Platforms: Their Structure, Benefits, and Challenges

Abstract

In this chapter, we describe platforms and their structure and how that structure differs from traditional linear value chains. We then discuss some of the key economic factors, including two-sided and multi-sided network effects, which underpin both the platform value proposition and the ability to create welfare for users. Platform and technology firms have grown to the point where their market capitalizations greatly exceed oil, gas, and financial services firms. We then explore some key governance and regulatory issues, including privacy, false information, and antitrust. We conclude with a discussion of emerging issues posed by Large Language Models such as ChatGPT, including their ability to create false information at scale and disrupt creative industries.

1 Introduction

A platform is a business based on enabling value-creating interactions between external producers and consumers.¹ The platform provides an open, participative infrastructure for these interactions and sets governance conditions for them. The platform’s overarching purpose is to match users and facilitate the exchange of goods, services, or social currency among those users in order to create value for all participants (Parker et al., 2016).

One challenge when describing platforms is that the term is used in different ways by different people and disciplines. For example, personnel who are responsible for an organization’s technology—such as a Chief Technology Officer—often describe a platform as the technology (computers, software, and communications systems) that are deployed by the organization to carry out its operations. Others might refer to a platform as a technology stack that provides reusable functionality to make product design, development, and delivery faster and cheaper (Levandowski et al., 2013). In this chapter, the term “platform” will refer to a business system that includes the functionality and architecture necessary for users to create and consume value as well as the rules of governance to promote and regulate interactions (Jacobides et al., 2019). For example, platform resources include the default insurance contracts that protect drivers and riders (in the case of Lyft and Uber) as well as hosts and guests (in the case of Airbnb). They facilitate user interactions while reducing the transaction costs, both monetary and non-monetary, of engaging in an economic activity or transaction.

2 Platforms

Platforms are widespread, and most people use them in their daily lives. Common platform uses include searching the web with Google, producing or consuming messages using Twitter, interacting with other professionals using LinkedIn, transporting through Uber or Lyft, renting a room through Airbnb, managing teams and engaging customers using Salesforce.com, purchasing goods and/or services on eBay or Amazon Marketplace, and purchasing applications on Google Play or Apple’s App and iTunes stores. While the firms mentioned have varying business models, they share the common characteristic of facilitating interactions among users. They are also relatively new companies that rely on relatively recent information and communications technology (ICT). Their platforms’ rapid adoption suggests that consumers derive significant benefit from the goods and services these platforms provide.

Despite the recent rise of large technology firms using platform business models, platforms have been present throughout human history. Medieval village marketplaces served as platforms for merchants to connect with consumers and exchange goods and services for some form of compensation. More recently, shopping malls have functioned as platforms by matching stores with consumers. Mall developers explicitly woo name brand (anchor) stores because of their ability to attract shoppers who then shop at smaller specialty stores (Yeates et al., 2001). Similarly, radio stations, television networks, and newspapers have long functioned as platforms, matching advertisers with consumers (Evans, 2009).

What is new is the growth and reach of platform firms. This occurred for many reasons, chief among them the (ICT) advances that have dramatically increased computing power, decreased the costs of storage, and increased network connectivity. In the past, transaction costs were sufficiently high that similar platforms could exist simultaneously, protected by geographic, cultural, linguistic, or other barriers. A medieval village was limited as a platform because it served only those merchants and consumers who could travel there over roads using human and animal power. Today’s digital technology has dramatically reduced the transaction costs of earlier platforms, expanding the scope and size of the markets served, sometimes quite dramatically. For example, both eBay and Craigslist greatly expanded the markets available to sellers of used (and new) items. Previously, these sellers depended on newspaper classified advertisements to reach buyers (Seamans & Zhu, 2014). As this example shows, an entire market that was once fragmented by region can now be served by one or a few firms.

Although the economics of platforms have been studied for decades, the term “platform” itself is relatively new. Previous research described the structure of these types of businesses as systems and networks (David, 1985; Katz & Shapiro, 1985; Farrell et al., 1998), with much emphasis on the concept of network effects. Network effects describe the impact of the number of users on the value generated for each user and the adoption rate of new technologies. For example, fax machines and telephones were (and are) one-sided networks that enable communication between a single type of user (e.g., sender and receiver of fax messages). As more users joined these platforms, they were able to facilitate a greater number of interactions, resulting in increased value for users Eisenmann et al. (2006) and Parker et al. (2016) describe these effects in detail.

Over approximately the past two decades, scholars have extended their analyses and formalized the understanding of two-sided markets/networks, network effects, and the impact on platform pricing decisions (Parker & Van Alstyne, 2000, 2005; Rochet & Tirole, 2003; Caillaud & Jullien, 2003). Two-sided markets are networks in which two distinct types of users (e.g., demand-side buyers and supply-side sellers) can interact with one another. Two-sided networks can add additional sides, and the literature often refers to these as multi-sided platforms (MSPs) (Hagiu & Wright, 2015). For example, LinkedIn is an MSP in that it connects individual users, recruiters, and advertisers.

Importantly, multi-sided platforms often begin as one-sided systems that can exhibit strong same-side network effects among a single type of user. Over time, the systems then often expand (i.e., open) to add additional types of users (Eisenmann et al., 2009). These additional users create cross-side network effects (i.e., between different types of users) if the value to users depends on the number of other types of users. Note that in the long run, one-sided platforms cannot provide unsubsidized free services; to sustain their operations, they need a revenue source. Even if there are strong same-side network effects, such as those often observed in social networks, these systems will still tend to open up to additional types of users, such as advertisers, to gather the necessary resources to fund ongoing operations and technology development.

The topic of pricing by platforms has been well studied by economists. The early literature laid out the conditions of market size, network effect strength, and elasticity of demand that could drop prices below those that would normally be set by profit-maximizing firms (Parker & Van Alstyne, 2005; Hagiu, 2006). This stream of literature showed that prices that might once have been viewed as predatory—that is, below some appropriate measure of marginal cost—were in fact perfectly rational. Firms can afford to give away goods and services to one type of user so long as that type of user’s participation on the platform attracts a sufficient number of paying users. Their fees can more than offset the cost of serving the free users. For example, early two-sided network literature showed that it can be profit maximizing to charge one side of a network zero or negative prices. This phenomenon, long observed in examples such as advertising-sponsored radio and television (Eisenmann et al., 2009), was becoming prevalent on the Internet in the late 1990s and early 2000s (Parker & Van Alstyne, 2005).

3 Platform Structure

As described above, a platform’s most critical function is enabling interactions among both similar and diverse users. Platforms do this by providing both the infrastructure for these interactions and the governance mechanisms to enforce rules about what users can and cannot do. This architecture is central to key aspects of platform operations, e.g., the nature and size of network effects, and whether platforms networks encourage users to affiliate with multiple platforms at once (multi-homing) (Choi, 2010). Similarly, understanding a platform’s governance, the extent to which the platform is open or closed, and the way(s) in which the platform can be monetized are all critical to understanding the platform’s underlying economics.

Before describing platform businesses, we first describe what business structure is not a platform. For example, a standard business arrangement to produce and deliver a product or service is often described as a linear supply chain, i.e., an arrangement where value accumulates from one stage of production or distribution to another. Figure 1 shows a stylized example of a linear automotive supply chain (Hayes et al., 1988). On the left are the upstream sources of raw materials such as metals and plastics. In the middle, raw materials are fabricated by suppliers into components such as paint, tires, and seats. These components are then combined in an assembly plant to form a complete automobile. At the end, the automobile is sold to a customer.

Fig. 1

Linear supply chain: styled automotive example

In this figure, value accumulates from left to right. Supply-chain partners are compensated from right to left for their value add. From a customer’s point of view, the supply chain is relatively invisible and largely irrelevant (Zeng, 2015). Customers care about what the supply chain can deliver in terms of a finished product or service delivered at retail. There are relatively minimal network effects, if any, in a standard supply chain system; one consumer’s purchase of an automobile does not significantly change the value that another consumer derives from the product. Note that electric vehicles are an exception, as consumers do care whether others are active on their network (Anderson et al., 2022a). But the linear supply chain discussed above is not a platform.

A true platform’s structure differs from that of linear supply chains in that it has a triangular structure that facilitates user interactions (Eisenmann et al., 2011). We begin with a description of one-sided platforms with a single type of user. We then describe two-sided platforms with both consumer-side and supply-side users. For illustration purposes, readers can think of consumer-side users as the buyers on the platform; supply-side users, by contrast, are the platform’s sellers. First, consider a one-sided platform where users interact with other similar users. When platforms first launch, they often operate as a one-sided platform. In this way, platforms can work out the value proposition to one type of user before the platform opens up to additional types of users. For example, when LinkedIn first began operations, it only facilitated interactions between professionals who wished to connect with one another. Figure 2 shows the one-sided network platform structure graphically.

Fig. 2

One-sided network platform structure

Telephone systems are examples of a one-sided network. The technology facilitates interactions between users; the interaction takes the form of a telephone call. At any given time, a user can be either the call originator or receiver. If there are only a few (or no) other users, then the system has minimal value. However, once adoption grows, the system can provide significant value to its many users.

In contrast, the structure of a two-sided network platform is triangular, as shown below in Fig. 3.

Fig. 3

Two-sided network platform structure

Two-sided platforms allow different types of users (Side “A” and Side “B” above) to use platform resources to directly transact with one another to exchange value (Hagiu & Wright, 2015). Uber provides a well-known example of this two-sided network platform structure. The Uber ride-sharing system (the platform) matches drivers and their cars (Side “A”) with riders (Side “B”) and then allows a ride to be exchanged (the “Direct Exchange of Value”). The platform provides significant functionality; it tracks the location of drivers and riders, matches them to one another, transmits pricing information, provides payment services, allows for bidirectional ratings, and more.

Multi-sided platforms facilitate interactions among even more types of users (Sides “A,” “B,” and “C”). For example, Uber has expanded to food delivery (Uber Eats) by adding restaurant users (Side “C”) who want access to Uber’s drivers (Trabucchi & Buganza, 2020). The new structure is illustrated by Fig. 4.

Fig. 4

Multi-sided network platform structure

Typically, a two-sided network has both demand-side users (e.g., consumers seeking to buy a good or service) and supply-side users (e.g., sellers seeking to sell a good or service). Consider as an example Sony’s PlayStation game console. Like many gaming systems, PlayStation functions as a platform that allows game players (demand-side users) to use software (games) provided by game developers (supply-side users). Also, Sony sets the platform’s rules. For example, it determines which games are allowed on the platform (Eisenmann et al., 2006). The PlayStation system facilitates both same-side and cross-side interactions. Gamers derive value by playing games, but they can also interact with other players (same-side interaction) in multi-player game settings, deriving additional value. In addition, PlayStation gamers value access to the wide variety of content (cross-side interaction) provided by authorized game developers.

One key source of platform value is the network effect. Here, value depends on the number of different-type users. Because this concept is so central to understanding platforms, we discuss network effects in detail below.

4 Network Effects and Value

Network effects exist when the value users derive from a platform depends on the number of platform users (i.e., the size of a user base). For an example of a one-sided network effect, consider a social network such as Instagram. It is more valuable to users when members of their family are also on the network. For a two-sided network effect example, consider how a merchant’s willingness to accept a given credit card (e.g., American Express) depends on the number of consumers who wish to pay with such a card. The greater the number of consumers wishing to pay with that card, the greater the value to the merchant in accepting it. Similarly, a consumer’s willingness to carry a given credit card depends on the number of merchants willing to accept it.

One-sided platforms manage interactions among just one type of user. Importantly, the value that users receive from the presence of other users and their interactions can be either positive or negative, as shown in Fig. 5. For this reason, platforms work to reduce or even eliminate negative interactions, as we discuss further in the section on governance.

Fig. 5

One-sided networks facilitate one main interaction. Positive (+) same-side: telephone and fax machine users benefit from being able to call one another and exchange information. Negative (−) same-side: receiving a harassing phone call is unwelcome and reduces a user’s value

By contrast, two-sided platforms facilitate two kinds of interaction between users: same-side user interactions and cross-side user interactions. From these interactions, network effect value flows. More formally, a two-sided market is one in which (1) two sets of agents interact through an intermediary or platform and (2) the decisions of each set of agents affect the outcomes of the other set of agents, typically through an externality (Rysman, 2009). In a two-sided platform, there are four potential network effects to consider, as shown in Fig. 6.

Fig. 6

Two-sided networks have four interactions. Positive (+) same-side: player-to-player contact in Xbox games, end user sharing of PDF files. Positive (+) cross-side: merchants and consumers for credit cards, application developers and end-users in Android or Apple iOS. Negative (−) same-side: competing suppliers on procurement platforms, harassment from other users on social media platforms, other diners who book your restaurant table. Negative (−) cross-side: advertising clutter to consumers. Source: T. G. Eisenmann, Geoffrey G. Parker, and Marshall Van Alstyne, “Strategies for Two-Sided Markets.” Harvard Business Review 84.10 (2006)

Let us return to the case of the PlayStation video game system. The platform sponsor is the console producer—Sony—while the two sets of users are consumers (i.e., game players) and video game developers. Neither consumers nor game developers are likely to find significant value in the PlayStation console unless the other party is present and active. More players attract more game developers, and more and better games attract more players. These are positive cross-side network effects. The value that users place on other users’ participation is exactly the network effect discussed above.

A highly stylized and simplified representation of the value that platform users gain from the system is shown in the equations below.² Note that the value users obtain from using the platform is not the same as the amount that the users pay for the platform. In some cases, users pay nothing, yet they can derive significant value from affiliating with platforms. For example, the users of the Google, Bing, and Baidu search engines are not directly charged. Instead, the platforms charge advertisers for those users’ attention.

4.1 User Value for One-Sided and Two-Sided Networks (with User Types 1 and 2)

User value in a one-sided network

$$ \mathsf{User}\ \mathsf{Value}=\mathit{\mathsf{V}}+{\mathit{\mathsf{e}}}^{\ast }\ \mathit{\mathsf{N}} $$

(1)

User type 1 value in a two-sided network

$$ \mathsf{Type}\ \mathsf{1}\ \mathsf{User}\ \mathsf{Value}={\mathit{\mathsf{V}}}_{\mathsf{1}}+{{\mathit{\mathsf{e}}}_{\mathsf{1}}}^{\ast }\ {\mathit{\mathsf{N}}}_{\mathsf{1}}+{{\mathit{\mathsf{e}}}_{\mathsf{2}\mathsf{1}}}^{\ast }\ {\mathit{\mathsf{N}}}_{\mathsf{2}} $$

(2)

User type 2 value in a two-sided network

$$ \mathsf{Type}\ \mathsf{2}\ \mathsf{User}\ \mathsf{Value}={\mathit{\mathsf{V}}}_{\mathsf{2}}+{{\mathit{\mathsf{e}}}_{\mathsf{2}}}^{\ast }\ {\mathit{\mathsf{N}}}_{\mathsf{2}}+{{\mathit{\mathsf{e}}}_{\mathsf{1}\mathsf{2}}}^{\ast }\ {\mathit{\mathsf{N}}}_{\mathsf{1}} $$

(3)

Equation (1) shows the value that a single type of user receives from a one-sided network. V is the stand-alone value for the platform that a single user enjoys. N represents the number of other users, and e represents the incremental value that each additional user contributes to a single user’s value from the platform. Thus, as N increases by one, a user’s value from the platform increases by e, which can be interpreted as the marginal network effect value from one additional user.

Equation (2) presents a user’s value in a two-sided network. V₁ is the stand-alone value that user type 1 (such as a consumer) derives from the platform (i.e., the value to user type 1 from her use of the platform, in the absence of any other users). The second term e₁* N₁ captures the same-side network effect, where parameter e₁ measures the incremental value that an additional Type 1 user provides to other Type 1 users, and N₁ is the number of other Type 1 users. The third term e₂₁* N₂ captures the cross-side network effect, where parameter e₂₁ measures the extra value that user Type 1 derives from the addition of another Type 2 user (such as a seller on a marketplace) and N₂ is the number of Type 2 users. Equation (3) parallels Equation (2) and describes the value that a Type 2 user derives from the core platform (V₂), the same-side network effect from additional Type 2 users (e₂* N₂), and the cross-side network effect from Type 1 users (e₁₂ * N₁).

When there are no network effects, parameters e₁, e₂, e₂₁, and e₁₂ equal zero, and users benefit only from the stand-alone value of the platform. Products such as stoves or hair dryers fit this description.

To summarize, the value proposition to users can be thought of as occurring in three main categories: stand-alone value (in the absence of other users); same-side network effect value, which depends on similar users; and cross-side network value, which depends on the interactions of different types of users (Anderson et al., 2022b).

5 Platform Openness

A critical decision that platforms must make is how open to be. Platforms can create significant value by coordinating the activities of external “ecosystem” partners, who in turn attach to the platform to reach customers and produce products or services (Jacobides et al., 2018). Examples include computer operating systems that are extended by both software developers and social media platforms (such as YouTube and Twitter) that depend on their users for content creation. Enterprise software systems such as those offered by SAP and Salesforce.com have significant functionality by themselves, but they also depend on external partners who can extend the platform’s functionality.

One challenge platforms face is determining how to provide access to users who can create value for other users while also excluding users who either provide poor-quality products and services or misbehave and destroy value for other users. An early example comes from the video game crash of 1983 (Aoyama & Izushi, 2003). During this time, Atari could not control who published games on its system; as a result, many inferior games were produced for Atari’s gaming console. The result: because customers could not easily distinguish good-quality games from bad, many fled the market—a classic “Market for Lemons” failure (Akerlof, 1970). The next company to lead the videogame market, Nintendo, learned from Atari’s mistake. To ensure quality, Nintendo excluded certain developers from publishing games for its system.

Controlling access to a platform is a key way that a platform’s owner has the control points needed to monetize participation, should it choose to. Platforms that are too open have had difficulty charging for the goods and services transacted across them (Parker & Van Alstyne, 2018). Platforms can open only slightly, allowing external supply-side partners to connect to their systems, by providing application programming interfaces (APIs) (Parker & Van Alstyne, 2009). More generally, a platform is open to the extent that (1) no restrictions are placed on participation in its development, commercialization, or use or (2) any restrictions—for example, requirements to conform with technical standards or pay licensing fees—are reasonable, non-discriminatory, and applied uniformly to all potential platform participants (Eisenmann et al., 2009).

5.1 Access to Demand-Side Use (as a Consumer)

Most platforms are open to any end user who wishes to connect. However, if payment is required for access, then a system will be designed to ensure that only paying users can use the platform. Access to the demand side can be tightly restricted. For example, when Apple’s iPhone was first released in the US market, the device was compatible only with the AT&T network. Users who wanted iPhones but were not already AT&T customers had to incur significant costs, first acquiring the device and then signing up with AT&T.

5.2 Access to Extend Platform (Supply Side)

Supply-side users including developers may have open access to platforms such as Linux, Microsoft Windows, and Apple macOS. None of these systems impose restrictions on the ability of demand-side users to load new applications. In contrast, Sony limits the games that can be published on its PlayStation game consoles. Sony does this primarily so it can screen games for quality. As detailed above, the “crash of 1983” shows that firms can suffer significant losses if they do not manage the quality of all elements of their platforms or systems. In the case of business software, systems such as Salesforce also limit access to only those developers who can pass a quality screen (Sawhney & Nambisan, 2007). Similarly, Apple restricts access to the supply side for its smartphone operating system, iOS (Hagiu, 2014).

5.3 Access to User Provision (e.g., Android and Apple Devices)

Platforms must also often provide access technology to their users. This leaves the platform owner with a decision: whether to provide the technology itself or whether to open the ability to supply technology to third parties. In the game console examples described above, the major suppliers—Microsoft (Xbox), Nintendo (Wii), and Sony (PlayStation)—are the exclusive console providers, and they do not allow third-party firms to supply hardware. However, in other settings, firms do open this role to third parties. Even within a single industry, firms can make different decisions. For example, in smartphones, while Apple restricts supply-side access, Google allows any hardware manufacturer to use Android Open Source Project software on its phones (Gandhewar & Sheikh, 2010).

5.4 Access to Change Technology/Contracts (Decision-Making)

Another critical decision facing platform sponsors is the degree to which they open their platforms to multiple decision-makers on issues regarding, for example, the nature and form of platform content. This decision can drive who gets access to the platform, what technology the platform will deploy, and if (or even how) the platform will capture value.

In some cases, these companies have opted to reserve all platform decision rights for themselves. This has been the case with Apple (iOS), Microsoft (Xbox), Nintendo (Wii), Salesforce, SAP, and Sony (PlayStation). However, there are also cases where decisions are made by multiple actors, whether a small number of partners or (as is the case with open-source software) a relatively large number of decision-makers. For example, both the Mastercard and Visa credit card networks are owned by member banks who control their decisions (Akers et al., 2005). Similarly, open-source systems such as Apache and Linux are inherently open at the user and developer level. And at the decision-making level, these open-source systems allow any of their users to propose changes to the core technology. Further, open-source governing bodies such as the Apache Software Foundation have multiple membership levels determining who gets to make changes (Lerner & Tirole, 2002). Membership levels are typically allocated to those who make greater contributions (O’Mahony & Ferraro, 2007). When a disagreement over technology direction cannot be resolved, a party is free to “fork” the software to make a version that incorporates their vision. Other factors may encourage forking, too, as user-developers may have conflicting interests as to the technology’s evolution (Lerner & Tirole, 2002).

6 Platform Governance

Platform governance encompasses the set of rules and decisions platforms make to determine who gets to participate in an ecosystem, how value gets divided, and which mechanisms are used to settle disputes, whether among users or between the platform and its users. Platforms must determine how to provide access to users and encourage activities that generate value while also excluding certain users or specific actions that do not contribute to value creation. Failures in platform governance can both prevent value creation and destroy existing value (Boudreau & Hagiu, 2009).

In addition, online platforms that attract advertisers with their user data must confront additional privacy concerns. To make this data useful, online platforms want their users to share as much as possible about themselves. However, this can increase privacy worries among the users themselves, such as whether their information is being sold, stolen, or otherwise misused. Therefore, online platforms must strike a balance between sharing user data with advertisers and implementing appropriate privacy controls and data security. In addition, governments have passed laws designed to protect consumer privacy, by restricting both the use of data and online tracking techniques used by websites.³ Online platforms must comply with such regulations, even in the face of research that suggests even moderate privacy regulation can reduce the effectiveness of online advertising (Goldfarb & Tucker, 2011).

One governance area that has gained attention recently is the challenge of managing the dramatic increase of false information online (Vosoughi et al., 2018; see also the chapter in this volume on content moderation by Prem and Krenn). This includes the potentially deadly consequences of false information such as attacks on the Rohingya in Myanmar (Fink, 2018). Election integrity is another area under attack across the globe as numerous actors, including nation states, seek to influence electoral outcomes (Henschke et al., 2020). The explosion of fake news online has led some platforms to invest in capabilities that can identify and remove false information and problematic content such as fake user profiles, offensive language, and hateful or discriminatory comments (Aral, 2020). These platforms must also decide whether to impose disciplinary actions on users who fail to comply with such guidelines, which can include temporary or permanent removal from the platform (Conger & Isaac, 2021). As the more recent reversal of the ban on Donald Trump from Twitter shows, this is a rapidly changing area.⁴ Given the implications for social stability, there is likely to be significant policy innovation as regulators across the world attempt to reign in the harmful effects of misinformation. This is a particularly thorny issue in the United States because the Constitution’s First Amendment prohibits the government from abridging the freedom of speech and of the press. One novel solution (proposed by Van Alstyne, 2021) is to focus on the amplification of speech instead of the original speech itself. The issue of amplification has been identified as critical (Syed, 2017). In Van Alstyne’s scheme, speakers might be required to warrant that their speech is true if they wish to have it disseminated beyond the original publication.

7 Platform Growth and Power

Platforms emerged as an important type of organizational form as early as 2006 (Eisenmann et al., 2006) when the conditions for winner-take-all or winner-take-most markets began to become clear. One important question is when firms can overcome the entry barriers. Platforms can enjoy these barriers to competition due to network effects and switching costs. But platforms can also overcome these barriers through a strategy known as platform envelopment, in which one platform provider can enter another’s market by bundling its own platform’s functionality with that of the target, leveraging shared user relationships and common components (Eisenmann et al., 2011). This strategy is particularly effective for a platform that has overlapping user bases with a rival. The platform can then bundle the rival’s functionality into its own offering.

Over the past 25 years, platforms have taken on a larger presence in the global economy as compared to energy, pharmaceutical, and manufacturing firms that once dominated the top firms by market capitalization.⁵ Today, the top seven publicly traded firms (as of 2020) in three sectors are shown in Fig. 7 with market capitalizations from March 2023. Note the prominence of Apple, Microsoft, Google, and Amazon.

Fig. 7

Top 7 publicly traded companies by market cap by sector. Source: Author using data from Yahoo Finance (Accessed March- 30, 2023)

One important platform feature is the ability to enter the markets of traditional linear value-chain (pipeline) firms. One key method of entry is the aforementioned platform envelopment, the process by which a platform leverages overlapping user bases to enter new markets (Eisenmann et al., 2011). Because of the nature of the technology infrastructure—which tends to be modular so that ecosystem partners can connect—platforms find it relatively easy to add functionality and put forth a compelling value proposition. Product and services firms that have only stand-alone value propositions can find it difficult to match a platform offering. As shown by Fig. 8, stand-alone offerings can be threatened by the entry of a platform such as Google Android or Apple iOS.

Fig. 8

Stand-alone firms are vulnerable to entry by platforms

The rapid growth of platforms has also become a central concern for regulators. Partly, this is because of the potential for market dominance, especially by the largest technology companies (Shapiro, 2019). The pricing and business models typically used by platforms can make it difficult to apply existing regulations. For example, many online platforms offer goods and services for free. This renders meaningless the significant and non-transitory increase in price (SSNIP) test to identify the relevant market (Hesse, 2007). Theories of harm include excessive prices, inferior quality, reduced incentives for innovation, predatory pricing, and self-preferencing (Parker et al., 2020). What’s more, the merger and acquisition behavior of Google, Apple, Facebook, Amazon, and Microsoft (GAFAM)⁶ can add complementary functionality in the core business, add new functionality in the vertical value chain, merge with substitute products and services such that competition might be reduced, and seek to acquire human capital (Parker et al., 2021). The regulatory concerns that emerge from mergers and acquisitions focus essentially on market foreclosure, competitive bottlenecks, and distortion of upstream competition. Despite the potential for harm, very few of the 855 GAFAM mergers completed during the years 1988 to 2020 were investigated, suggesting that regulators did not yet have an effective way to control the potentially negative effects of these deals.

The European Union has taken significant steps toward understanding the issues, and that has included promulgating regulations such as the Digital Markets Act (DMA) and Digital Services Act (DSA).^7,8 The DMA and DSA have a particular focus on “gatekeepers” or “Very Large Online Platforms” which, for consumers, are defined as platforms with over 45 million users as of 2023.^9,10 One key question is whether such regulation will be enforced ex ante through a series of obligations that platforms must adhere to or whether it will be enforced ex post, should a platform be found in violation. The ex post regulatory framework has come under criticism for taking too long, being too expensive, and having seemingly random outcomes (Parker et al., 2020). Instead, the DMA takes a different approach; it lays out a series of over 20 ex ante directives designed to curb potential abuse.¹¹ A panel of economic experts commissioned by the European Union Joint Research Centre (The Cabral panel) analyzed the proposed regulations, and while it found that most were well grounded in economic theory, it still proposed that some prohibited activity should instead be “gray.” In these cases, the practice would be prohibited, but arguments could still be made by large platforms for why the practice should be allowed (Cabral et al., 2021).

8 Conclusions

As can be seen from the growth of large technology firms and the resulting regulatory scrutiny, there is great concern over the impact such firms are having on both the economy and the global population. Such concerns are not new; during the industrial revolution, the growth of manufacturing, transportation, and financial institutions drew similar attention. But the relatively recent growth and public awareness of accessible Large Language Models (LLM) such as ChatGPT (Chat Generative Pre-trained Transformer) show that the platform space remains as dynamic as ever. The pace of change in LLMs has been nothing short of extraordinary, with capabilities seeming to grow by the week. However, a large number of potentially negative uses for the technology have already been identified, and more are coming every day.¹² These include AI taking academic and professional exams such as the US Bar Exam for legal certification; AI authoring academic articles, student exams, and papers; and deep video fakes. There is also the potential for significant disruption to industries such as music and video. These disruptions could change the economics of these industries in ways that transform the existing competitive landscape (Gupta & Parker, 2023).

Discussion Questions for Students and Their Teachers

• 1. Consider sectors that have yet to face significant competition from platform firms. Is it only a matter of time or are there structural reasons that protect those sectors from platform entry?

• 2. Have platforms become a threat to innovation and competition? If so, are interventions such as the European Union’s Digital Markets Act and Digital Services Act a step in the right direction, or do they go too far?

• 3. What are the implications for regions of the world without significant platform firms? For example, Europe is home to only a few platform firms. Should governments seek to invest in emerging technologies such as Large Language Models, or should they leave it to private enterprise?

• 4. How are individuals impacted by large technology platform firms? Consider access to information and useful services as well as the impact on the social fabric.

Learning Resources for Students

Platform Structure

• 1. Amrit Tiwana, 2013. Platform ecosystems: Aligning architecture, governance, and strategy. Newnes.

  This work lays out the structural elements of platforms and how they interoperate.

• 2. Geoffrey Parker, Marshall Van Alstyne, Sangeet Choudary, 2016. Platform revolution: How networked markets are transforming the economy and how to make them work for you. WW Norton & Company.

  This work is a comprehensive treatment of platforms from their economics, structure, launch conditions, business models, governance structures, and regulatory concerns.

Network Effects

• 1. Rochet, J.C. and Tirole, J., 2003. Platform competition in two-sided markets. Journal of the european economic association, 1(4), pp. 990-1029.

  This is one of the two seminal works that describe the theory of two-sided network effects from the lens of the credit card industry.

• 2. Parker, G.G. and Van Alstyne, M.W., 2005. Two-sided network effects: A theory of information product design. Management science, 51(10), pp. 1494-1504.

  This is the second of two seminal works describing the theory of two-sided network effects from the lens of multiple industries, including computer operating systems and software complements.

Openness

• 1. Joel West, “How Open Is Open Enough? Melding Proprietary and Open Source Platform Strategies,” Research Policy 32, no. 7 (2003): 1259–85.

  This work examines the decisions that Apple, IBM, and Sun made in computer operating system markets to examine how openness impacts a firm’s ability to capture value (appropriability) and user adoption.

• 2. Eisenmann, T.R., Parker, G. and Van Alstyne, M., (2009). “Opening platforms: How, when and why.” In Platforms, markets and innovation, 6, pp. 131–162.

  This work defines platforms, lays out the different structures of sponsorship, and makes predictions of openness over a platform’s life cycle.

• 3. Kevin Boudreau, “Open Platform Strategies and Innovation: Granting Access Versus Devolving Control,” Management Science 56, no. 10 (2010): 1849–72.

  This work explores the decisions that mobile platform sponsors made with respect to allowing access to the platform and lays out subsequent adoption statistics.

Governance

• 1. Andrei Shleifer and Robert W. Vishny, “A Survey of Corporate Governance,” Journal of Finance 52, no. 2 (1997): 737–83, esp. 737.

  This paper lays out the broad principles of governance at a firm level.

• 2. Geoffrey G. Parker and Marshall Van Alstyne, “Innovation, Openness and Platform Control,” 2018, Management Science.

  This paper develops a model of sequential innovation that shows how ecosystem participants can be collectively better off if a platform enforces rules that require ecosystem partners to share intellectual property with other partners after a closed proprietary period.

Platform Growth and Power

• 1. Cusumano, M.A., Gawer, A. and Yoffie, D.B., 2021. Can self-regulation save digital platforms?. Industrial and Corporate Change, 30(5), pp. 1259–1285.

  This work explores which areas where platforms can be trusted to regulate themselves and where the threat of government intervention might be necessary to improve social welfare.

• 2. Cennamo, C., Kretschmer, T., Constantinides, P., Alaimo, C. and Santaló, J., 2023. Digital platforms regulation: An innovation-centric view of the EU’s Digital Markets Act. Journal of European Competition Law & Practice, 14(1), pp. 44–51.

  This work provides an additional analysis of the European Union’s Digital Markets Act with a special focus on innovation.