarily implied growth in types of sellers, the expansion of the product database drove the expansion of the network. Each time a new box appeared, so too would a new group of sellers flock to the market, in turn attracting a new swath of consumers.

Early in the marketplace’s evolution, a certain practice called “category splitting” became quite common among Taobao’s employees to achieve growth. For example, in the women’s apparel category, employees would notice a type of product that was particularly popular, e.g., dresses. By creating a subcategory of dresses like “miniskirts,” employees enabled buyers and sellers to find each other much more easily. Browsing efficiency increased, which naturally improved sales. At the same time, sellers could further specialize their products and gained better and better access to precision volume, increasing conversion rates and business. From the platform’s perspective, more categories meant that transactions from more and more industries were going online. Taobao thus expanded the boundaries of the marketplace through datafication.

For category managers, as long as the website was growing, splitting categories was an easy and effective way to drive business growth. In the words of one of Taobao’s early category managers, “split a category, and GMV spiked.” Separating and defining categories had such an effect because, even as early as 2005, the growth in buyers greatly exceeded the growth in sellers. In those first earlier years, the main channel that consumers used to find products was the category listings. Give buyers a new category to explore, and they would buy like gangbusters. And for sellers, each new category meant greater specialization.

At the same time, category splitting laid the groundwork for the platform’s future data intelligence. The data produced as a side effect of category splitting became richer and deeper, and category managers could see when demand for certain categories was growing quickly. But there was a bigger problem with the category system: increasingly fine-tuned differentiation between products ultimately did not improve efficiency. As Taobao expanded, it became immediately obvious that the existing categories were far from adequate to keep up with the dizzying variety of new products and new consumer demands.

Fundamentally, the category managers’ decisions were imperfect human decisions. By the time the market had grown to include millions of buyers, even the category managers had to admit that their job was next to impossible. How could one or two employees responsible for an entire category with up to hundreds of thousands of storefronts hope to select products efficiently? Splitting categories without restraint or principles would eventually mean efficiency losses and would perhaps even hurt the marketplace. The network had grown, but it needed stronger tools of data intelligence to keep up its efficiency.

Logistics

As the e-commerce industry began to pick up, Taobao’s consumers, merchants, and employees quickly had to cope with a pressing social problem: the sorry state of Chinese logistics. As China’s economy modernized and opened up in the 1980s and 1990s, industries emerged from nothing. But the development of supporting commercial infrastructure often lagged behind. Logistics, especially courier service, was no exception. Though China’s national postal service covered the country, its service even to this day is often too slow for effective commercial use. International couriers such as FedEx and DHL sent packages into and out of China, but not to any significant degree within the country. For the better part of two decades, businesses had little choice when it came to consumer shipping. (This lack of delivery infrastructure was one of many factors that prevented Chinese firms from creating companies and business on a national scale before the 2000s.)

Beginning in the late 1990s, several entrepreneurs from a county called Tonglu, located in Zhejiang Province to the southwest of Hangzhou, began to offer courier service for companies and individuals in the commercial corridor around the Shanghai River delta. These companies very quickly began to expand across the country to fill a nascent need for quick and affordable delivery. More importantly, as these companies expanded in the booming years after the turn of the new century, they also caught the wave of e-commerce.

When Taobao was founded in 2003, it unwittingly stepped into a complicated, shifting map of courier service providers across the countries. The so-called Tonglu Mafia, all from the same county in Zhejiang, founded four of the largest companies. SF Express, a high-end express-delivery service originally founded to handle deliveries between Guangdong Province and Hong Kong, began to expand quickly into the rest of China’s high-end market. Another dozen or so smaller players operated throughout the country, staking various regional claims while aiming for national dominance. How was a growing e-commerce marketplace to decide how to work with this many companies operating at vastly different scales and levels of expertise?

By the time the basic structures of Taobao’s marketplace had solidified in 2006, company leadership quickly saw that the complicated state of the logistics industry created serious difficulties for both buyers and sellers. Because of regional differences across China and a convoluted system of franchising within the logistics industry, pricing and service quality varied drastically across geographical regions. Even in the same area, different logistics companies could offer very different quotes for the same service. Sellers operating on Taobao had to contract their own logistics partners, which operated independently from the website.

The situation for consumers was a minor nightmare. Even once the seller had quoted shipping prices and times, consumers had to use badly designed and unstable websites created by the logistics companies to access tracking data related to their purchase. The problems spawned by this innocuous-sounding arrangement could be truly treacherous. For one, because the logistics companies possessed less-than-stellar IT capabilities, rapidly escalating traffic from Taobao consumers checking delivery statuses soon gave rise to regular server crashes and blackouts. More dangerously, often exploited by hackers, security vulnerabilities hidden in the logistics companies’ web pages were stealing user information, engaging in phishing, or even infecting consumer computers with Trojan viruses.

Taobao needed to provide better consumer experiences on fulfillment, but what to do? It could emulate Amazon, well known throughout the industry for developing in-house logistics capabilities. But the cost and difficulty of managing a fulfillment fleet was not feasible given the exponential growth of Taobao’s platform model. In 2006, Alipay experimented with an optional service whereby sellers could send pricing inquiries to logistics companies. The idea was to make prices more transparent and standardized. (Alipay also intended to reserve the right to fine logistics companies if they did not meet their promised standard of service, including time to fulfillment.) The service didn’t take off. Only two major providers and one minor company located in Shanghai were willing to work with Taobao, most likely because it meant ceding no small amount of control to an outside party. Sellers had little incentive to use the optional and very limited service, and consumers felt little impact when they did use it.

In response, Taobao took a different approach. It would embed the act of fulfillment into the core transaction flow of the platform and, in doing so, would implicate third-party logistics companies into the larger ecosystem.

To understand what this new approach meant, consider Alipay’s escrow system. In Chinese e-commerce, once the consumer finishes a purchase and completes the checkout, the money paid is held in Alipay’s system until the transaction completes. The flag for completing the transaction is again incumbent on the consumers, who confirm that they have received their item as requested. Once Alipay receives that confirmation—and only when it does so—the merchant receives payment. The fulfillment process is deeply intertwined with the escrow system. Many situations where the consumer is not satisfied and refuses to complete the transaction have nothing to do with the seller. For example, the parcel never arrived or it was damaged in transit.

Taobao’s retooling of the transaction flow embedded shipping into the escrow workflow. In 2006, Taobao consolidated transaction software and product under a single team, including portions of Alipay transactions, and required sellers to enter a package tracking number into Taobao’s system after shipping physical goods. Without a tracking number, consumers could not complete the transaction. By integrating tracking numbers, Taobao pulled tracking information from the logistics companies, allowing consumers and merchants alike to track their parcels en route. This change had profound implications for the platform, for consumers, and for the logistics companies.

At first, the vast majority of the logistics companies were less than willing to share core operating data with Taobao. But six months after the first company successfully integrated its fulfillment platforms with Taobao in 2007, that company’s delivery volume spiked. The others took notice and warmed to the idea. By the end of 2008, all of China’s major logistics companies had integrated with Taobao. Though sellers still had to contract with logistics companies on their own, the process became more standardized and consumer complaints decreased. More importantly, from the perspective of the platform, logistics companies had been integrated into Taobao’s core mechanisms. The ecosystem had expanded. As Taobao began to experience exponential growth, the logistics companies that were now connected to the platform rode the same wave of growth.

From Yahoo! China to Alimama

Utilizing Yahoo! China’s technology, Taobao founded Alimama, a marketing technology platform, in 2007. The platform first brought advertising into the Taobao ecosystem. Alimama’s advertisements came in three general forms. First, keyword-based ads embedded in Taobao search pages allowed sellers to post ads within the Taobao marketplace. Second, Alimama created an ad exchange that allowed ads to be posted to external websites. Third, the Taobaoke platform created a market for affiliate marketing, quantifying and paying for ads on third-party websites, that drew traffic in from every corner of the web. (See the later discussion of Taobaoke in this appendix.)

Online advertising differs from offline advertising in its precision and ability to quantify behaviors and reactions. If a user views an ad you have posted through Baidu, Baidu will record that impression. If a user clicks on an ad from Taobao and eventually goes on to complete a transaction, Alimama’s advertising engine will alert you. More-advanced advertising technologies can also give ad buyers a clear picture of who their audience is, the audience’s preferences and interests, and even what sorts of products people might be interested in. It is next to impossible to measure the exact impact from, and return on, offline advertising expenses. By contrast, online advertising directly measures and tracks connections between sellers and buyers.

Yahoo! China possessed yet another important technology that had significant influence on Taobao’s ecosystem: search. Not surprisingly, in 2008, Taobao began to continuously upgrade its rules for popularity search. For example, the upgraded search assigned more weight to sales, picture click rates, and conversions. Taobao began to encourage certain seller behavior through its design of search algorithms, encouraging the healthy growth of sellers.

The New Fusion: Search and Ads

In combination with ads, search took on a new meaning as a metaphorical internal combustion engine for the platform. By 2009, many sellers had begun using Zhitongche, Alimama’s keyword-based advertising product. Bidding for search keywords meant exposure to buyers, which meant more sales and improved metrics. Improved metrics meant better search rankings, and again more exposure to buyers through search. Combining search and ad products in this way brought explosive growth for enterprising sellers.

To understand how the combustion engine works, consider how retail products are positioned on Taobao. To avoid cannibalizing offline sales and even damaging the brand, the average company in China does not sell the exact same goods in online and offline channels. Every retail channel needs a dedicated product strategy to deliver a distinctive experience for the customer—an experience that is customized for each channel. In the average Taobao storefront, a seller will typically offer at least three tiers of products, which are differentiated mostly on margin: “exploding” items, margin-driven items, and showcase items. (In women’s accessories, for example, an “exploding” pair of tights might be priced at a 1.6 ratio to cost, while margin-driven and showcase items might instead offer a ratio of 1.8 and 2.1, respectively.)

Besides contributing to the bottom line in different ways, each of these tiers of products has a specific function within the store and on the platform. Margin-driven items are easy to understand; these are the higher-margin items that every retailer wants to sell. Showcase items are often placed in a store to shape the general brand identity and positioning, with many sellers keeping very low quantities in stock.

The most interesting of these categories is the exploding tier, a term that originally arose on Taobao in the early years of the platform and soon grew to represent the basic operational strategy for SKUs in associated categories. Exploding items (baokuan) are cheap, but with quality surpassing their price. In fact, they are often priced at a level that would in and of itself be unsustainable for the seller. The attraction to heavily price-conscious customers is obvious, but exploding items are not offered to exploit economies of scale. They work to build reputation for a store and gain higher weight in search algorithms.

In e-commerce, placement and traffic is chiefly determined by reputation, which is quantified by historical sales volume and product review scores. The best way to quickly accrue reputation is to offer a high-quality product at low-quality prices, thereby attracting a critical mass of customers and positive responses. The ensuing reputation data in turn boosts the placement of, and traffic to, the store, ensuring that more and more consumers can visit the store. From then on, the seller’s job is to drive traffic into conversions of its higher-margin items.

This tripartite tier of items works in Chinese e-commerce because of the technological mechanism by which reputation accrues and how it influences operating strategy. In an offline shopping mall, a store is perfectly within its rights to display a high-quality product sold at a low price in its window. Odds are, it will attract consumers, but the long-term payoff for the brand will probably be limited. But online, a store’s very placement in the online mall is inextricably tied to its reputation and performance. It is almost as if the aforementioned store, after selling numerous sale items, could move closer to the escalator.

Many sellers quickly learned that they could use advertising and search to quickly grow their brands. First, they would invest in advertising, funneling traffic into stores to purchase exploding items. Sales activity would boost storefront metrics and afford better search placement, which in turn brought more traffic and sales. Sellers who understood the system could quickly take advantage of the platform’s resources of data intelligence to grow their businesses.

As time went on, this internal combustion engine, too, began to impair the ecosystem’s long-term growth, as larger sellers grew larger and larger, creating imbalances on the platform. Thus, Alibaba would invest in recommendation technologies in later years to promote more-balanced seller growth across the platform.

Taobaoke: The Growing Network

The growth of advertising technology was in many ways the beginning of Taobao’s most important period of explosive growth and the first of many factors that allowed the small Taobao ecosystem to exert influence on the business environment outside the platform. In addition to external advertisements hosted through Alimama’s exchange, Taobao’s other network of affiliate marketing, Taobaoke, also did considerable work in expanding the network.

Taobaoke began as a fairly simple idea. If I bought a piece of clothing or a household good that I particularly liked, I could recommend the product to my friend. If the friend bought it, I could earn a small commission from the seller as a way of saying thanks for helping the merchant gain a new customer. In 2008, when the Taobaoke product first went online, this logic of affiliate marketing was already established in markets like Japan and the United States, with the prime example being Google AdSense. This platform allows websites to embed advertisements on their web pages and monetize the ads’ content. If a visitor to the website clicks on one of Google’s ads, the website earns a small commission. For many small websites that lack the resources to develop advertising and sales teams, AdSense provides a crucial lease on life.

Taobaoke differs from AdSense in one fundamental way: the Chinese commissions are based on sales, not clicks. (In technical terms, this is the difference between cost per click and cost per sale.) If a visitor to a small media portal is looking at an article on what sort of porridge to brew for weight loss and radiant skin, the individual might see a link to Taobao selling cloud-ear mushrooms. If that user clicks on the link and looks at the page but doesn’t end up buying a product, the website gets nothing.

This mechanism design completely changed the incentives for external websites. It was not enough to ensure that an ad got an enormous number of clicks: users needed to enter Taobao and actually buy products for the website to earn its commission. In practice, it was most profitable to guide as many people into the Taobaoke platform as possible in the hopes that some fraction of them would end up buying. Thus, Taobaoke encouraged savvy webmasters to drive traffic to Taobao.

As a result, when Taobaoke was introduced into the Taobao ecosystem, it brought far more than extra business for sellers. It expanded the boundaries of the entire network and opened up Taobao to countless external sites across the internet. As smaller websites began to rely on Taobaoke for more and more income, their traffic poured into Taobao, rapidly growing the network and feeding seller demand. These new sources of traffic meant more and more business for the store owners that participated in Taobaoke’s commission program. As more consumers entered the Taobao ecosystem, they became even “stickier” buyers, purchasing from more and more sellers. The network grew thicker and denser as it expanded.

Tao Brands: The Origin of Species

In just a few years, Taobao experienced the natural evolution of markets that in an offline setting might have taken decades. Roles began to separate, evolving new forms of cooperation and competition. Like cells dividing, the market grew more complex. One of the most interesting and exciting examples of that complexity was the Tao brands, brands incubated and grown completely online. (See the sidebar “A Case Study of UNIFON.”)

The Birth of Singles Day

In late 2009, employees at the newly separated Tmall sat down to brainstorm how to bring their new marketplace into the mainstream. Tmall was positioned as Alibaba’s platform for large and established offline brands, but many of those brands did not see a compelling reason to join the marketplace. A sale might be a worthwhile carrot to lure these flocks of brands to the platform.

Traditional Chinese retail cycles focused around the Chinese New Year in January or February, as well as changing seasons across the country. After copious research, the staff decided that the month of November would serve as an ideal bridge between fall and winter seasons in various categories and regions of China. This was the opportunity to create a new megasale, a Black Friday for China. Except, there was one problem. Between the National Day holiday on October 1 and late December (when the rarely celebrated Christmas can be invoked as a pretense for sales), China had no important holidays analogous to an American Thanksgiving Day.

The search persisted to no avail, until somebody suggested November 11, known among some young people as a tongue-in-cheek holiday for the lonely, unattached masses: Singles Day, named for the depressing singularity represented in the date’s numerical written form.

And so, Singles Day was born. The first year, with little preparation and only twenty-seven sellers participating, 52 million RMB worth (US$8 million) of goods were sold in a single day, to the utter stupefaction of all involved. Clearly, they were onto something: in 2010, the new holiday’s sales numbers continued to skyrocket, racking up 936 million RMB (US$144 million), more than the entire volume of products sold in Hong Kong in a single day.

Yet even as Taobao and Tmall were growing at a bewildering pace, e-commerce had still not entered the average retailer’s field of vision. For the offline brands that comprised an increasing percentage of sellers on Tmall, e-commerce was still only a minor part of their operations, often handled by a small portion of their marketing staff. On Singles Day in 2010, when demand first greatly exceeded the volume of goods that sellers had prepared ahead of time, many companies had to scramble to divert products from other channels. Yet despite the huge sales numbers created on Singles Day, physical retailers and malls still felt e-commerce was beyond the horizon.

But by 2011, the e-commerce tide was clearly coming in. Many local Chinese brands at all levels, and even foreign brands, had begun to open stores on Tmall. To avoid conflict with their offline channels, these brands began to offer differentiated goods only available online, or special packages and assortments for online consumers. Tao brands (e.g., UNIFON) also came of age around then. That year’s Singles Day saw 3.36 billion RMB (US$517 million) in transactions, including the first time a brand’s sales broke the 10 million RMB (almost US$2 million) ceiling in a day. (That apparel brand, GXG, in fact shattered the ceiling, selling 44 million RMB worth [US$7 million] of products.)

In the same way that coastal inhabitants know that an offshore earthquake presages a tsunami, everyone could see what was coming. When the curtain closed on Singles Day in 2012 with nearly 20 billion RMB (US$3 billion) in transactions, the phones began to ring. Indeed, the next day, I received a call from a traditional retailer who asked me almost verbatim, “Tell me again, how do I get online and start selling?” The sea change of e-commerce had arrived, and by 2012, everyone could see that those who hadn’t grabbed a spot on the boat might end up sleeping with the fishes.

A Mall That Evolves: The New Mobile World

In the fall of 2013, a sense of crisis had set in among Alibaba’s management. Mobile penetration was skyrocketing, and Tencent’s killer app WeChat was steamrolling its way across China’s smartphones. Yet across Alibaba, nearly all its products, platforms, and policies were designed for a world of PC computing. Both inside and outside the company, comparisons were made to companies like Nokia and Motorola, whose insensitivity to changing market conditions caused those companies’ precipitous decline from market leadership to obsolescence.

In late October of that year, Alibaba’s business units were restructured and resources were diverted from PC-era product lines. Management moved to redesign its marketplace and product line for a mobile environment. Taobao’s technical and product teams began to tackle the enormous technological and infrastructural challenges of converting a PC-based platform into a mobile application that would perform equally well on myriad devices across developmentally disparate areas of China. These reforms, though difficult at the time, ushered in a new wave of growth for the marketplace. Now, the vast majority of Taobao users are now completely acclimated to mobile shopping and payment, as evinced by the enormous share of transactions (more than 90 percent) that were completed on mobile devices during the first hour of Singles Day in 2017.

For PC users, the Taobao customer experience was overwhelmingly dominated by search, static category lists, and the occasional curated promotion page. The Taobao app, on the other hand, became, and still is, a multifarious collection of banner ads, live streaming, flash sales, recommendations from power users, written articles (the Taobao Headlines), social networking, crowdfunding, crowdsourcing, and more. In this one app, a user can find a cornucopia of product channels, from worldwide brands to deep discounts to fresh food to luxury goods. Sellers, products, and even business models drastically differ in different sections of the app.

Most importantly, data intelligence operates pervasively behind the scenes, in the app’s operation and in its constantly evolving structure. The Taobao app, like the ecosystem as a whole, is constantly changing and getting smarter and smarter. It is like a virtual mall where every storefront is customized to the customer and where even the layout of the mall itself can be rearranged. Flash-sale channels, for example, rank their products dynamically according to the customer. Even before a consumer begins browsing a particular channel, data intelligence is already working behind the scenes to guide the buyer to his or her desired product. To give another example, many of the pictures that users see to guide them into categories or specialty channels are already highly personalized. Different users that encounter the same channel will see different pictures.

Deeper network coordination and more-robust data intelligence drove GMV growth but, significantly, also incubated more and more new functions within the mobile ecosystem. In Taobao’s PC era of growth, the platform evolved from a forum of buyers and sellers to a marketplace composed of all sorts of supporting functions for merchants. The growth of these ISVs and the development of market mechanisms and technological infrastructure to support their growth proceeded organically as the market developed. After 2013 and Taobao’s transition to mobile, the marketplace continued to evolve. Even more roles and business models flocked to the platform. Taobao created another series of mechanisms and infrastructure for a new generation of ISVs.

The first group of new players in the mobile marketplace evolved from affiliate marketers and product recommenders that had done business on the platform for years. These new players operated by collating products on behalf of sellers and earning commissions. (Their offline analogue are buyers in the fashion industry.) In the mobile era, more and more of these aggregators began to exploit social media platforms to find and circulate their product listings. The distinction between recommenders and the group of users normally called influencers began to blur—very quickly, social media influencers and websites started to make money off commissions from Taobao. Taobao categories like Ai Guang Jie (“Love Window Shopping”) and You Hao Huo (“Good Find”) provided specialized virtual “racks” where these marketers could showcase their wares.

Behind many of these affiliate marketing channels, a common infrastructure for payment settlement allows entrepreneurs to make money on commission. When a third-party writer (a new type of ISV) creates marketing copy, they inevitably embed links to Taobao products in their articles. Conversions that occur from these links generate commission on a cost-per-sale basis, with 20 percent going to the writers and 10 percent going to the platform. For many sellers, this commission is a acceptable marketing spend, especially for traffic that, thanks to data intelligence, can be more accurate than other traditional search and advertising channels.

Astute readers may notice that this payment infrastructure resembles the Taobaoke product described earlier. (Taobaoke allows affiliate marketers to earn commission from sales generated through the links they circulate.) The infrastructure between Taobaoke and these affiliate marketing platforms is in fact one and the same: the spread of cost-per-sale commission models to new types of content is a hallmark of Taobao’s evolution in the mobile world. Much like the API, which continues to facilitate network coordination on Taobao to this day, new uses of Taobaoke’s infrastructure also shows how core platform infrastructure continues to expand and be reused in new situations for the benefit of different users.

From Search to Recommendation

Certain sections of the platform showcase product content that is completely based on recommendation. These sections take data intelligence even further.

Flash-sale channels were once operated by Taobao employees, who curated the content in the channel through negotiations with merchants or by selecting from standard pools of product. Yet as Taobao’s mobile technology improved, the Taobao app began to focus on purely algorithmic channels. In those channels, the products shown to consumers are completely chosen by data intelligence, from specialized recommendation-based areas aimed at higher-spending consumers to individual items displayed at the very bottom of the Taobao app’s main screen. (This item recommendation area, dubbed “Guess What You Like,” is nothing to sneeze at. In an industry where area within an app is worth more than real estate, “Guess What You Like” is extensive, easily displaying more than a hundred precisely related products for users who are willing to continue scrolling.)

These recommendations are now ubiquitous throughout the Taobao app. When a consumer enters a store or views an individual item, the system will also recommend other goods from the same store. At the bottom of the user’s shopping cart, other items purchased by consumers who bought the same item will also appear as the user views their cart. After checking out, the user will see extra recommendations for products from other categories and sellers that might interest the consumer. When a user opens a recent order to check a delivery status, if he or she gets bored and scrolls down past the order information, there is ample data intelligence waiting with more recommendations. In many internet products, such as search, advertisements are intrusive and even detrimental to the user experience. However, because the Taobao app is already dedicated to shopping, product suggestions placed in relatively unobtrusive locations are not distracting to the consumer. (After all, the person is browsing, ultimately, to buy.) Consequently, the Taobao app works hard to turn every possible nook and cranny of the app into smart products, helping merchants create as many selling opportunities as possible. Through these merchant actions, consumers encounter, in increasingly smarter ways, the goods they are most interested in.

Data intelligence operates everywhere behind the scenes of the Taobao app, often in areas that the user might not expect. Consider, for example, banner ads, one of the marketing mainstays of e-commerce websites. In years past, advertising copy and visual advertisements were the province of a multitude of designers within the industry. Some designers are employed in-house by merchants, but many operate independent companies that provide specialized services for e-commerce operators. Year-round, these companies are busy with design work, but when large sales occur (and, especially, Singles Day), many of these designers find themselves overwhelmed with a spike of demand for work product. Enter Alibaba’s AI banner-design software, LuBan, named after ancient China’s most famous craftsman. This software uses existing product information and pictures, as well as those uploaded by merchants, to automatically generate effective and aesthetically pleasing banner ads. During Singles Day 2017, LuBan created four hundred million banner ads—that’s eight thousand per second. Coupled with the Taobao app’s personalized ad engines, LuBan ensures that just about every consumer sees different advertisements even for the same product.

From Retail to Content

Starting in 2014, Taobao began to move further into content, reaching out to writers, online content producers, and established media outlets and publishers. In addition to retail products in the narrow sense, Taobao fostered an environment where users could consume content related to retail. The offerings and business models in Taobao’s new, mobile incarnation came to take a very different form than did those of its PC predecessor, which tended to focus on commerce more narrowly defined.

The first type of content to appear on the Taobao app was sponsored articles collected in a prominently placed section of the app called Taobao Headlines. These feature-length articles are written by third-party affiliate marketers about certain industries or certain categories of product, for example, “Full-Length Winter Coats for Short Girls” or “In the Battle of the Toners, Who Wins, China or France?” The authors may not get high marks for their refined aesthetics, but Taobao Headlines aren’t meant to be literary masterpieces. The articles, a form of indirect advertising copy, make an effective marketing tool for sellers. (Authors get paid using Taobaoke infrastructure for commissions, as mentioned above.)

The Taobao Headlines section of the Taobao app builds data intelligence into the design of its channel the same way other specialized sections of the Taobao app do. The first application of data intelligence comes even before consumers enter the channel. As befits a headline, the channel gateway on the front screen of the Taobao app previews a rotating selection of articles. Which articles are displayed on the front screen is based on the user’s previous buying history and browsing behavior. Once the user enters the channel itself, the order in which content is displayed is also customized, much like search results. Yet because the Taobao Headlines channel is matching consumers to content rather than to product information, its algorithms naturally must consider more facets of the consumer to effectively pair user and article. More-sophisticated algorithms and data technology maximize matching efficiency and drive sales, generating commission income for a growing cohort of content producers. The result is that the Taobao Headlines channel is itself an intelligent product whose value constantly evolves in response to consumer behavior and feedback.

After the success of written content in 2014 and 2015, Taobao continued to move into video content in 2016, comprehensively investing in video technology across all parts of the app. In China, technology for online video began to boom around 2015, and live streaming soon took the platform by storm. From the mobile app, users can browse live streams featuring social media influencers, Chinese celebrities, industry experts, and representatives from brands and merchant storefronts. These live streams might take consumers on tours of factories and storehouses in China, might show viewers what buyers around the world are buying at discount outlets, might feature singing and dancing—the sky is the limit. What most attracts China’s netizens to these live-streaming sellers is the opportunity to ask direct questions of brands, catch deals and sales specially designed for live-stream viewers, and even tease and flirt with the streaming host. For merchants, live streams support brand marketing and increase consumer engagement.

Video now appears all across the Taobao app. Product detail screens—previously static text and images—now often commonly feature short videos produced by the brand. Especially in categories where the look and feel of a product are important, such as in apparel and cosmetics, video content greatly enhances the consumer experience. Buyers can even upload short videos (generally a few seconds long) when leaving reviews for products they have bought, for example, capturing how a pen writes on paper or how easily they assembled a coat rack. Although video on social media websites is not a new technology, allowing any user to upload created content at scale, when the platform hosts hundreds of millions of active buyers, is a significant technical challenge. (Not surprisingly, ISVs specializing in video production now do a thriving business for merchants on the platform. New point positions abound in a mobile world.)

Taobao has even partnered with well-known directors and production companies across China to run sales in the form of short original videos. In 2016, Taobao quietly introduced a “secret” area in the app called the Taobao Loft, which users can access by swiping downward on the home screen after ten o’clock every night. The app’s “second floor” originally featured an online video series, One Thousand and One Nights, created exclusively for Taobao. Each episode tells the story of a different person enjoying a different midnight snack at a magical restaurant and reliving the past. At the end of each fantastical video, viewers can buy the featured food from merchants on the platform.

The series was a runaway hit and became a classic case study in Chinese marketing circles. More importantly, it also sold a prodigious amount of food. The first episode tells the story of a young woman from Qingdao (in the north of China) working far from home in Shanghai. One rainy night, unable to catch a cab, she happens upon a magical restaurant that is larger inside than it appears from outside. The chef serves her a certain fish dumpling from her hometown and calls her by name, wishing her a happy birthday. The fish dumplings, made with mackerel native to Qingdao, are unheard-of outside the city. The flavors of home bring back memories of home cooking and childhood.

In the first two hours after the video went live, consumers bought two hundred thousand dumplings. Throughout the sixteen-episode run of the series, users again and again demonstrated unexpected appetite for the flavors of the stories. A merchant selling imported Iberian ham, normally an expensive and extremely niche product, sold out of the store’s entire stock by midnight after the product was featured in an episode. (When China’s Spanish consulate got wind of the story, an embassy official directly called Taobao’s marketing team to learn more about the series. Understandably, the phone call left the young, unprepared employees in a state of apoplexy.)

But more than just creating a huge volume of sales for a few merchants, the video series reminded users that Taobao offers more than just business or commercial products. It reminded users that Taobao began as a forum, a place where you could find anything under the sun, sold by anyone, anywhere around the world. Nights sought to bring users back to a Taobao they only vaguely remembered, a Taobao a decade removed, when the merchants were less “professional” but more genuine, where each product came with a genuine story, and where you might shoot the breeze online with a stranger in an afternoon lull at the office when neither person had anything better to do. It meant a return to a Taobao that felt small and warm, where new things were happening and where the scent of the future wafted through the virtual air.

Taobao continues to evolve. The only limits of the platform are the limits of technology and the limits of your imagination.