In business school, we teach that platforms create value through coring. Coring is the steps a digital platform takes to make sure that interactions between different user groups go well, and that as a consequence they wish to return to the platform and use it again. As such, platforms typically take on a governance role and actively manage interactions between different user groups. This article discusses two implications of this business school concept for competition economics: Market definition and the recent Supreme Court Decision for Amex.
By Catherine Tucker[1]
I. WHAT IS CORING?
Platforms enable multiple distinct groups of users to interact with one another, and platform economists refer to the practice of actively managing these interactions to ensure they go well as “coring.”[2]
A platform acts as a “core” for these interactions by adopting technology. policies, and procedures that facilitate the interaction taking place, build trust in the interactions, and provide incentives for interactions to stay on the platform. This often requires transaction platforms to establish governance mechanisms based on observable transactions that ensure successful interactions.
With interactions that are observable, transaction platform operators can track the quality of the interactions and user satisfaction based on the nature of the transactions (e.g., guaranteeing secure transfer of payment information or identifying accurate matches for user searches). Transaction platforms can continually adjust their policies so that all sides will continue to use the platform to transact through high-quality interactions rather than turn to an alternative.
Coring efforts may increase quality, promote trust, or increase security, and may restrict the ways in which one or both sides can use the platform. Because multi-sided platforms must balance the preferences and actions of all user sides when designing their policies and processes, coring may result in policies restrictive for one side in isolation, but that are pro-competitive and improve the platform as a whole and the quality and value of the interaction for all user sides.
Examples of coring include the creation of reliable rating systems that give insight into the often-unobservable quality of platform participants, such as when device users who download an app from an app store leave a rating or a review. This transparency helps platforms to maintain their reputations and brands, even in the face of some lower quality platform participants. Other examples of coring are constraints on who can use the platform, and policing behavior by some users that may spill over negatively on other users.
Coring can also include limitations on what kinds of products are available, for example limitations on what kind of firms are allowed to advertise at trade shows, if these products lead to dilution of the focus of the trade show. These are all examples of platform competition requiring restriction to better serve the needs of users and respond to potential competitors and new entrants that would exploit negative spillovers and offer better alternatives. The transaction platform creates value from interactions between sides, and requires coring by the operators to enhance overall user welfare.
For example, Uber riders’ willingness to use Uber depends on their trust that rides will be safe and punctual. Absent this trust, riders will be less willing to use Uber, which decreases the value of Uber to drivers, and the overall value of Uber’s platform in the long term. A coring activity that builds user trust in the quality of interactions is removing users who undermine trust in the quality of interactions on the platform. For example, Uber may remove drivers with bad reviews, as these drivers undermine rider trust in the safety and punctuality of rides.
In terms of the underlying economics, usually coring activities by platforms are aimed at reducing search and transaction costs.[3] Search costs are incurred by users to identify someone on the other side of the platform with whom a beneficial match is possible. Transaction costs are incurred by users after they have found each other and started to interact.
The reduction in search and transaction costs helps unlock the value inherent in network effects that characterize platforms by making it easier to find the appropriate match and cheaper to execute the resulting interaction as the number of users on the platform increase.
For instance, AirBnB, through its integrated search technology, makes it easier for a guest to find the right accommodation at the right time and read reviews to ensure that the place they choose is likely to be a good fit. AirBnB also reduces transaction costs for hosts by providing a secure environment for consumers to pay hosts, and helping settle any disputes that arise between hosts and guests during the stay. The more hosts available on AirBnB, the higher the value that AirBnB’s technology, payments and review system creates for guests as they will be able to find better options and transact with them with lower costs.
Effective coring, designed to enhance the attractiveness of the platform for both sides, is also critical to the platform’s ongoing viability and innovation. A platform profits from its ability to capture some of the value that its facilitation technology creates. The payments made to the platform, either by one or both sides must be used to fund the platform and to continually innovate to compete with other platforms. As such, a platform must encourage the use of its platform by both customer groups to facilitate valuable interactions from which to capture a portion of the surplus. Coring strategies are therefore central to any antitrust analysis of platform market definition, market power, and conduct.
Coring needs evolve as platforms compete, learn about what works, observe their competitors, and monitor evolving interactions on their platform. Continuous coring is particularly critical when a platform is facing competition. While product improvements may occur with or without competition, competition increases the speed with which platforms must respond through such efforts to attract and retain users on both sides in response to competitive pressure.
For example, Uber has increased its coring efforts over time, to build and maintain flexible driving supply capacity and a healthy rider base. In response to issues surrounding rider safety, Uber implemented increased driver screening protocols such as recurring background checks in 2017, adding onto existing driver screening requirements such as completion of a city knowledge test and a quality inspection of cars. In 2018, Uber introduced a safety center for riders, a dashboard with safety tools such as a button to share details and location with others and a panic button to call emergency services, giving customers more tools to ensure their personal safety.
Coring efforts were not limited to the rider side and were also more directly related to competition from other ride-share services. On the driver side, Uber has also introduced new policies over time to attract and retain drivers. In 2014, Uber provided new commercial insurance to close the insurance gap when drivers were on the road without passengers. In 2019, Uber created new driver benefits to attract and maintain talent, such as a new driver lease program and a driver-appreciation bonus coinciding with its IPO.
The trust by buyers and sellers alike that the other party will fulfill its side of the bargain will erode absent corrections and policies from the platform operator to maintain trust and ensure high quality interactions. Specifically, buyers must trust that products will have the advertised characteristics and will not be associated with costs beyond those advertised. Similarly, sellers must trust that payment will be received as agreed for the products they sell and that their reputation will not be damaged by poor behavior by others or systemic failures by the platform. Finally, both buyers and sellers must trust that the interaction will not result in unwanted follow-on transactions or experiences and that the hardware through which they are completing the transaction will not be damaged.
Diminished trust will erode the number of interactions on the platform as well as the number of sellers and buyers. This will in turn erode indirect network effects and leave the platform at a disadvantage relative to other, better-managed platform competitors.
II. HOW DOES THIS BUSINESS SCHOOL CONCEPT OF CORING RELATE TO THE CONCEPT OF TRANSACTION PLATFORMS?
The Supreme Court decision in American Express.[4] has introduced the terminology of a transaction platform into the broader debate about the application of antitrust principles to two-sided platform markets. As defined in the academic article cited by the Supreme Court in its decision, transaction platforms are platforms where the interactions (or “transactions”) between both sides of the platform are observable and the transaction itself is also immediately observable. As a result, the platform cannot serve one side without simultaneously serving the other, and the interaction can be facilitated and tailored by the platform.
In contrast, the non-observability of interactions in other contexts severely limits and may eliminate a platform’s ability to reduce search and transaction costs in a way that can unlock the same level of value for the platform users. For example, one may argue that a newspaper could be considered a two-sided platform because it connects advertisers with readers and exhibits network effects since advertisers benefit from a greater number of users and users from a greater number of advertisers (assuming advertisers provide information).
It would not, however, be by this definition a transaction platform because although it brings together readers and advertisers, the interaction between readers and advertisers is not observable at the transaction level, nor something the newspaper manages and controls directly. Traditional newspapers cannot tailor advertisements to specific readers and cannot observe whether, for example, a reader opens or reviews an advertising insert or disposes of it. As such, in contrast to operators of two-sided transaction platforms, newspapers do not actively manage the interactions of advertisers and readers and therefore are not primarily focused on managing or optimizing the value of the underlying network effects of the system.
Simply put therefore, a transaction platform is one where the platform conducts serious “coring” activity to facilitate interactions between two groups of users. A non-transaction platform is one where the platform does not attempt to take serious “coring” actions and facilitate interactions between different user groups.
III. IMPLICATIONS FOR MARKET DEFINITION
Market definition is a tool economists use in antitrust cases to help courts evaluate a firm’s conduct. By defining a market and understanding what alternatives exist for customers, it is possible to then analyze the likely positive or negative effects of a firm’s conduct on consumers.
The key challenge this author sees for market definition in digital platforms is the conflation of a platform’s coring actions with a market.
To illustrate this it is useful to take a non-digital example. For example, when I teach platforms, I often refer to the example of an Auckland fish market that brings together fish buyers and fish sellers as a non-digital platform. This encourages some useful discussion of what really is a platform. We talk about how non-digital platforms possess many of the same features as digital platforms, but are often unexciting because they are not easy to scale.
We then turn to talk about coring. Often students have the picture of a fish market as being one where shoppers wander around piles of fish on stands and choose between them. Such an image suggests very little coring activity. This preconception, though, could not be more wrong. The Auckland fish market actually engages in a great deal of coring activities. In particular, it runs an electronic “Dutch auction.” In this Dutch auction, a large bidding clock displays a description of each lot on offer. The auctioneer started the clock at about $2 above the expected price per kilo. The price drops every 10 seconds – until a buyer enters a bid on their keypad. At the point the buyer’s name appears above the clock. In other words, this is an example of a non-digital platform where the platform takes its job of coring seriously, and reduces search and transaction costs for participants.
In this case the product market should be defined as the facilitation of interactions between fish buyers and fish sellers. To see why this makes sense, think about the real competitive constraints on the fish market – it is the potential that fish sellers and fish buyers could find other ways of interacting.
However, what I see happening again and again in multiple antitrust cases, is that the market is defined as something analogous to “the provision of Dutch auction services at the Auckland fish market,” or, even more specifically, “the provision of digital bidding clocks at the Auckland fish market.” In other words, the coring actions of the platforms are conflated with a market definition. This has two problematic implications. First, simply because of the nature of platforms, this means that using an incorrectly narrow market definition will almost by definition produce a conclusion of monopoly power. Second, because the market definition is incorrect, when this wrong market definition is used to evaluate conduct, that evaluation will almost certain miss out on analyzing the benefits that this coring action brings to users of this platform in terms of reducing underlying search and transaction costs.
Therefore, this article will end with a plea to courts, which is to understand that the business of platforms is to bring together different groups of users together and make their subsequent interactions go well through coring. The coring actions that the platform takes, however, are not a market, nor an after-market, nor any kind of market. Instead, they should simply be viewed as the actions the platform takes which aim to make interactions go well.
[1] Catherine Tucker is the Sloan Distinguished Professor of Management Science at MIT Sloan School of Management. She has consulted for many technology companies – please see https://mitmgmtfaculty.mit.edu/cetucker/disclosure/.
[2] Gawer, Annabelle & Michael A. Cusumano. ”Platform leaders.” MIT Sloan management review (2015): 68-75.
[3] Hagiu, Andrei, “Strategic decisions for multisided platforms,” MIT Sloan Management Review, Vol. 55 No. 2, Winter 2014.
[4] Ohio et al v. American Express Co. et al, Supreme Court of the United States, Docket No. 16-1454, decided June 25, 2018.
