Chapter 3. Evolving E-Commerce Business Models

E-commerce facilitates trade across borders, increases convenience for consumers, and enables firms to reach new markets. Despite its short history, the e-commerce landscape has rapidly evolved through the development of new business models, which often integrate new and emerging digital technologies as well as new online payment mechanisms. This chapter analyses evolving e-commerce business models, focusing in particular on business-to-consumer (B2C) e-commerce, and includes examples from existing businesses to highlight firm-level innovations. It concludes by identifying key areas for policy action.


E-commerce has emerged as an economic and social phenomenon over the last 25 years, facilitating trade across borders, increasing convenience for consumers, and enabling firms to reach new markets. Over that short history, the e-commerce landscape has rapidly evolved through the development of new business models, which often integrate new and emerging digital technologies as well as new online payment mechanisms. This change has occurred in tandem with the wider social and economic implications of digital transformation.

Many e-commerce business models use online platforms, facilitating purchases between often unknown and dispersed buyers and sellers. Another emerging trend is the growth of subscription e-commerce business models, whereby users access goods and services in a continuous, recurring stream. Firms are also developing online-offline e-commerce business models that integrate digital ordering mechanisms alongside physical infrastructures, including within brick-and-mortar stores. This chapter discusses these business models as well as the digital technologies and payment mechanisms that enable them.

3.1. What is a business model?

A business model is a “term of art” without precise definition that has emerged from the business and economics literature (Ovans, 2015[1]). Broadly, the term describes the strategies and mechanisms firms use to succeed in competitive marketplaces. The Oslo Manual, a statistical manual for the measurement of innovation, defines a business model as “all core business processes … as well as the main products that a firm sells, currently or in the future, to achieve its strategic goals and objectives” (OECD/Eurostat, 2018[2]).

A variation on this definition includes “the rationale of how an organization creates, delivers and captures value” (Osterwalder and Pigneur, 2010[3]). Notably, this lens of value creation has been used across the OECD, including in analyses of taxation and environmental policy (OECD, 2018[4]; Hilton et al., 2019[5]).

In this chapter, “business models” refer to commonly observable business processes, purposes, strategies and means of generating revenue through e-commerce, although the precise formulation of these features differ. The chapter also highlights the role of new and emerging digital technologies and online payment mechanisms in these business models.

New business models push out the e-commerce frontier in two ways (see Chapter 2). First, new business models can enable more transactions to move online in a given market or for a given set of participants, an effect known as the intensive margin of e-commerce. Second, new business models can enable whole new markets to emerge for goods and services that previously could not have been bought or sold online, or allow new participants to enter the market. This effect is referred to as the extensive margin of e-commerce.

Digital technologies enable e-commerce innovations and often serve as the backbone of business model developments. Some of these technologies, like smart assistants enabled by artificial intelligence (AI), constitute new channels for selling or purchasing products over electronic networks. Other emerging technologies, like big data analysis, foster the growth of new data-driven business models for e-commerce and can support transactions moving online.

Similarly, online payment innovations can help to unlock e-commerce potential by promoting trusted online transactions between unknown parties. Such innovations make e-commerce more convenient and can facilitate trade online along the intensive margin. Similarly, innovations that enable new groups to participate in e-commerce can facilitate more trade online along the extensive margin. This chapter considers evolving e-commerce business models, focusing in particular on the business-to-consumer (B2C) space, and uses examples from existing businesses to highlight firm-level innovations.

3.2. Data: Enabling innovation in e-commerce

Data are both an input and an output of e-commerce transactions. Online transactions provide some actors with the ability to collect a range of detailed data, including:

  • Individual data (e.g. name, gender, age, location, place of residence),

  • Customer history (including, in some cases, the ability to track transactions over time and across multiple vendors),

  • In-store movements (in the case of some of the online-offline interactions) and other spatial information, and

  • Browsing history, Internet Protocol (IP) address, and data from connected devices (discussed further below).

As e-commerce is undertaken through digital means, such data can be more easily captured and utilised than they would have been for an offline transaction.

Data generated and gathered in the course of an e-commerce transaction are often particularly lucrative because they can help target product offerings, improve matching and create a better consumer experience. This data can also drive firm-level strategies to increase and extend user participation. As e-commerce expands, these flows are increasingly large and rich, yielding combinatorial insights and innovations (OECD, 2015[6]) and creating new sources of value (OECD, 2018[4]). The private use of data is sometimes controversial and is the focus of past and future work across the OECD (OECD, 2017[7]; OECD, forthcoming[8]).

Online platform e-commerce business models

An analysis of e-commerce would be incomplete without a discussion of online platforms, which act as both retailers and providers of digital marketplaces that bring buyers and sellers together online. Online platforms have become significant actors in the e-commerce landscape, facilitating a range of interactions between disparate actors. In the context of e-commerce, online platforms typically bring together buyers and sellers for the sale/purchase of either digital or physical products, a phenomenon known as a multi-sided market. Ongoing OECD work analyses online platforms and the resulting multi-sided markets that facilitate interactions between multiple groups or users, or “sides” (OECD, 2019[9]).

Increased scale and scope of goods and services

As multi-sided markets, online platforms benefit from both direct and indirect network effects, whereby economies of scale benefit users on both sides of the market. In the context of e-commerce, these sides can be understood as buyers and sellers. Typically, buyers gain utility from the presence of more sellers, assuming there is an expansion in the scope and/or variety of products for sale. Similarly, sellers are attracted to buyers to whom they can sell their products. As digital services, platforms are characterised by relatively higher fixed costs and comparatively lower marginal costs, meaning that the additional cost of hosting another buyer or seller can be close to zero.

In the context of e-commerce, online platforms act as intermediaries between buyers and sellers to facilitate the exchange of goods and services over the Internet. The large number of actors in a digital marketplace allows a potentially infinite variety of goods and services available for sale, in contrast to the more limited scope of products available in physical stores. For example, the average physical “super centre” store of the American retail chain Walmart typically holds approximately 120 000 items for sale while Walmart’s online store offers 35 million items for sale (Bloomberg, 2017[10]). As additional products can be added to online platforms at very low marginal cost, the scope and variety of previously unprofitable products increases, pushing out the extensive margin of e-commerce (Ellison and Ellison, 2018[11]).

The sheer scope of the goods and services that can be made available on online platforms can also enable the provision of so-called “long-tail” goods and services (Anderson, 2004[13]), namely very niche products with small markets where it might be otherwise unprofitable for offline sellers to operate. Further academic work identifies the consumer surplus resulting from increased scope and variety of product offerings from online platforms (Brynjolfsson, Hu and Smith, 2003[12]).

3.3. Transforming back-end operations using digital technologies

The sheer abundance of products available for purchase through some online platforms creates significant back-end logistical, distributional and supply-chain challenges for online retailers. Digital technologies can transform back-end operations to increase efficiency and enable efficient delivery, despite increasing volumes and rates of processing and production. Machine learning algorithms that improve with use can also optimise autonomous machines and robotics to improve the warehousing, fulfilment and logistics processes.

Online retailers have a range of applications for machine learning, including optimising back-end scheduling and optimal resource allocation, predicting product demand, classifying products and identifying keywords, and developing ideal inbound and outbound logistics. Machine learning can also help in the parcel sorting process. For example, the Chinese retailer has incorporated robotics into logistics centres and warehouses to improve parcel sorting, packaging and distribution. The company unveiled a fully automated sorting centre in 2018, which features the completely automated processing of 16 000 packages per hour with an accuracy rate of almost 100% (, 2018[14]).

Connected devices are also extensively used throughout the inventory, logistics and fulfilment systems of e-commerce firms. For example, radio frequency identification (RFID) is a technology that enables the transmission of information over radio waves to Internet-connected devices, while a variety of sensors can be embedded in products to help to track items for transit. These can help optimise the supply chain process, predict shipping times and provide real-time monitoring of levels of stock. Nonetheless, RFID technologies have yet to be widely adopted by enterprises, suggesting unrealised potential (OECD, 2017[18]).

In other cases, online platforms have created new markets for goods and services that were previously not traded online, again increasing the extensive margin of e-commerce. For example, new business models like the American company Uber and the Chinese company Didi Chuxing enable online ride-sharing services, a market that did not exist prior to 2011. Similarly, home-sharing platforms like Airbnb enable access to accommodation in a novel sharing arrangement that would otherwise not take place, particularly on a global scale, without the existence of an online platform.

Matching buyers and sellers

As online platforms scale, they must provide some form of matching mechanism to enable buyers and sellers to find each other in a way that is satisfactory to both parties. Reducing friction, as well as transaction and search costs for both buyers and sellers, is essential to improving utility on both sides, and thereby increasing the chance of a successful online sale/purchase. Notably, these matches can take place between both local and geographically dispersed actors, and for the sale of both physical and digital goods and services.

Where a firm’s business model is based on user engagement (revenue from advertising) or on successful sales (a commission for intermediating services), the firm has incentives to develop mechanisms to ensure successful matches. Unsuccessful matches also cause actors on both sides to leave the marketplace, thinning out the market (Fradkin, 2017[19]; Horton, 2014[20]).

The type of matching mechanism differs across online platforms, and typically relates to the nature of the good or service on offer and the ability of the buyer to form or express a preference for a good or service. For homogenous, non-differentiable services where buyers have few preferences (or difficulties in assessing and expressing preferences), some online platforms have had success with centralised matching.

Uber, for example, algorithmically offers rides to drivers based on location and the buyer’s preferences (e.g. a private limousine or a ride shared with others). As noted by Einav et al. (2016[24]), this choice rests on the on-demand nature of Uber’s services which prioritises safe transportation with few transaction costs, low prices and short wait times. Customers indicate their preferences and their destination, and an algorithm matches the customer with a driver. While customers are not given a choice about specific drivers or car model, they have the option to reject their match. Other inputs to the matching algorithm include passenger and driver location, as well as supply (number of drivers) and demand (number of passengers), which influences the price.

In contrast, search and filter mechanisms may be more effective in decentralised marketplaces with more heterogeneous goods and services for sale for which customers may have a range of preferences (Fradkin, 2017[19]). For example, those looking for an apartment to rent on Airbnb will typically have different preferences for location, room characteristics and price in a given area.

3.4. Using big data analytics and AI to improve matching

To match buyers and sellers, or indeed consumers and content, e-commerce firms can make use of data gleaned from their customers to algorithmically optimise and personalise matching and product recommendations. These data could include browsing patterns, the length and nature of user engagement with particular features, responsiveness to design or format changes, or the behaviour of other similar users.

For example, Amazon uses AI, in the form of neural networks (Amazon, 2016[21]), to train algorithms to generate recommendations based on user purchase history, product recommendations and ratings history. Personalised recommendations play a large part in steering consumer behaviour online – for example, some have estimated that 80% of Netflix activity is driven by algorithmic recommendations (Financial Times, 2016[23]).

Academic work has found that changes in algorithmic design can alter the rate of matching between buyers and sellers in the context of online platforms (Fradkin, 2017[19]), thereby improving overall engagement and the likelihood of matches. This is particularly important where online platforms earn revenue through successful transactions, as failed matches can cause actors to exit the marketplace.

Consumers typically express an initial preference by outlining core characteristics of the product or service they wish to buy (e.g. a product name or a location). The online platform’s search algorithm then offers a set of results for consideration for the consumer. This set of results may be derived from the online platform’s indexed knowledge of product and seller characteristics (see Box 3.2), and could be personalised based on the online platform’s existing knowledge of the buyer. The consumer will then typically use features in the search interface, including filter mechanisms, to refine the search results to determine the product or service that best matches their preferences. A range of academic work finds that the ranking of search results influences a consumer’s choice of what to purchase (Ursu, 2016[25]).

In general, when the online platform only benefits from user engagement (advertising revenue) or from a successful transaction (commission or selling fee), matching mechanisms are essential for the platform’s business model. Matching mechanisms are also often associated with pricing arrangements (see Box 3.5).

3.5. Pricing on online e-commerce platforms

Pricing strategies on online platforms differ in accordance with the nature of the good or service on offer and buyer and seller characteristics. Some pricing arrangements are closely associated with the matching process on online platforms.

For example, Uber uses a pricing mechanism that operates in tandem with the process of centralised matching. Uber acts as the price-setting agent, dynamically changing prices in response to market conditions to facilitate matches in real time. “Surge pricing” of this nature can help clear markets faster and minimise transaction costs for buyers and sellers, thereby enabling e-commerce, although some consumers have complained about such practices (Riley, 2017[26]). Dynamic pricing is evident in other markets, including the airline industry.

As noted above, e-commerce enables the online sale of goods and services in niches that were previously unprofitable to offer for sale offline. However, as these goods and services typically also have comparatively smaller markets, pricing such goods is not necessarily obvious. Auction mechanisms are useful as a means of price discovery in these cases, particularly for unique or used products (Einav et al., 2017[27]), although transaction costs for buyers typically increase. Auctions are also often used for heterogeneous services, such as task-mediating services like Upwork and Thumbtack, where the specificity of the task results in a lack of an alternative price-setting mechanism with lower frictions (Fradkin, 2017[19]).

In the early days of e-commerce, auction mechanisms were prevalent on online platforms. Today, they have declined in use in the retail market, particularly for standard goods with large markets. On eBay, an e-commerce platform and online auction house, the share of auctions in the active listings declined from over 90% in January 2003 to just approximately 15% in January 2013 (Einav et al., 2017[27]). A notable exception is the use of advertising placement auctions for digital marketing purposes, as used by Google and Yahoo, and the use of bidding and auction systems to determine the ranking of products on many price comparison platforms.

The simplest and most common prices found on online platforms are those determined and posted by the seller. Seller-determined prices remove the physical and cognitive transaction costs associated with bidding in an auction. Seller-determined prices enable sellers to change easily their prices in accordance with their preferences (e.g. Airbnb hosts can easily change prices in accordance with their own schedule or requirements). The decline in the relative share of auctions on retail e-commerce platforms, and the growth in seller-determined prices, suggest that buyers may also prefer this format in many cases.

Online intermediaries are in a unique position to collect data about users on both sides of the market, including but not limited to transaction data, demographic data, and data on user behaviour, among others. This data may be offered, observed or inferred. When used in combination and potentially matched with other data, the price-setting agent may be in a better position to determine the consumer’s reserve price – namely, the highest possible price that each individual may be willing to pay.

This practice is known as personalised pricing, whereby some consumers may pay less for a given product and others may pay more than they would have done if all consumers were offered the same price (OFT, 2013[28]). While there is little empirical evidence of the use and extent of personalised pricing, by virtue of being personalised, such practices may be difficult to detect and monitor. Nonetheless, recent OECD work suggests that firms have the technical potential to personalise prices, and consumers report having experienced the practice, including on online platforms (OECD, 2018[29]). Personalised pricing practices have uncertain impacts of overall competitive conditions and consumer welfare.

The OECD Recommendation on Consumer Protection In E-Commerce (OECD, 2016[30]) notes that businesses should “provide information about the … costs associated with a transaction that is sufficient to make an informed decision.” When firms personalise prices, they should therefore disclose to consumers that they are doing so. However, such disclosures may have uncertain impacts on consumer behaviour (OECD, 2018[31]). Future experimental work from the OECD will aim to identify which approaches to disclosure are effective in enabling consumers to identify and comprehend personalised pricing, and the extent to which disclosure has a material impact on consumer decision making.

Approaches to increase trust

E-commerce platforms bring together buyers and sellers who may be dispersed geographically and involve parties that have not met before. Some sellers on online platforms are large and may have established brands that buyers may trust. In contrast, smaller, potentially unknown vendors may have more difficulty establishing conditions that would lead buyers to transact willingly with them. Third-party providers and sellers operating on multi-sided markets may be further unsure about payment or reliability from buyers who purchase their products.

Online platforms can provide mechanisms that help to resolve this essential information asymmetry, build trust on both sides of the market, and ensure that transactions are safe and reliable to foster e-commerce. Online platforms have the ability to easily collect, store and communicate information to both sides of the market in a verifiable way, particularly when there have been repeat transactions. This enables the entirety of the engagement with the platform, and not simply transaction history with a single seller on the platform, to build trust for online transactions.

In general, analysis on peer platform markets finds that consumers generally trust platforms, sometimes to a greater extent than conventional businesses in the same market (OECD, 2017[32]). Notably, this analysis finds that consumer trust in an online platform is generally higher than trust in unknown third-party sellers that may operate on the platform, and that consumer trust in the platform is the most commonly cited reason for proceeding with transactions even when the buyer may not trust the seller. This indicates that trust mechanisms developed by online platforms can enable more transactions, pushing out the extensive margin of e-commerce.

For many online platforms, a significant ex ante mechanism for building trust includes the performance of verification checks of the identity and/or the credentials of either buyers or sellers. For example, Uber, Lyft and other ride-sharing platforms confirm the identities of drivers through government-issued documentation like a valid driver’s license, vehicle registration, reviews of criminal records, and checks of the suitability of the vehicle. Airbnb verifies the identity of both guests and hosts on the platform through passports, while other online platforms confirm professional certifications and licenses. The verification of identities online can enable e-commerce (see Box 3.6).

3.6. Digital identity and its potential for e-commerce

Digital identity refers to the set of information used by a computer to authenticate an identity. This identity could be anonymous or pseudonymous, and relate to a civil identity, a corporate identity, or be linked to information provided by a user (age, gender, etc.), or established through browsing, selling or purchasing history. As such, it is possible for individuals or organisations to have multiple digital identities across various platforms and websites that may or may not be connected.

In response to this fractured digital identity landscape, both public and private schemes that seek to create unified digital identities are becoming increasingly prevalent. Government digital identity efforts usually consist of a unique number that links a digital profile to civil information like demographic or biometric data. For example, India’s Aadhaar programme issues a unique number to every Indian citizen which is a valid means of identification vis-à-vis the government as well as private Internet sites including Airbnb, Uber and digital wallet services (Nadadhur, 2018[33]).

Private firms also offer digital identity authentication services. Similar to government schemes, private firms provide users with online identities through mechanisms like unique numbers or accounts that verify a user’s identity across a variety of Internet sites. In contrast to centralised government efforts, this identity may be separate from a user’s civil identity or may be linked to only certain aspects. Facebook is one such private actor that increasingly acts as a digital identity provider. For example, many sites give users the option to use Facebook account information as identity verification instead of creating a new username and password.

Digital identities present a number of opportunities for buyers and sellers alike. For one, when sellers have access to information about buyers’ diverse tastes through their digital identity, they can make better use of algorithms that tailor their advertising and content specifically to the buyer’s needs. Such tailoring decreases bounce rates (the proportion of users that leave a website without making a click) and reduces cart abandonment (the phenomenon of placing an item in a digital checkout cart but not carrying out the sale), while increasing overall conversion rates (the proportion of users that make a purchase) (Deloitte, 2017[34]). Overall, digital identity schemes that share information with e-commerce firms can make sellers more efficient and profitable.

For consumers, these digital identity mechanisms can also represent an advantage in terms of ease-of-use and security. When users have a digital identity across many sites, they do not need to manually re-enter personal data such as shipping details, and are less likely to make mistakes when filling out online forms. With a unified digital identity, users also no longer need rely on an ever-growing list of passwords or the individual security systems of many different and potential unsecure sites to protect sensitive data (Deloitte, 2017[34]; Eggleton, 2016[35]).

Digital identities also pose certain drawbacks, particularly for buyers. Privacy issues and lack of control of personal data may be exacerbated by digital identity schemes, which could share personal information with e-commerce sites. For example, those taking part in the Aadhaar digital identity scheme have been reticent to link this data to certain functions like their digital wallets (see below) due to privacy concerns (Nadadhur, 2018[33]). To a large extent, the real impact of digital identity on e-commerce will depend on the nature of the digital identity systems adopted as well as how much user information is deemed appropriate to share with e-commerce firms. Policy makers therefore play a pivotal role in the development of digital identity in e-commerce and the level of control that buyers have over how their personal data is shared.

Security risks are another potential drawback of digital identities. Although digital identity schemes may offer buyers the opportunity to consolidate password data, these mechanisms themselves are not impervious to digital security threats. Particularly in the case that data associated with a digital identity are not anonymised, digital identities may be easily linked to sensitive administrative information like social security numbers or national identification data, and then exploited by third parties. Even when data are anonymised, digital identity schemes could in fact leave users more vulnerable. If a data breach of a centralised digital identity occurs, third parties could access users’ most sensitive data all in one place (Domingo and Enriquez, 2018[36]).

Beyond these potential advantages and disadvantages, interoperability issues create further challenges for the relevance and efficiency of digital identity initiatives. As both centralised government initiatives and private digital identity firms multiply, the ability of these systems to interact with one another influences the role that digital identity can play in e-commerce. Namely, the relevance of the benefits (and risks) of digital identity in the e-commerce sector may be muted if interoperability issues are not resolved.

Another significant ex-ante mechanism for building trust includes the imposition of minimum quality standards. Drivers wishing to participate as sellers with Uber in its American marketplace must pass a 19 point vehicle inspection test with a registered mechanic before their account can be activated (Uber Partner Help, 2018[37]). Third-party sellers on online platforms may be required to meet a minimum standard of service as a condition of using the platform; for example, some third-party sellers on Amazon must meet minimum requirements for returns and delivery times (Levy, 2016[38]; Levy, 2017[39]).

The most common ex post trust-building mechanism on online platforms is the use of reputation and review systems, whereby one or both sides of the transaction is able to rate or review the transaction through a mechanism that can be viewed by other users in the marketplace. Both reputation and review systems are a form of public good. Recent OECD work on trust in peer platform markets finds that reviews and ratings are important for consumer engagement with online platforms (OECD, 2017[32]).

Reputation systems track the performance of a user over time and help other users determine whether they wish to transact with them. They are an important mechanism of reducing moral hazard and information asymmetry between different sides of the market. Reputation systems were initially pioneered by eBay but have subsequently spread to most online platforms (Luca, 2016[40]), indicating their utility as a trust-building tool in e-commerce.

Research on online platforms like eBay and Amazon finds that the reputation of a seller impacts demand from buyers (Tadelis, 2016[41]). In the case of Airbnb, analysis suggests that incentivising reviews by offering a USD 25 coupon could improve the usefulness of the review (Fradkin et al., 2017[43]), while other research on Alibaba’s Taobao platform found that rewarding feedback with a coupon could be used as a signal of the quality of the seller (Li, Tadelis and Zhou, 2016[44]). However, there is a risk that reviews that are compensated could be biased (Cabral and Li, 2014[45]). Sellers may also have perverse incentives to manipulate their reviews (Mayzlin, Dover and Chevalier, 2014[42]).

Product reviews are another form of feedback often used by online platforms. They enable buyers to review the quality and utility of the product (as opposed to the seller). These ratings are a means of ensuring that product quality on online platforms remains high, as low ratings impact sales (Chevalier and Mayzlin, 2006[46]), particularly when the ratings have been judged as useful by others (Chen, Dhanasobhon and Smith, 2008[47]). Private surveys also find that consumers trust online reviews to a similar degree as personal recommendations (Bright Local, 2017[48]).

Product reviews have significant impacts on offline behaviour as well. Pew Research finds that 45% of consumers check reviews before purchasing in a physical store (Pew Research Centre, 2016[49]), while Google finds that mobile searches for product reviews increased by 35% from 2015 to 2017 (Google, 2018[50]). This underscores the degree to which digital technologies are embedded in offline purchasing (see below).

Consequently, online platforms have experimented with mechanisms to ensure the review system’s quality, including using machine learning to display reviews that are more useful to other consumers (Rubin, 2015[51]). Some online platforms have also taken action against third-party sellers over attempts to manipulate ratings with fake or misleading reviews (Perez, 2016[52]; Broida, 2019[53]; Statt, 2019[54]).

Some online platforms have used other features to improve trust between unknown parties. Airbnb and Uber offer insurance to sellers for property damage that occurs during the course of a booking or transaction. In addition, some online platforms act as an mediating party should a dispute occur between the buyer and seller in a transaction, while others have customer service and return options for dissatisfied buyers. Some online peer-to-peer lending platforms, including Lending Club and Prosper, use algorithms to detect and protect against fraud (Xu, Lu and Chau, 2015[55]).

3.7. Disintermediating e-commerce: The potential of blockchain to enable trusted transactions

Blockchain is a technology that enables applications to authenticate ownership and carry out secure transactions for a variety of asset types (OECD, 2019[56]). For so-called public or unpermissioned blockchains, actors are able to verify transactions without the need of an existing trusted relationship through both cryptographic and market-designed incentives. Blockchain technologies are emerging and have yet to be applied in many domains or integrated into policy frameworks. However, as a radical means of facilitating disintermediated transactions, Catalini and Gans (2017[58]) note that blockchain could significantly bring down the costs of verification and networking associated with economic transactions, including e-commerce.

In the context of e-commerce, blockchain removes the need to use an intermediary for third party verification for trusted transactions. This could facilitate the development of distributed, peer-to-peer networks with multiple sides without the need for a centralised online marketplace. One emerging peer-to-peer marketplace of this nature is OpenBazaar, which features no fees for listings, selling or commissions, and accepts over 50 cryptocurrencies as payment (OECD, 2017[18]). Marketplaces like this could theoretically challenge the business model of firms that act as intermediaries between different sides of a market.

Other potential applications of blockchain related to trust could involve the development of a portable and decentralised reputation system. As noted above, reputation and review mechanisms have emerged as important enablers of e-commerce by increasing trust through the publication of previous transaction history. However, these rankings are not transferable from marketplace to marketplace, which could theoretically increase switching costs for users. Chlu, a new start-up, aims to provide a transferable record of a vendor’s reputation, using the immutable and public nature of blockchain, while also keeping them private using advanced cryptographic solutions (Chlu, 2018[59]). Such a system could also be used to aggregate rankings and performance from multiple online platforms or communities.

Mechanisms to facilitate firm participation

While the specific design of online platforms differ, each platform has incentives to add users, meaning that the entry costs for sellers that wish to participate on online platforms are typically low. This gives small and medium-sized enterprises (SMEs) and, in some cases sole traders, the ability to compete alongside more established firms on online platforms. When online platforms operate in multiple international markets, being active on the platform can give sellers access to new markets overseas.

However, there are a range of complementary investments that sellers may need to make to effectively buy and sell online. Trading at a distance, including potentially across borders, requires significant upstream and downstream investments in supply chain management, secure payment systems, delivery and fulfilment mechanisms, and customer-facing services like dispute resolution mechanisms and customer service. For e-commerce across borders, these activities may need to be conducted in foreign languages. Notably, these are all factors mentioned by firms as influencing their decision not to participate in e-commerce (see Chapter 2).

As a result, online platforms have begun to offer complementary services for firms that trade on their platform, including fulfilment, logistics, customer service and software-as-a-service (SaaS) offerings. SMEs disproportionately benefit from these services because they would otherwise require significant upfront fixed costs, while platform-enabled services can transform this fixed cost into a variable cost. These new solutions push out the extensive margin of e-commerce, enabling new participants to enter the marketplace.

Many major online platforms such as Amazon, Alibaba and Rakuten, offer fulfilment services to sellers that operate on the marketplace. These services are often integrated with the broader sales management interface of the platform and compete with other international fulfilment service providers, including Shipwire, Whiplash and Cloud Fulfilment. Such fulfilment services may be particularly useful in countries without a well-developed postal system or with a limited number of third party logistics service providers (see Chapter 2).

In the case of Amazon, there are over 2 million active sellers worldwide on Amazon Marketplace, and they are responsible for over half of the sales made on Amazon’s platform (Amazon Investor Relations, 2015[60]). Amazon’s fulfilment service, known as “Fulfillment by Amazon”, enables these sellers to use Amazon’s supply chain for downstream services including storage, packaging, shipping and foreign language customer services. Membership in the Fulfillment by Amazon programme enables third-party sellers to gain access to Amazon’s lucrative Prime subscriber base (discussed below).

In 2016, Amazon shipped more than 2 billion items for third-party vendors through Fulfilment by Amazon (Amazon, 2018[61]). In 2017, Amazon reported a 70% year-on-year increase in vendors using Fulfillment by Amazon. Revenues from services provided to third-party sellers, which includes commissions and any related fulfilment and shipping fees, and other third-party seller services, had grown to almost USD 43 billion by 2018, more than triple its value in 2014 (Amazon, 2018[61]).

Other online platforms include services that are specifically aimed at cross-border e-commerce, thereby enabling SMEs to access international markets. This appears to be a notable feature of online platforms that operate multiple, country-specific marketplaces. Thus, Amazon operates over 11 marketplaces in different legal jurisdictions and regions (Amazon, 2018[62]). Amazon customers from over 180 countries are then able to purchase through these marketplaces. Amazon sellers must meet certain listing and registration requirements prior to operating on an international marketplace (Amazon, 2018[62]). Amazon also provides legal and regulatory information specific to each jurisdiction for sellers, and directs international sellers to third-party solutions providers for advertising, customs brokerage and tax assistance.

Another large firm using an online platform is eBay, which operates almost 30 international websites, but offers no formal restrictions on listing in overseas marketplaces aside from the buyer’s indication of willingness to ship overseas. For US and UK-based sellers, eBay has developed an export logistics service for international sales, shipping and customs clearance, called the eBay Global Shipping Programme. Sellers send goods purchased on the platform to a domestic intermediary, who then handles the customs, taxes, tracking, insurance and shipping associated with the export process to over 80 countries (eBay, 2018[63]). This service also enables sellers to avoid any negative feedback (see the sub-section above on reputation mechanisms) or any damage that results from international shipping.

An analysis of the impacts of this programme finds that cross-border trade increased exports at the extensive margin by targeting SMEs, for whom export costs were otherwise prohibitive (Hiu, 2016[64]). The analysis also finds that the programme increased trade in products with smaller price-to-shipping ratios, while foreign consumers benefited from an increase in marketplace quality and product variety.

Other online platforms have pursued similar cross-border strategies for SMEs in other parts of the world. For example, the Chinese online platform Alibaba recently announced plans to open an electronic world trade platform in Malaysia. This online resource aims to provide Malaysian SMEs support with tariffs, customs and logistics associated with international trade. The electronic resources are complemented by a regional distribution centre established by Alibaba’s partners, and additional co-operation with respect to payment mechanisms (Alibaba, 2017[65]).

More generally, an analysis by eBay found that among the micro, small and medium-sized enterprises (MSMEs) from 18 advanced and developing economies operating on its platform, about 90% to 100% sold internationally, and were also likely to serve multiple markets (eBay, 2016[66]). Recent analysis by the OECD and the World Trade Organization (WTO) also identified online platforms as useful mechanisms for enabling international trade for SMEs in developing countries (OECD/WTO, 2017[67]).

Subscription e-commerce business models

Another business model that is of increasing interest in e-commerce is the subscription model. This business model is characterised by regular and recurring payments for the repeated provision of a good or a service. In the e-commerce context, this encompasses a range of new and emerging businesses, from streaming services like Netflix to recurring purchases of consumer products like Dollar Shave Club. The subscription business model can relate to recurring purchases of digital goods and services, or a combination of both digital and tangible products (such as a newspaper subscription that includes access to digital content).

Subscription e-commerce business models typify a broader trend towards more continuous, digitally-enabled access to or provision of goods and services. Digital technologies enable easy ordering of goods and services, removing transaction costs associated with the purchasing process, thus improving convenience for consumers. Firms benefit from regular and ongoing revenue streams. The OECD E-Commerce Guidelines recognise the increasing importance of subscription models to e-commerce, and note that firms should disclose to consumers: “terms, conditions, and methods of payment, including … recurring charges, such as automatic repeat purchases and subscription renewals, and ways to opt out from such arrangements” (2016[30]).

Access to digital products by subscription

One aspect of the emerging trend of e-commerce subscription business models is subscription access to digital products. This encompasses the ordering of intangible products that can be delivered online, including software, media and services like cloud computing. These services increasingly form part of larger supply chains for some firms, or are bundled or complementary to other purchases.

Subscription models have become a feature of successful media companies, and provide a means of diversifying revenue streams away from advertising. As digital media is easily reproducible and shared, subscription services are predicated on access to a wide range of media content rather than individual purchases of specific content.

Many emerging subscription services offer access to digital products that are only tradeable as a result of digital transformation, like software services. As noted in economic literature, the pricing of non-rivalrous digital goods with low or zero marginal costs is not always obvious for firms. One solution is bundling many digital products and charging a single price (Bakos and Brynjolfsson, 1999[68]; Bakos and Brynjolfsson, 2000[69]). Some current e-commerce subscription business models, such as Spotify or Netflix, are practical examples of this theoretical insight (Goldfarb and Tucker, 2017[70]).

3.8. Using blockchain for digital rights management

Blockchain enables new kinds of decentralised and radically transparent markets and business models. For example, business models that aim to monetise content creation, like traditional music labels or film studios, have been negatively impacted by easy duplication and sharing of digital products. Some firms have created new digital media business models that earn revenues through subscription access to bundles of digital products, who funnel a share of revenues to content creators.

As the use of assets and transactions can be automatically captured in a detailed and transparent way on a blockchain, it becomes increasingly easy to price and deliver access to information and digital goods with a high degree of precision. This in turn may provide a new business model for the purchase and sale of digital content. For example, the Open Music Initiative (2018[72]) is an example of a start-up that aims to use blockchain technologies to promote and enforce digital rights management to enable compensation for artists.

Some digitalised subscription models pursue a so-called “freemium strategy” whereby usage of or access to content is provided with some limitations, or only when users accept exposure to advertising. Those who pay the relevant subscription fee may enjoy a more extensive level of service, which may include additional content or the absence of advertising. This model can help new and small firms gain market share by enabling the consumer to experience the service without initial up-front costs. As those users who pay for premium services are also likely to use the service more, firms are able to scale up responsively (European Commission, 2015[73]).

An example of the freemium model is used by The New York Times, which aims to convert revenues from subscriptions of printed newspapers into digital subscription services. In this model, there is a a paywall that emerges after a user accesses a certain amount of free content (Kumar et al., 2011[74]). The New York Times’ digital subscription revenues exceeded USD 1 billion in 2017 (Ember, 2018[75]).

In 2018, the United States Recording Industry Association estimated that music streaming accounted for 75% of American music industry revenues (RIAA, 2018[76]). At the end of 2018, the music streaming firm Spotify, one of the vanguards of the streaming business model, had 207 monthly active users, of which 116 million were freemium (ad-supported users). Approximately 96 million users had subscribed to a full subscription package (Spotify, 2018[78]). Consistent growth in premium subscriptions enabled Spotify to make a profit at the end of 2018 for the first time in its history. However, the company’s business model relies on paying royalties to content holders, which amounted to approximately 70% of its revenues in 2018 (Financial Times, 2018[79]).

Software and information technology applications are other intangible products that are increasingly ordered online as an ongoing service through a subscription business model. Typically, the model includes the licensed provision of software and other information technology applications over the Internet. Ubiquitous computing and the extensive use of cloud hosting underpin this business model (see Box 3.9).

Subscription e-commerce business models for software and information technology applications increasingly replace legacy pricing arrangements, which typically involved a single, perpetual license for a specific version of a product. One example is Adobe, a software provider that converted its full service offering into the cloud-based Adobe Creative Cloud suite (Sanitago, Gnanasambandam and Bhavik, 2018[81]). This enables dynamic service provision with ongoing updating and re-versioning of features and functionality, rather than a two- or three-year production cycle. Ongoing collection of data on customer use and preferences enables more tailoring to specific user needs.

Such business models are also more scalable, and in some cases can be offered at a price point competitive enough to attract new customers. This could include SMEs which may be able to convert the cost of investing in ICTs into an ongoing operating cost. Downward trends in information and communication technology (ICT) investment in OECD countries over time suggest that firms may indeed be using cloud computing to substitute for other kinds of investment (OECD, 2019[56]).

3.9. Cloud computing underpins digital service models

Cloud computing refers to a service model that provides clients with flexible on-demand access to a range of computing resources (OECD, 2014[80]; OECD, 2019[56]). As a set of combined technologies, cloud computing provides individuals and organisations with the ability to access resources, including software applications, storage capacity, networking and computing power) through an online interface. Some well-known variants of the cloud computing service model (OECD, 2018[4]) include:

  • Infrastructure-as-a-service (IaaS): IaaS refers to the delivery of infrastructure such as computing capacity. Also known as hardware-as-a-service, IaaS encompasses all of the physical computing resources that support delivery of applications as a service, such as computing services, database storage and networking capabilities. IaaS provides major cost savings to customers, as it provides access to additional computing capacity on demand, without the need for major capital investments in additional hardware.

  • Platform-as-a-service (PaaS): PaaS is a method by which an entire computing platform can be utilised remotely over the Internet via cloud computing. PaaS refers to a broad collection of application infrastructure, including operating systems, application platforms and database services. PaaS provides a way for customers to outsource their platform infrastructure needs and therefore avoid the need to purchase and implement a new platform. This service model typically allows cloud computing companies to charge customers only for the share of the resources they use, which is especially useful for a business that requires a specific application it would only use on occasion.

  • Software-as-a-service (SaaS): SaaS is a software model that incorporates the delivery and management of a software application to a remote client via the Internet. SaaS relies on the centralised hosting of a software application that is typically accessed via a web browser application. SaaS can be configured to allow public access or private access, where only users with the proper credentials are granted access to a particular hosted software application.

These computing resources can be priced on-demand and used in a flexible, scalable and adaptable manner, enabling users to reduce the costs of fixed investment in ICTs. Cloud computing therefore increases the availability, capacity and ubiquity of computing resources in a way that enables other digital technologies, including AI and autonomous machines like drones (OECD, 2015[6]; OECD, 2017[18]). It also allows users, like SMEs and individuals, to access computing resources that would otherwise be prohibitively expensive.

An interesting extension of this trend is the growing competitive marketplace of PaaS models for businesses wishing to sell online – namely, a form of e-commerce that has emerged from the growth of e-commerce itself. Shopify, BigCommerce, Lemonstand and Magento are all PaaS solutions for businesses that wish to build online stores. While they range in features, these packages typically offer website hosting, thematic design options, payment gateway services, analytics, as well as inventory and fulfilment integration, and they tend to integrate with multiple online sales channels, including social media. Shopify, one of the earliest and best-known e-commerce PaaS vendors, hosts more than 600 000 active online stores and has facilitated USD 55 billion of gross merchandise volume (Shopify, 2018[82]). In 2015, Amazon closed its own native PaaS offering (Amazon Webstore) and directed its merchants to Shopify, which remains the case today.

Subscription access to tangible and bundled goods and services

A recent e-commerce trend has been the growth of subscription business models for tangible goods, including in categories like beauty supplies (Birchbox), minerals (Celestial Minerals), groceries (Blue Apron, Hello Fresh), snack foods (Nature Box), cosmetics and self-care products (Dollar Shave Club, Harry’s), and many more.

This sector of e-commerce has grown in value over the last five years, and the largest of these retailers generated more than USD 2.6 billion in sales in 2016 (Chen et al., 2017[83]). While this space is volatile, new entrants have drawn significant interest from venture capitalists in recent years and there have been some notable success stories. For example, Dollar Shave Club, a subscription service for shaving materials founded in 2011, gained 5% of the US market share in razor cartridges within five years when it was acquired by Unilever for USD 1 billion (Cao and Mittelman, 2016[84]).

3.10. Connected devices can facilitate recurring purchases of goods that require replenishment

Many of the subscription business models for tangible goods and services relate to goods and services that deplete with use and require replenishment (Chen et al., 2017[83]). An interesting development in this space has been the use of connected devices that utilise streams of data through sensors, software and network connections associated with physical goods to make continuous or recurring purchases of tangible goods. For example, over the last five years, smart home speakers have emerged as a popular consumer good. These smart home speakers use a combination of AI and cloud computing to respond to natural language requests.

Many of the leading retailers of these devices are associated with popular online e-commerce brands, including Amazon, Google, and Alibaba. These devices are able to make digital and physical purchases in response to voice commands (Amazon, 2018[85]; Bateman, 2016[86]; Cadell, 2017[87]). Recent reports have noted that users of these devices appear to have increased their purchases of standardised products that require replenishment, including paper towels and diapers (Kim, 2017[88]; Pandolph, 2017[89]). Some firms that sell smart home speakers have encouraged this trend by offering voice-exclusive deals for specific products (Warren, 2017[90]).

An extension of this trend is the use of connected devices that automatically order goods that require replenishment. In some cases, “smart” appliances can automatically detect when they run low on essential supplies and automatically purchase additional supplies. For example, the Amazon Dash programme (Amazon, 2016[91]) enables the automatic replenishment of supplies for connected devices like dishwashers, washing machines, printers and water filters.

Other forms of e-commerce subscription services rely on a mix of bundled digital services and tangible goods and services. This has been a common form of diversification for conventional media companies – for instance, in return for a fee, many newspapers offer regular delivery of physical newspapers alongside open digital access to their website. Notable media companies who pursue this model include The Economist (The Economist, 2019[92]), Le Monde (Le Monde, 2019[93]) and the Financial Times (Financial Times, 2019[94]).

An example of this model is Amazon’s subscription service know as Amazon Prime, which enables access to a range of tangible services (free shipping, food delivery) alongside access to digital services (free streaming of video and music content). Amazon Prime offers fast delivery (usually under two days) for all purchases on eligible products without additional cost. Amazon Prime subscribers in some major cities are able to receive shipping within two hours without additional cost. Depending on the country, Amazon Prime membership also offers access to additional services including free streaming of video and music content, free access to popular e-book titles, restaurant delivery and data storage. There were over 100 million paid Prime members across the world in 2017 (Amazon, 2018[95]), who had access to unlimited free shipping on over 100 million items (Amazon, 2018[95]). Amazon gains an increasing amount of revenue from retail subscription services, which have more than quintupled in revenue since 2014 to almost USD 14.2 billion in 2018 (Amazon, 2018[61]). Retail subscription services captures not only revenues from Amazon Prime but also other Amazon subscription services.

3.11. The last mile: Using technologies to make the delivery of tangible products more efficient

As more transactions go online, fast and efficient delivery of tangible products becomes more important. Many of the biggest e-commerce providers have succeeded in improving their delivery time by optimising the logistics and supply chain fulfilment process. However, consumer demand for free and fast shipping services and the high rate of returns for some items have resulted in significant costs. Amazon, for example, posted net shipping expenses of USD 27.7 billion in 2018 alone and has driven further innovations in the delivery process (Amazon, 2018[61]).

In particular, firms have attempted to innovate in the provision of delivery services for the final leg of delivery to the requested location, or the so-called “last mile”. The costs of the final last mile of delivery can comprise up to half of the product’s total transportation costs and is subject to considerable expenses in terms of fuel, vehicle, labour and time costs (McKinsey, 2016[97]).

Some e-commerce firms have experimented with the use of autonomous delivery devices, including the use of unmanned aerial vehicles (commonly known as drones) or autonomous ground vehicles and robots. Autonomous devices are connected devices that make use of developments in cloud computing, machine learning and data to make sense of constant streams of data from a wide variety of sensors.

In the People’s Republic of China (hereafter “China”), won a national pilot for the use of its smart drone technology, which has been in development since 2015 (Meredith and Kharpal, 2017[98]). Seven types of drones are in use across four Chinese provinces to deliver packages weighing from 5 to 30 kilogrammes, with plans to test the delivery of packages up to 1 000 kilogrammes. The drones are battery powered, can fly up to 100 kilometres per hour, and could reduce shipping costs by up to 70%. has plans to open more than 200 drone airports in mostly rural areas across China (, 2017[99]), potentially driving access to a larger scope of goods and services through e-commerce. In collaboration with the Japanese e-commerce platform Rakuten, will deploy this technology for local deliveries in Japan in 2019 (Lim, 2019[100]). However, many countries and regions still ban the commercial use of drones in urban areas (ITF/OECD, 2018[101]).

Another interesting development is the use of autonomous ground vehicles and robots for last mile deliveries. Some forms of lightweight autonomous robots are being tested to make deliveries in the city of Washington, District of Columbia (D.C.). These pilot robots are approximately 70 centimetres tall, weigh approximately 23 kilogrammes and can carry approximately 10 kilogrammes of products, but are restricted to speeds of just 6.5 kilometres per hour. The robots use cameras, computer vision and artificial intelligence to navigate sidewalks and urban environments, albeit currently with human supervision (Lonsdorf, 2017[102]), and just five such robots made over 7 000 deliveries over a period of 18 months. Most transport and zoning regulations do not have provisions that enable the use or testing of such devices. Recently, a pilot program was extended to enable further testing of so-called “personal delivery devices” in Washington, D.C. (Washington DC District Department of Transport, 2018[103]).

Online-offline e-commerce business models

Consumers increasingly leverage online tools throughout the lifecycle of the commercial process, including for research, price comparison, delivery and customer service (Verhoef, Neslin and Vroomen, 2007[104]). This includes the ordering process, which is at the core of an e-commerce transaction (see Chapter 1). Digital technologies, including mobile applications, self checkouts, electronic kiosks and smart shelf technology, have become increasingly embedded in each stage of the retail process. To a large extent, the ubiquity of these tools is driven by the rise of mobile commerce and the spread of high-speed broadband in many parts of the world.

Many firms have taken advantage of the ubiquity of digital technologies to develop business models based on a combination of both online and offline features. These business models serve as extensions of e-commerce, pushing the frontier of online purchases into physical stores (see Box 3.12). Some business models combine online ordering with offline distribution, which may be useful to enable the online purchase of products whose quality may not be assessed from a distance, such as perishable goods like groceries. Other online firms are moving offline by adding brick-and-mortar elements to enable the online sale of other goods, like clothing, where fit may be difficult to assess from a distance. Finally, some business models combine offline and online mechanisms by bringing online ordering very close to the point of purchase, including within brick-and-mortar stores.

3.12. Mobile technologies are helping e-commerce to flourish in brick-and-mortar stores

Mobile technologies enable consumers to conduct a range of digital activities, including online shopping. Consumers use digital technologies throughout the commercial process, but smartphones enable shoppers to research, compare prices, and ultimately make transactions from any networked location. A Google survey found that 82% of surveyed consumers research products on their smartphones before making purchases in brick-and-mortar stores (Google, 2016[105]). The actual act of purchasing through a smart, connected device is referred to as “m-commerce,” whereby consumers complete a commercial transaction using a mobile device (OECD, 2013[106]).

OECD work highlights m-commerce is an emerging trend (OECD, 2013[106]) with ramifications for consumer protection (OECD, 2018[107]). The ubiquity of smartphones is the foundation for many of the online aspects that can be increasingly found in offline commerce. Smartphone features also enable other innovations in the retail space, including location-based offers (Grewal et al., 2016[108]). Similarly, the ability to purchase goods and services by mobile application is central to some of the business model innovations outlined in more detail below.

Online-offline distribution mechanisms

As e-commerce has become more prevalent, many conventional firms and retailers have experimented with the inclusion of online distribution channels alongside their existing brick-and-mortar operations. However, leveraging the Internet, or other electronic networks, to integrate e-commerce into an existing firm-level business model often requires a range of complementary investments and capacities. This can include, but is not limited to, supply chain and fulfilment arrangements and consolidated inventory systems.

The most common form of e-commerce consists of simple direct shipping to the customer’s home after they have purchased online from a distance and without the physical inspection of the product. However, this carries delivery costs arising from dispersed fulfilment and delivery chains, or dependence on third-party logistics services (see Box 3.11). Thus, many firms are experimenting with mechanisms for fulfilment through existing physical infrastructure or retail locations.

For example, many firms have developed “click-and-collect” mechanisms to enable consumers to order and purchase online, while collecting the relevant items in a local brick-and-mortar store or another location such as a locker. This allows consumers to immediately purchase the good or service online, but to save on shipping costs, delays and inconveniences associated with delivery. Notably, this mechanism enables firms to retain their current centralised inventory system and reduces operational costs associated with physical brick-and-mortar stores, while also acquiring useful data about the user.

To the extent that click-and-collect mechanisms are located in a brick-and-mortar store, this mechanism may also allow consumers to check the quality and assess the colour, style and size of the product within the store itself. In addition, consumers are able to make returns in store, which may increase their propensity to purchase online. A survey by the United Parcel Service (UPS) finds that consumers were more willing to purchase online if they could return in a brick-and-mortar store (United Postal Service, 2018[109]).

Other interesting developments in this space include curbside fulfilment, whereby consumers can order groceries online and then drive to their local brick-and-mortar store for immediate fulfilment (Howland, 2016[110]). This model enables retailers to minimise expensive investments in home-delivery supply and logistics systems, and has been adopted by major retailers like Walmart, Amazon, Target and Nordstrom.

Online groceries – A new e-commerce frontier

In its most conventional form, e-commerce consists of committing to buy a product online, usually using a device that is not close to the product purchased. However, many products are perishable, or they may have quality characteristics that are difficult to assess online, through pictures or at a distance. For example, private surveys have found consistently high estimates (69% to 84%) of surveyed US consumers prefer to purchase fruits and vegetables in brick-and-mortar stores in order to physically assess the quality and suitability of the product, which may be because fresh products are of variable quality and individuals have subjective tastes (Griswold, 2017[111]; Nielsen, 2017[112]). Perhaps as a result, this category of online shopping has seen comparatively little growth so far (see Chapter 2).1

Developing an effective method of selling groceries online could be an opportunity for firms, as groceries, including food and non-alcoholic beverages, accounted for 14% of household expenditure on average across 34 OECD countries in 2017 (OECD, 2019[113]). Some private sources find that consumers take 1-2 trips to the supermarket each week, representing significant friction and inconvenience that could be mitigated through e-commerce (Food Market Institute, 2017[115]).

Many firms have attempted to develop processes to successfully sell perishable food and groceries online. Some online business models offer direct shipping of purchased groceries alongside guarantees related to quality and customer satisfaction to give consumers confidence in the purchase of perishable goods and services. However, this model typically necessitates the development of a cold-chain direct fulfilment and logistics system, which is expensive. Instacart, an emerging start-up in this space, enlists contract workers to physically select the relevant groceries ordered by the consumer and drop them off at their desired location. Instacart provides detailed instructions to its contractors about how to select produce in order to meet minimum quality standards (Griswold, 2017[111]).

Some firms include online and offline components in order to successfully sell groceries online. For example, Alibaba, a Chinese e-commerce online platform, has refined its business model to facilitate the sale of perishable goods and services by opening brick-and-mortar stores. Over 25 Hema supermarkets have been opened across China for the sale of perishable goods including live seafood, fruits and vegetables. The supermarkets also act as fulfilment centres for purchases made online through the Hema application. Customers within a 3 kilometre radius can order products online and have their orders hand-selected by store workers and then delivered to their home within 30 minutes using smart logistics technology (Xiaohan, 2017[116]).

Data and insights from its online business enable Alibaba to optimise the brick-and-mortar experience inside Hema stores, resulting in 300% to 500% more sales per unit area than other supermarkets (Najberg, 2017[117]). Customers who use the application make purchases up to 35% of the time they use the application (conversion rate), while online orders account for more than 50% of all sales from Hema. Notably, the rival Chinese e-commerce firm is following suit by opening new digitally infused brick-and-mortar experiences and also has plans to develop fully automated convenience stores (see below) (Bloomberg, 2018[118]). A global survey found that Chinese customers may be more willing than other consumers to buy groceries online (Figure 3.1).

3.1. Consumer preferences for buying groceries, 2017
 3.1. Consumer preferences for buying groceries, 2017

Note: See Chapter notes.1

1. Figure 3.1: The total global sample was 24 073; the Chinese sample was 894.

Source: PricewaterhouseCoopers (2017[119]) Total Retail Report 2017,


In Alibaba’s Hema supermarkets, a companion mobile application is available to enable shoppers to determine the origin and provenance of perishable fruits and vegetables. Another new technology trialled for product safety and quality includes the use of distributed ledger technologies to track product provenance (see Box 3.13).

3.13. Distributed ledger technologies help trace product provenance

Existing applications of distributed ledger technology in e-commerce largely relate to tracking the origin and provenance of goods and services to ensure product safety and quality. For example, IBM has collaborated with a range of retailers including Walmart and Costco (Aitken, 2017[120]) to develop a permissioned distributed ledger database that tracks products from supplier to retailer, including information on production, inspection, farm of origin, factory and processing, expiry date, storage temperatures and shipping (Unuvar, 2017[121]).

This is of particular use in systems with multiple intermediaries, including producers, wholesalers, retailers, couriers and regulators, each of which have incomplete information about the upstream and downstream production process. Distributed ledger technologies transform a previously cumbersome and manual process into a system that can provide specific information about product provenance in less than 3 seconds (Unuvar, 2017[121]). A system like this could also be used to ensure that illegal products are not traded, or to ensure that foods labelled “organic” have indeed passed the requirements to be labelled as such.

Going offline to ensure the right fit

Some consumer goods, like clothing, can vary in their suitability based on their fit to the consumer. Thus, while clothing is one of the most popular products bought and sold online (see Chapter 2), the sector has struggled with significant returns. As noted in Chapter 2, approximately 5% of given online transactions were refunds, of which 57% are related to the fashion industry. Consumers may often buy multiple items of a similar size, shape or colour, with the intention of trying them on and returning those that do not fit (Orendorff, 2019[122]).

As a consequence, an interesting emerging e-commerce business model for online fashion businesses is the inclusion of offline features to enable the sale of fit-critical goods and services online. An offline distribution channel re-introduces frictions to the business model and may increase costs, but also increases the extensive margin of e-commerce by enabling new types of products to be sold online. In particular, firms that sell heterogeneous or bespoke products like clothing may benefit from this kind of business model.

For example, several online apparel retailers have opened brick-and-mortar stores that allow consumers to try on products before ordering them online. Bonobos, an American online retailer, has opened over 30 “guideshops” to enable consumers to try the product for fit and quality (Waldron, 2019[123]). Consumers then place their order online, a process that increases conversion, minimises returns and increases the average purchase value. A similar model is pursued by bespoke tailor Indochino, whose service involves an in-person suit fitting service with a second follow-up and in-person alteration. A physical component of the traditional e-commerce experience has been added by online brands including Amazon Books, Birchbox, Bonobos, Daniel Wellington, Harry’s and Warby Parker.

A unique feature of the offline stores provided by these online retailers is that they are purely show-rooms – that is, spaces that are purely for fitting purposes, and the clothes or goods on display at the shop cannot actually be purchased on the spot. The inventories themselves are typically highly curated with a limited selection of inventory to try on. Data from the online business regarding consumer choices and preferences facilitate this selection. This enables the online retailer to avoid complicated “balance on hand” questions about store-level inventory that typically require an extensive separate inventory and logistics chain. This also erodes the so-called institutional capability of traditional retailers to choose the optimal product mix throughout the supply chain, a previous source of comparative advantage (Hodson, Perrigo and Hardman, 2017[124]).

As the actual ordering and purchase of goods and services is typically conducted online (via application), the brand can retain a centralised inventory and distribution chain, and avoid the need for an additional payment system. Moreover, the process increases the conversion of online sales while reducing the additional costs associated with returns of unwanted products. The spaces are often smaller than other brick-and-mortar retailers, and often leased on demand. This approach also increases the productivity of the space itself – as much as five times the revenue per square foot in comparison to legacy retailers – while reducing overall overhead costs (Taggart and Granville, 2017[125]).

These initiatives are also notable as they have been adopted by vertically-integrated brands, whose low price point and value offering is predicated on the ability to cut overall production costs through the curation of the entire product production lifecycle from design to sales. While the offline features re-introduce frictions into the purchasing process for consumers, the seller is able to leverage insights from their core online service offering to maximise the utility of the offline space while minimising the expensive and inefficient burdens associated with traditional retail, like inventory management and point-of-service purchasing. Tellingly, some surveys show that consumers already use brick-and-mortar stores as showrooms in order to browse and understand product characteristics before purchasing online (Freeman, 2014[126]; Khan, 2018[127]).

Online ordering in brick-and-mortar stores

Other firms are increasingly experimenting with the inclusion of online ordering mechanisms within or very near brick-and-mortar stores themselves in order to boost sales, enable customisation and increase efficiency. This appears to be a salient trend for businesses that rely on convenience or service on demand.

Ordering, purchase and payment by application or kiosk for almost immediate pick-up has been adopted by many restaurants. For example, the American fast food chain McDonald’s has installed digital self-order kiosks in all 14 000 of its US stores (Hafner and Limbachia, 2018[128]). These kiosks rely on touch-screen technology to relay information via wireless networks from customer orders to the kitchen, where the meals are made on demand. For some restaurants, the use of these kiosks has increased revenues (Hafner and Limbachia, 2018[128]; Garber, 2014[129]). Users tend to spend more time considering their options when using an automated kiosk, which can result in selecting more items for purchase (Houser, 2018[130]). Similarly, increased revenues may result because users are more likely to customise their orders, which typically carries an additional fee. Academic work also finds that online ordering resulted in 14% more customisation requests than orders made in person (Goldfarb et al., 2015[131]).

Other firms have experimented with the use of ordering by mobile application prior to almost immediate pick-up from the store. McDonald’s, for example, has plans to further enable ordering and payment by mobile application, while Starbucks allows consumers to “skip the queue” by purchasing via application and picking up in store. Some reports suggest that orders made through mobile application can be up to 20% higher than other orders (Wong, 2015[132]).

An emerging and innovative example of embedding online ordering mechanisms within brick-and-mortar stores is the partially automated grocery store pioneered by Amazon (Amazon, 2018[133]). A mobile application facilitates entry into the store, after which point consumers can simply select the products they wish to buy, and then immediately leave the store without a formal checkout process. The aim of their business model is to increase the efficiency of the shopping experience by partially automating the payment process.

The partially automated grocery store, known as Amazon Go, uses geo-fencing technology, embedded sensors, computer vision, infrared detection and deep learning to automatically detect when consumers take products from shelves. Purchase and payment is undertaken automatically through the digital wallet associated with the customer’s Amazon account (see section on innovative payment mechanisms below). The process is more efficient and convenient than other grocery stores (McFarland, 2018[134]), and removes frictions from the typical brick-and-mortar grocery experience by partially automating the payment process. Notably, the goods and services for sale are mostly pre-packaged (e.g. sandwiches), with few items with significant variation in quality like fresh produce. Alibaba and major online firm Tencent have also developed unmanned convenience stores that use quick response (QR) codes for entry and purchase (Soo, 2017[135]; Zhang, 2018[136]), while other automated stores in Korea have experimented with palm vein scanning for custom verification (Chang-won, 2018[137]).

On the intensive margin, the Internet of Things and connected devices have been used as mechanisms for digital marketing to boost sales within physical stores. Digital advertising via mobile phones can be used to provide shoppers with discounts based on proximity, providing personalised offers (Grewal et al., 2016[108]). For example, Coca Cola has developed new forms of digitally-interactive marketing that uses browsing history, IP address, approximate age and gender to develop custom advertising and discounts to consumers inside some American supermarkets (Darrow, 2017[138]).

Innovative payment mechanisms

Safely and remotely exchanging money online, including across borders, is fundamental to e-commerce. Safe and effective online payment mechanisms facilitate trusted online transactions, boosting the growth of e-commerce between unknown actors. Many online payment mechanisms are closely associated with the rise of e-commerce. In fact, one of the earliest online payment models – Paypal – emerged in conjunction with the pioneering online auction house and e-commerce platform eBay, to enable safe online payments between peers.

Online payment service innovations are forms of “Fintech”, a broad term that encompasses the application of digital technologies and processes to financial services. Recent OECD work considers other Fintech applications at more length (OECD, 2018[139]). For consumers, safe online payment mechanisms increase convenience and trust online. Conversely, payment security concerns are among the most cited reasons for not transacting online (see Chapter 2), and these concerns remain prevalent across many countries. In the EU28, the share of people reporting not having bought online due to payment security or privacy concerns declined by almost 10 percentage points from 2017 to 2009 (OECD, 2019[52]).

For firms that sell online, enabling consumers to purchase more efficiently can help to convert and actualise payments. Some payment mechanisms may also help consumers access new markets. E-commerce firms tend to offer a mix of payment options, including through local or established payment mechanisms, to provide maximum convenience for their customers. For example, an e-commerce firm may choose to offer credit and debit card payments through a secure payment gateway on their website as well as other options, such as Paypal or other online payment mechanisms. There are also established mechanisms for enabling payments using local currencies through the establishment of local banking relationships.

Looking to the future, this section focuses on three innovative forms of holding and conducting payments that could facilitate e-commerce going forward: Digital wallets, mobile money and cryptocurrencies. These mechanisms are not necessarily discrete – indeed, mobile money and cryptocurrencies are both stored in forms of digital wallets – but together, they have the potential to drive future developments in the e-commerce landscape. Just as initial innovations in online payments facilitated the early online commercial transactions, new payment mechanisms could boost e-commerce along both the intensive and the extensive margin and drive changes to existing e-commerce business models.

Digital wallets can foster e-commerce

An online payment can be broadly considered to be a “purchase order placed using devices connected to the Internet”, a definition that is relevant to many forms of e-commerce (OECD, 2018[139]). One mechanism of enabling online payments includes the use of digital wallets, also known as “e-wallets” or “electronic wallets”. These wallets act as intermediating application layers that hold financial information about the relevant funding source on both sides of the transaction (e.g. credit card details) (Cheok, Huiskamp and Malinowski, 2014[140]). Essentially, digital wallets tokenise financial information such that the information does not have to be directly shared with an unknown party. Many digital wallets can hold, send and receive funds.

Digital wallets enable e-commerce by facilitating trusted transactions online, without which most e-commerce purchases would not be possible. Consumers may be more willing to undertake a transaction online using a digital wallet as opposed to directly sharing financial information with online retailers. For e-commerce firms, therefore, enabling payments through a digital wallet service can help sellers reach more potential customers.

Digital wallets vary in their service offerings and features. Some wallets directly process payments, transferring money between buyers and sellers (e.g. Paypal); others simply transfer financial details between the payment processors of either party (e.g. Google Wallet). Digital wallets can hold a variety of currencies, including cryptocurrencies (see below).

Some digital wallets include specific features, such as deferred payment options or limited credit programmes and fraud prevention schemes. Often these specific features are closely associated with a particular business model or platform. For example, Alipay, the digital wallet associated with Alibaba, is used for payments on Alibaba’s e-commerce platform, known as Taobao. Payments made through Alipay can also be held in escrow until the buyer confirms the satisfactory receipt of an order, a function also performed by other services like Paypal. Similarly, sellers on Alibaba are only required to ship the product after the payment details of the transaction have been confirmed through Alipay.

Digital wallets can be used from any connected device, including mobile phones and other smart devices, like smart watches. Mobile wallets are a sub-type of digital wallets, with mobile-specific features and services (a discussion of mobile money follows below). Some are simply existing digital wallets that have added mobile functionality through a mobile application, such as Paypal Mobile. Others are specific to the particular hardware or software of a particular mobile device, such as Google Pay, Apple Pay or Samsung Pay.

Mobile wallets can be used to make purchases online, but they are also increasingly used at point-of-sale transactions in brick-and-mortar stores using connected devices. Depending on the nature of the transaction – specifically, the nature of the ordering process – purchases using a mobile wallet may constitute e-commerce transactions (see Chapter 1). More generally, the rise of mobile wallets for payment epitomises the increasing digital transformation of all aspects of the commercial process, including payments. This is true in many areas of the world, including China (see Box 3.14).

3.14. Digital wallets: Digitalising payments in the real world

China is the world’s most developed market for mobile payments through digital wallets, driven by the ubiquity and functionality of two digital wallet systems, each associated with major e-commerce platforms. Alipay is a digital wallet system developed by Alibaba, the Chinese e-commerce platform, and WeChat Pay is a digital wallet system developed by Tencent, another Chinese e-commerce platform. Purchases made on Tencent (e.g. bike sharing services) or an Alibaba platform (e.g. Taobao) are often mediated through each respective digital wallet.

Both digital wallets are notable for the scope of their functionalities beyond the simple mediation of the purchase and sale of consumer goods. For example, Alipay can be used to pay credit card bills, buy mobile top-up credit, transfer money and pay for food delivery. Alipay is also increasingly integrated with a range of Chinese public services, meaning that Alipay can be used to pay for expenses including utility bills, transportation fines and school fees. These features may help push out the extensive margin of e-commerce into transactions that were previously analogue.

But both the Alipay and WeChat Pay digital wallets are also notable for the way in which their digital wallets are increasingly used for purchases in the “real world”, namely through their functionalities in brick-and-mortar stores. Both systems work in similar ways. They are typically used through a smartphone and a mobile application that is linked to an existing bank account. Transfers of money from the linked bank account can be made by using the mobile application to scan the QR code of the vendor. The success of this system has reportedly driven the wide uptake of QR codes to street food vendors, cafés, convenience stores, restaurant and retail chains

In contrast to other places with a high share of unbanked people, 85% of those who make online purchases in China also pay for them through an online payment mechanism (World Bank, 2017[141]), suggesting a degree of trust in online payment mechanisms in China. The Chinese market for mobile payments has grown to approximately USD 16 trillion (Jacobs, 2018[142]) and together Alipay and WeChat Pay are used for approximately 93% of mobile payments (Taylor, 2019[143]). Alipay has over 500 million monthly active users, while WeChat Pay boasts over 900 million monthly active users (Jacobs, 2018[142]) though anecdotal reports suggest that Chinese mobile payment users utilise both (The Economist, 2018[144]). A recent study found that 92% of people in a sample of Chinese cities used either of the two mobile wallets as their primary payment method, rather than cash.

Mobile money can help extend e-commerce to the unbanked

Another form of payment innovation that enables e-commerce is the rise of a specific form of mobile payment, also known as “mobile money”, particularly for unbanked people (i.e. those without access to financial services). Mobile money differs from digital wallets insofar that the mechanism for payment is conducted via mobile communication networks, and does not necessarily require an existing relationship with a financial services provider.

Mobile money is mediated by mobile network operators who use a system of agents to accept regular (fiat) currency in the form of cash and store an equivalent value in a digital wallet, which can then be transferred to other users or withdrawn later. Mobile money is typically associated with a mobile phone number and often uses two-factor authentication through a personal identification number issued at the point of registration. Mobile money can typically be transferred to others who are also registered with the same mobile money system, exchanged with merchants for goods and services, or can be withdrawn as cash from a mobile money agent. Mobile money can therefore act as a means of storing and transferring value in a secure and convenient way for unbanked people.

Mobile money is growing in importance, particularly in places with less developed financial markets. In December 2017, there were over 690 million registered mobile money accounts (GSMA, 2017[145]). There are over 270 mobile money systems across 90 countries in the world, and 19 countries have more mobile money accounts than formal bank accounts (GSMA, 2017[145]). There are 30 mobile money systems with more than one million active mobile money customers and overall revenues of the mobile money industry rose to USD 2.4 billion in 2017 (GSMA, 2017[145]).

3.15. M-Pesa: Enabling e-commerce for the unbanked

The deployment of the mobile money service M-Pesa, used ubiquitously in Kenya, is the canonical examples of the successful take up of mobile money. M-Pesa is also one of the few mobile money systems used for e-commerce, indicating how mobile money can enable e-commerce in developing economies.

M-Pesa developed as a means of enabling microfinance payments and easier remittances through basic mobile phones with 2G cellular capability. M-Pesa has come to be used in a variety of online and offline transactions, including for point-of-sale payments. The mobile application features an additional extension known as Lipa na Mpesa, which enables mobile money payments to other users, including to enterprise bank accounts.

In addition, a competitive marketplace of payment services has emerged that enable mobile money to be used for online payments. For example, Safaricom (the mobile network operator that operates M-Pesa) and Paypal recently collaborated in order to enable the transfer of funds from Paypal accounts to M-Pesa wallets (Safaricom, 2018[146]). Safaricom has also launched an e-commerce platform that enables payments via M-Pesa (Masoko, 2018[147]).

In 2017, M-Pesa served nearly 30 million customers through 287 400 agents across 10 countries (Vodafone, 2017[148]). In Kenya, M-Pesa is the dominant form of mobile money, and it is used in 81% of transactions, while 65% of all mobile money transfer transactions related to mobile commerce payments (Communications Authority of Kenya, 2018[149]). M-Pesa can be used to purchase a variety of goods and services online, including a flight on Kenya Airways (Kenya Airways, 2018[150]). M-Pesa has also had transformative impacts on Kenya’s overall economic development, driving increased long-term consumption and a reduction of the households in extreme poverty, particularly for women (Suri and Jack, 2016[151]).

Mobile money is relatively underutilised in e-commerce today, even though e-commerce is increasingly prevalent in areas where mobile money accounts are prevalent (see Chapter 2). The greatest share of transactions undertaken through mobile money was for direct person-to-person transfers, meaning that firms generally do not transact online using mobile money (GSMA, 2017[145]). Just over half of all e-commerce transactions in developing countries are paid for using cash on delivery according to a recent World Bank study, with the exception of China (World Bank, 2017[141]). This may be because consumers in developing countries prefer to inspect the product before payment. Hortacsu et al. (2009[152]) finds that same-city sales on eBay and a Brazilian e-commerce platform known as Mercado Libre are disproportionately high because users prefer to make purchases of products that can be observed and delivered in person.

However, mobile money has the potential to facilitate e-commerce transactions and online purchases for the otherwise unbanked. To do so, better integration and collaboration will need to take place between actors in this space, including mobile network operators, banks and e-commerce firms, to ensure that mobile money can be used for other use cases (Osafo-Kwaako et al., 2018[153]). In particular, reliable integration between mobile money systems and online interfaces for e-commerce firms, for example through application programme interfaces, may better enable the uptake of mobile money as a payment mechanism for online shopping (Camner, 2016[154]).

Similarly, enabling interoperability and transactions between mobile money systems could facilitate peer-to-peer e-commerce. Efforts are underway to ensure that users of different mobile money services can transact directly with each other in Kenya (Central Bank of Kenya, 2018[155]), Bolivia (Novais and Sanin, 2018[156]), and Ghana (Ghana Web, 2018[157]), among other places.

Blockchain-enabled payment mechanisms may boost e-commerce

Another emerging payment mechanism involves the use of distributed ledger technologies as a payment mechanism, also known as cryptocurrencies. Cryptocurrencies like bitcoin operate through a distributed database independent of central banks or financial institutions and provide a means of making anonymous, validated transfers of value. However, more potential for e-commerce may lie in other extensions of blockchain-enabled payments, including the potential use of “smart contracts”, namely self-executing and deterministic software protocols that only transfer value after particular conditions are met.

Cryptocurrencies can be used as fiat currency, and can be managed through digital wallets (see above) or on a connected device. All transactions are permanently recorded on a distributed ledger, thereby preventing fraud by ensuring the same currency cannot be spent twice. In general, cryptocurrencies are generated by “mining” through the use of computing power to solve increasingly difficult algorithms. Cryptocurrencies can therefore act as an algorithmically trustworthy but anonymous mechanism of transferring value online. Notably, they have been used to facilitate the online purchase of contraband or counterfeit goods and services on marketplaces like the Silk Road and TOR (OECD, 2017[18]).

The use of cryptocurrencies in more legitimate e-commerce transactions has been mixed. Notably, payment gateway service Stripe and online gaming store Steam both recently stopped supporting cryptocurrency payments, citing high transaction costs (Russell, 2018[158]). More recently, there has been an increasing uptake of payment services that convert cryptocurrencies to fiat currencies through more traditional e-commerce sites. For example, Coinbase recently announced a partnership with Shopify, an e-commerce hosting site with over 500 000 merchants worldwide (Shopify, 2018[159]), which enables these merchants to accept payments in over 300 cryptocurrencies. Coinbase argues that there are reduced processing fees and fewer risks of charge-backs for merchants who accept cryptocurrencies as payment, while shoppers enjoy more simplicity, anonymity and low transaction fees.

However, extensions of blockchain technology other than cryptocurrencies may hold more promise for enabling e-commerce. In particular, so-called smart contracts may be of use in some contexts. Smart contracts are often used to implement escrow systems, which could be useful to facilitate trusted e-commerce transactions between unknown parties. For example, escrow contracts on OpenBazaar (see Box 3.7) ensure that an e-commerce seller is only paid once the buyer confirms that they have satisfactorily received the product (OECD, 2017[18]).

On the extensive margin, smart contracts could also help to enable the safe online ordering of many different kinds of products currently outside the remit of e-commerce. For example, a blockchain-enabled smart contract could be used to execute a rental agreement, where the conditions of payment are predicated on the property owner and renter agreeing to certain rules or conditions like the rental amount or the date of the transfer of keys. To date, transactions of this nature have been carried out in person, but blockchain-enabled smart contracts could make such transactions a form of e-commerce.

Smart contracts could also hold particular promise for e-commerce when combined with connected devices. For example, a blockchain-enabled, connected washing machine could initiate an e-commerce transaction through a smart contract when it detects that it is out of detergent (OECD, 2017[18]), as an extension of the e-commerce subscription business model outlined earlier (see Box 3.10). Nevertheless, connected devices could also potentially transact with each other in an autonomous fashion using smart contracts, thereby facilitating a completely new kind of e-commerce. A bitcoin-based start-up called 21 has outlined a model whereby environmental sensors could passively collect data and sell it to other machines or institutions for micropayments of cryptocurrencies, like bitcoin (Pate, Kun and Srinivasan, 2016[160]). Blockchain technology could therefore enable e-commerce transactions between connected devices, rather than simply between individuals and firms. Such an innovation could drive new e-commerce business models.

Key areas for policy action

As digital transformation progresses, new business models will arise in ways that are difficult to predict. Business model innovations that make use of data and digital technologies often challenge traditional policy frameworks, particularly for firms that use new business models to innovate across and between sectors, new payment mechanisms, or technologies in a new and innovative way.

Remove regulatory barriers that preserve artificial distinctions between online and offline commerce

Technological changes have blurred the boundaries between online and offline activities, as well as between goods and services. This has an impact on policy settings that often rely on an increasingly artificial distinction between traditional commerce and e-commerce. Because firms are increasingly combining the most promising aspects of both worlds, the level of ambiguity will rise.

For example, the increasing use of a mix of online and offline distribution models means that brick-and-mortar stores increasingly perform functions beyond the simple point-of-service purchase of products. Instead, physical stores often act as an extension of the online experience facilitated by e-commerce, and vice versa. Innovative business models may use brick-and-mortar stores as a point of collection or return of products bought online, or as a temporary storage facility before delivery. Existing licensing, permitting or zoning rules – particularly at the local level – may not allow such functions, and in doing so constrain the development of promising e-commerce business models (e.g. omni-channel models).

As consumers choose to order tangible goods online and have them delivered to a specific location, firms come under pressure to integrate digital and physical supply chains to ensure fast and responsive delivery. Consequently, firms are experimenting with logistics and fulfilment solutions to enable the physical delivery of products over the last mile, or the final leg of the delivery journey to a particular location. Some innovations in this area, including the use of autonomous robots and unmanned aerial vehicles, may not have been anticipated in existing road and sidewalk rules. Many such rules are local, which further underscores the need for a consistent and co-ordinated whole-of-government approach to e-commerce policy making at all levels of government.

Encourage regulatory flexibility, experimentation and transparency

Regulatory uncertainty can reduce the incentive to invest, and it may constrain the ability to scale as investors may be unwilling to invest in a firm with an untested product, service or business model. Policy experimentation can help ensure a firm’s ability to innovate while remaining within the spirit of existing laws. Outcome or performance-based regulations that specify required outcomes or objectives can be useful in this regard. Regulatory sandboxes are another mechanism of ensuring regulatory flexibility that enable selected firms to test innovative products or services through approaches and competences that differ from existing policy frameworks. In the e-commerce context, such sandboxes have been used to test the use of drones for delivery and digital payment mechanisms. As e-commerce business models evolve, additional applications may arise with respect to new technologies, like blockchain or 3D printing.

At the same time, policies that focus on a particular type of e-commerce business model should be avoided. Given the dynamism of the e-commerce landscape, some e-commerce business models that are dominant today may not be dominant in the future. For example, while e-commerce business models that use online platforms are among the most prominent in the current e-commerce landscape, advances in decentralised structures like distributed ledger technologies may diminish this role in the future. An alternative approach is to ensure that particular business functions conform with the regulatory framework, while better accounting for the interlinkages across business functions.

Increased transparency, including through better communication of existing regulations and their specific application to e-commerce, is another important step in reducing uncertainty for innovative firms. The Interpretative Guidelines on Electronic Commerce and Information Property Trading first formulated in 2002 with regard to e-commerce activities in Japan (METI Japan, 2016[161]) provide a useful example of such an approach.


Aitken, R. (2017), “IBM forges blockchain collaboration with Nestlé and Walmart in global food safety”, Forbes, (accessed on 19 April 2018). [120]

Alibaba (2017), “Alibaba signs MoU with Malaysia’s MDEC and Hangzhou municipal government to facilitate global trade for SMEs under eWTP”, Alibaba, (accessed on 19 April 2018). [65]

Alizila (2016), “Cainiao network overview”, (accessed on 19 April 2018). [16]

Amazon (2018), Amazon Annual Report - 2018, (accessed on 12 April 2019). [61]

Amazon (2018), Amazon Go website, (accessed on 23 October 2018). [133]

Amazon (2018), “Amazon letter to shareholders - 2017”, (accessed on 19 April 2018). [95]

Amazon (2018), “Amazon services”, (accessed on 4 April 2018). [62]

Amazon (2018), “Managing voice purchasing settings”, (accessed on 19 April 2018). [85]

Amazon (2016), “Amazon Dash replenishment”, (accessed on 19 April 2018). [91]

Amazon (2016), “Generating recommendations at Amazon scale with Apache Spark and Amazon DSSTNE”, Amazon, (accessed on 19 April 2018). [21]

Amazon Investor Relations (2015), “Amazon sellers sold record-setting more than 2 billion items in 2014”, (accessed on 19 April 2018). [60]

Anderson, C. (2004), “The long tail”, Wired, (accessed on 19 April 2018). [13]

Bakos, Y. and E. Brynjolfsson (2000), “Bundling and competition on the Internet”, Marketing Science, Vol. 19/1, pp. 63-82, [69]

Bakos, Y. and E. Brynjolfsson (1999), “Bundling information goods: Price, profits and efficiency”, Management Science, Vol. 45/12, pp. 1613-1630, [68]

Bateman, J. (2016), “Behold China’s answer to Amazon Echo: The Linglong Dingdong”, Wired, (accessed on 19 April 2018). [86]

Bloomberg (2018), “Online retailer JD follows Alibaba into the supermarket game”, Bloomberg, (accessed on 19 April 2018). [118]

Bloomberg (2017), “Wal-Mart to discount one million online items picked up in stores”, Bloomberg, (accessed on 19 April 2018). [10]

Bright Local (2017), “Local consumer review survey”, (accessed on 19 April 2018). [48]

Broida, R. (2019), “How to spot fake reviews on Amazon, Best Buy, Walmart and other sites”, CNET, (accessed on 19 April 2018). [53]

Brynjolfsson, E., Y. Hu and M. Smith (2003), “Consumer surplus in the digital economy: Estimating the value of increased product variety at online booksellers”, Management Science, Vol. 49/No.11, pp. 1580-1596, [12]

Business Insider (2017), “Amazon Prime subscribers hit 80 million”, (accessed on 19 April 2018). [96]

Cabral, L. and L. Li (2014), “A dollar for your thoughts: Feedback-conditional rebates on eBay”, [45]

Cadell, C. (2017), “Alibaba launches low-cost voice assistant amid AI drive”, Reuters, (accessed on 19 April 2018). [87]

Camner, G. (2016), “Launching GSMA mobile money APIs to raise industry capabilities”, GSMA, (accessed on 11 April 2019). [154]

Cao, J. and M. Mittelman (2016), “Why Unilver really bought Dollar Shave Club”, Bloomberg, (accessed on 19 April 2018). [84]

Catalini, C. and J. Gans (2017), “Some simple economics of the blockchain”, [58]

Central Bank of Kenya (2018), “Press release: Mobile money interoperability”, (accessed on 10 April 2018). [155]

Chang-won, L. (2018), “7-Eleven store manned by robot equipped with palm vein scanner and AI”, Aju Business Daily, (accessed on 10 April 2019). [137]

Chen, P., S. Dhanasobhon and M. Smith (2008), “All reviews are not created equal: The disaggregate impact of reviews and Rreviewers at”, [47]

Chen, T. et al. (2017), “Thinking inside the subscription box: New research on e-commerce consumers”, McKinsey, (accessed on 19 April 2018). [83]

Cheok, L., W. Huiskamp and A. Malinowski (2014), “Beyond payment – E-commerce trends and payment challenges for online merchants expanding e-commerce operations internationally”, ModusLink Global Solutions, [140]

Chevalier, J. and D. Mayzlin (2006), “The effect of word of mouth on sales: Online book reviews”, [46]

China Money Network (2017), “JD.Com CEO says AI will cut staff by 1/3, grow revenue 1000% in ten years”, (accessed on 19 April 2018). [15]

Chlu (2018), “Reviews and ratings verified by payments on blockchain”, (accessed on 19 April 2018). [59]

Communications Authority of Kenya (2018), “Second quarter sector statistics report for the financial year 2017/2018”, Communications Authority of Kenya, (accessed on 19 September 2018). [149]

Darrow, B. (2017), “Coca-Cola looks to Google to boost retail sales”, Fortune, (accessed on 19 April 2018). [138]

Deloitte (2017), “Digital identity: Identity is the fabric of the digital economy”, Deloitte Touche Tohmatsu Limited, (accessed on 15 June 2018). [34]

Deloitte (2017), “Digital media: The subscription prescription”, Deloitte Touche Tohmatsu Limited, (accessed on 19 April 2018). [71]

Domingo, A. and A. Enriquez (2018), “Digital identity: The current state of affairs”, (accessed on 15 June 2018). [36]

eBay (2018), “Global shipping program”, eBay, (accessed on 19 April 2018). [63]

eBay (2016), “Small online business growth report: Towards and inclusive growth economy”, (accessed on 19 April 2018). [66]

Eggleton, M. (2016), “Creating one digital identity could make us more secure”, Australian Financial Review, (accessed on 15 June 2018). [35]

Einav, L., C. Farronato and J. Levin (2016), “Peer-to-peer markets”, Stanford Institute for Economic Policy Research, [24]

Einav, L. et al. (2017), “Auctions Versus Posted Prices in Online Markets”, Journal of Political Economy, Vol. 126, No. 1, pp. 178-215, [27]

Ellison, G. and S. Ellison (2018), “Match quality, search, and the Internet market for used books”, NBER Working Paper, NBER Working Paper No. 24197, [11]

Ember, S. (2018), “New York Times Co. subscription revenue surpassed $1 billion in 2017”, New York Times, (accessed on 19 April 2018). [75]

European Commission (2015), “Freemium: Zero marginal cost”, [73]

Financial Times (2019), “Choose the subscription that is right for you”, Financial Times, (accessed on 18 April 2019). [94]

Financial Times (2018), “Revenue streams: Spotify’s bid to generate a profit”, (accessed on 19 April 2018). [79]

Financial Times (2016), “The algorithms using data to suggest your next purchase”, (accessed on 19 April 2018). [23]

Food Market Institute (2017), “US grocery shopper trends 2017: Food retail implications for US grocery shopper trends”, (accessed on 19 April 2018). [115]

Fradkin, A. (2017), “Search, matching, and the role of digital marketplace design in enabling trade: Evidence from Airbnb”, [19]

Fradkin, A. et al. (2017), “The determinants of online review informativeness: Evidence from field experiments on Airbnb”, [43]

Freeman, M. (2014), “Showrooming: It’s an opportunity for retail, not a threat”, The Guardian, (accessed on 19 April 2018). [126]

Garber, M. (2014), “Chili’s has installed more than 45 000 tablets in its restaurants”, The Atlantic, (accessed on 10 April 2019). [129]

Ghana Web (2018), “What you need to know about the mobile interoperability”, [157]

Goldfarb, A. et al. (2015), “The effect of social Interaction on economic transactions: Evidence from changes in two retail formats”, Management Science, Vol. 61/12, pp. 2963-2981, [131]

Goldfarb, A. and C. Tucker (2017), “Digital economics”, [70]

Google (2018), Product Reviews and Mobile Searches, (accessed on 19 April 2018). [50]

Google (2016), “5 ways consumers connect to store with mobile shopping”, (accessed on 19 April 2018). [105]

Grewal, D. et al. (2016), “Mobile advertising: A framework and research agenda”, Journal of Interactive Marketing, Vol. 34, pp. 3-14, [108]

Griswold, A. (2017), “There’s still one big reason why people aren’t buying their groceries online”, Quartz, (accessed on 19 April 2018). [111]

GSMA (2017), “Distribution 2.0: The Future of mobile money agent distribution networks”, GSMA Association 2018, (accessed on 7 November 2018). [145]

Hafner, J. and D. Limbachia (2018), “McDonald’s: You buy more from touch-screen kiosks than a person. So expect more kiosks”, USA Today, (accessed on 10 April 2019). [128]

Hilton, M. et al. (2019), Extended Producer Responsibility (EPR) and the Impact of Online Sales, OECD Publishing, Paris, [5]

Hiu, X. (2016), “E-commerce platforms and international trade: A large-scale field experiment”, Management Science, (accessed on 19 April 2018). [64]

Hodson, N., C. Perrigo and D. Hardman (2017), “PwC Retail Trends 2017”, Strategy& PwC, (accessed on 10 April 2019). [124]

Hortacsu, A., F. Martinez-Jerez and J. Douglas (2009), “The geography of trade in online transactions: Evidence from eBay and Mercadolibre”, American Economic Journal: Microeconomics, Vol. 1/1, pp. 53-74, [152]

Horton, J. (2014), “Misdirected search effort in a matching market: Causes, consequences and a partial solution”, Proceedings of the fifteenth ACM conference on economics and computation, p. 357, [20]

Houser, K. (2018), “Ordering food via touchscreen is so fun you spend more money when you do it”, Futurism, (accessed on 10 April 2019). [130]

Howland, D. (2016), “How retailers can use curb side pickup to build customer loyalty”, Retail Dive, [110]

ITF/OECD (2018), (Un)certain Skies? Drones in the World of Tomorrow, ITF/OECD, Paris, [101]

Jacobs, H. (2018), “One photo shows that China is already in a cashless future”, Business Insider, (accessed on 19 September 2018). [142] (2018), “ launches highly automated warehouse in Shanghai, sets new standard for China logistics”, (accessed on 4 April 2019). [14] (2017), “Drone delivery program”, (accessed on 19 April 2018). [99]

Kenya Airways (2018), “Mobile money payments for Kenya Airways”, Kenya Airways, (accessed on 13 November 2018). [150]

Khan, H. (2018), “Consumers are showrooming and webrooming your business, here’s what that means and what you can do about it”, Shopify, (accessed on 10 April 2019). [127]

Kim, E. (2017), “Amazon Alexa users buy more stuff, study shows”, CNBC, (accessed on 19 April 2018). [88]

Kumar, V. et al. (2011), “The New York Times paywall”, (accessed on 19 April 2018). [74]

Le Monde (2019), “Abonnements”, Le Monde, (accessed on 18 April 2019). [93]

Levy, A. (2017), “Amazon is telling booksellers to speed up deliery times by up to 40 percent”, CNBC, (accessed on 9 April 2019). [39]

Levy, A. (2016), “Amazon rolls out another policy that has sellers on edge”, CNBC, (accessed on 15 June 2018). [38]

Li, L., S. Tadelis and X. Zhou (2016), “Buying reputation as a signal of quality: Evidence from an online marketplace”, [44]

Lim, S. (2019), “ and Rakuten combine forces to become drone powerhouse in Japan”, The Drum, (accessed on 12 April 2019). [100]

Lonsdorf, K. (2017), “Hungry? Call your neighbourhood robot”, NPR, (accessed on 16 October 2018). [102]

Luca, M. (2016), “Designing online marketplaces: Trust and reputation mechanisms”, [40]

Masoko (2018), “Masoko E-Commerce Platform”, (accessed on 19 April 2018). [147]

Mayzlin, D., Y. Dover and J. Chevalier (2014), “Promotional reviews: An empirical investigation of online review manipulation”, American Economic Review, Vol. 104, pp. 2421-55, [42]

Ma, Z. et al. (2018), “Blockchain for digital rights management”, Future Generation Computer Systems, Vol. 89, pp. 746-764, (accessed on 18 April 2019). [164]

McFarland, M. (2018), “Inside Amazon Go: The camera-filled convenience store that watches you back”, CNN Business, (accessed on 19 April 2018). [134]

McKinsey (2016), “Parcel delivery: The future of last mile”, McKinsey & Company, (accessed on 18 April 2019). [97]

McKinsey (2013), “How retailers can keep up with consumers”, (accessed on 19 April 2018). [22]

Meredith, S. and A. Kharpal (2017), “Chinese e-commerce giant is developing a drone that can deliver packages weighing as much as one ton”, CNBC, (accessed on 19 April 2018). [98]

METI Japan (2016), Results Compiled of the E-Commerce Market Survey, Japanese Ministry of Economy, Trade and Industry, [161]

Nadadhur, S. (2018), “Linking Aadhaar with digital wallets: It’s not all black and white!”, The Hindu, (accessed on 25 June 2018). [33]

Najberg, A. (2017), “Hema Supermarket offers shoppers a ‘new retail’ experience”, Alizila, (accessed on 19 April 2018). [117]

Nielsen (2017), “What’s in store for online grocery shopping”, (accessed on 19 April 2018). [112]

Novais, T. and J. Sanin (2018), “Mobile money interoperability in Bolivia: A case study for Latin America”, GSMA, 2018, [156]

OECD (2019), An Introduction to Online Platforms and their Role in the Digital Transformation, OECD Publishing, Paris, [9]

OECD (2019), Going Digital: Shaping Policies, Improving Lives, OECD Publishing, Paris, [56]

OECD (2019), “Household spending”, Household accounts, (accessed on 5 April 2019). [113]

OECD (2018), “Consumer protection enforcement in a global digital marketplace”, OECD Digital Economy Papers, No. 266, OECD Publishing, Paris, [107]

OECD (2018), Financial Markets, Insurance and Pensions: Digitalisation and Finance, OECD Publishing, Paris, [139]

OECD (2018), “Improving online disclosures with behavioural insights”, OECD Digital Economy Papers, No. 269, OECD Publishing, Paris, [31]

OECD (2018), “Personalised pricing in the digital era”, [29]

OECD (2018), Tax Challenges Arising from Digitalisation – Interim Report 2018: Inclusive Framework on BEPS, OECD Publishing, Paris, [4]

OECD (2017), “Algorithms and collusion: Competition policy in the digital age”, OECD Publishing, Paris, [7]

OECD (2017), OECD Digital Economy Outlook 2017, OECD Publishing, Paris, [18]

OECD (2017), “Trust in peer platform markets: Consumer survey findings”, OECD Digital Economy Papers, No. 263, OECD Publishing, Paris, [32]

OECD (2016), Housing Related Expenditure of Households, [114]

OECD (2016), OECD Recommendation of the Council on Consumer Protection in E-Commerce, OECD Publishing, Paris, [30]

OECD (2016), OECD Science, Technology and Innovation Outlook 2016, OECD Publishing, Paris, [57]

OECD (2015), Data-Driven Innovation: Big Data for Growth and Well-Being, OECD Publishing, Paris, [6]

OECD (2014), “Cloud computing: The concept, impacts and the role of government policy”, OECD Digital Economy Papers, No. 240, OECD Publishing, Paris, [80]

OECD (2013), “Electronic and mobile commerce”, OECD Digital Economy Papers, No. 228, OECD Publishing, Paris, [106]

OECD (forthcoming), Enhanced Access to Data: Reconciling Risks and Benefits of Data Re-use Across Societies, OECD Publishing, Paris. [8]

OECD/Eurostat (2018), Oslo Manual 2018: Guidelines for Collecting, Reporting and Using Data on Innovation, 4th Edition, Eurostat, Luxembourg/OECD Publishing, Paris, [2]

OECD/WTO (2017), Aid for Trade at a Glance 2017: Promoting Trade, Inclusiveness and Connectivity for Sustainable Development, WTO, Geneva/OECD Publishing, Paris, [67]

OFT (2013), “Personalised pricing: Increased transparency to boost trust”, [28]

Open Music Initiative (2018), “About Open Music”, Open Music Initiative, (accessed on 19 April 2018). [72]

Orendorff, A. (2019), “The plague of e-commerce return rates and how to maintain profitability”, Shopify, (accessed on 10 April 2019). [122]

Osafo-Kwaako, P. et al. (2018), “Mobile money in emerging markets: The business case for financial inclusion”, (accessed on 11 April 2019). [153]

Osterwalder, A. and Y. Pigneur (2010), Business Model Generation: A Handbook for Visionaries, Game Changers, and Challengers, John Wiley and Sons; 1st edition. [3]

Ovans, A. (2015), “What is a Business Model?”, Harvard Business Review, (accessed on 19 April 2018). [1]

Pandolph, S. (2017), “Amazon Echo owners are making more voice purchases”, Business Insider, (accessed on 19 April 2018). [89]

Pate, T., J. Kun and B. Srinivasan (2016), “Sensor21: Earn bitcoin by collecting environmental data”, (accessed on 26 October 2018). [160]

Perez, S. (2016), “Amazon sues more sellers for buying fake reviews”, Tech Crunch, (accessed on 19 April 2018). [52]

Pew Research Centre (2016), “Online shopping and e-commerce”, (accessed on 19 April 2018). [49]

PwC (2017), “E-Commerce in China – The future is already here”, PricewaterhouseCoopers Limited, (accessed on 19 April 2018). [119]

RIAA (2018), RIAA Music Revenues Report - Mid Year 2018, (accessed on 12 April 2019). [76]

Riley, C. (2017), “Uber criticized for surge pricing after London terror attack”, CNN, (accessed on 4 June 2017). [26]

Rubin, B. (2015), “Amazon looks to improve customer reviews system with machine learning”, CNET, (accessed on 19 April 2018). [51]

Russell, J. (2018), “As Stripe backs away from crypto payments, Coinbase offers a new solution for e-commerce”, Tech Crunch, (accessed on 19 April 2018). [158]

Safaricom (2018), “Safaricom, Paypal and Transferto collaborate for international e-commerce, transfer of funds to and from M-Pesa”, (accessed on 19 April 2018). [146]

Sanitago, C., C. Gnanasambandam and S. Bhavik (2018), “From box to cloud”, (accessed on 19 April 2018). [81]

Shopify (2018), Shopify website, (accessed on 19 April 2018). [82]

Shopify (2018), “Shopify – Payment methods – Cryptocurrency”, [159]

Soo, Z. (2017), “Unmanned stores in China still working on a seamless shopper experience”, South China Morning Post, (accessed on 10 April 2019). [135]

Spotify (2018), “Spotify technology S.A. announces financial results for third quarter 2018”, Spotify, (accessed on 12 April 2019). [78]

Springer, C. (ed.) (2015), “P2P lending fraud detection: A big data approach”, Intelligence and Security Informatics. PAISI 2015. Lecture Notes in Computer Science, Vol. 9074, pp. 71-81, [55]

Statt, N. (2019), “Fake Amazon reviews draw fraud charges in groundbreaking FTC case”, The Verge, (accessed on 19 March 2018). [54]

Suri, T. and W. Jack (2016), “The long-run poverty and gender impacts of mobile money”, [151]

Tadelis, S. (2016), “Reputation and feedback systems in online platform markets”, Annual Review of Economics, Vol. 8, pp. 321-340, (accessed on 20 March 2018). [41]

Taggart, J. and K. Granville (2017), “From zombie malls to Bonobos: What America’s retail transformation looks like”, New York Times, (accessed on 19 April 2018). [125]

Taylor, T. (2019), “Alipay and WeChat Pay: The quest for a cashless China”, Tech Genix, (accessed on 11 April 2019). [143]

The Economist (2019), “Subscription”, The Economist, (accessed on 18 April 2019). [92]

The Economist (2018), “Mobile financial services are cornering the market”, The Economist, (accessed on 3 May 2018). [144]

The Economist (2016), “When life gives you lemons”, The Economist, (accessed on 19 April 2018). [77]

Uber Partner Help (2018), “Does my vehicle need to get inspected?”, (accessed on 19 April 2018). [37]

United Postal Service (2018), “UPS pulse of the online shopper survey”, (accessed on 19 April 2018). [109]

Unuvar, M. (2017), “Blockchain in food safety: The food industry gets an upgrade with blockchain”, IBM, (accessed on 19 April 2018). [121]

Ursu, R. (2016), “The power of rankings: Quantifying the effect of rankings on online consumer search and purchase decisions”, Marketing Science, Vol. 37/4, pp. 530-552, [25]

Verhoef, P., S. Neslin and B. Vroomen (2007), “Multichannel customer management: Understanding the research-shopper phenomenon”, Vol. 24/2, pp. 129-148, [104]

Vodafone (2017), “Vodafone marks 10 years of the world’s leading mobile money service,M-Pesa, (accessed on 23 October 2018). [148]

Waldron, J. (2019), “Bonobos builds brick-and-mortar ‘guideshops’ to ecommerce”, eTail, (accessed on 10 April 2019). [123]

Wall Street Journal (2014), “Zara builds its business around RFID”, (accessed on 19 April 2018). [17]

Warren, T. (2017), “Amazon’s Alexa voice deals return for this year’s Prime Day”, The Verge, (accessed on 19 April 2018). [90]

Washington D.C. District Department of Transport (2018), Hearing Record: Personal Delivery Device Act, (accessed on 2 November 2018). [103]

Wong, V. (2015), “How Taco Bell’s ordering app turns extra onions into real money”, Bloomberg, (accessed on 19 April 2018). [132]

World Bank (2017), The Global Findex Database, 2017, International Bank for Reconstruction and Development/The World Bank, [141]

Xiaohan, T. (2017), “Alibaba’s Hema supermarkets present the model for the future of retail”, ASEAN Today, (accessed on 19 April 2018). [116]

Zhang, M. (2018), “Tencent beats Amazon to launch unmanned shop, but lags other Chinese tech rivals”, South China Morning Post, (accessed on 19 April 2019). [136]


← 1. This may be part of the reason why Amazon, which uses a business model that features an online platform, acquired Whole Foods Market, a relatively expensive food retailer that focuses on high quality rather than low prices.

End of the section – Back to iLibrary publication page