Sign In

Communications of the ACM

Economic and business dimensions

Fintech Platforms and Strategy

FinTech Platforms and Strategy, illustration

Credit: Anton Khrupin

Ten years from now, will Google and Amazon play more of a role in managing your investment portfolio than Fidelity? Finance has traditionally been about trust. We trust banks to hold our money and give it back to us when we want it; and we trust brokerage firms to buy the securities we want at market prices and to debit and credit our accounts accordingly. Because trust is so important, we have historically held banks and financial firms to much higher standards of compliance and control than other businesses. Financial institutions are required to follow well-defined processes with oversight and failsafe plans aimed at minimizing risk and maximizing public trust. These processes have traditionally involved humans, even as they have been increasingly augmented with technology over time.

But the last 20 years have seen the emergence of a fundamentally new and different phenomenon enabled by Internet connectivity and access: the "platform" business. In platform businesses, many complex processes, including compliance and checks-and-balance procedures, are performed securely by machines. Along with this platform phenomenon, newer generations of humans have co-evolved to be more comfortable trusting machines to make or support decisions. It would have seemed unthinkable 10 years ago to imagine trusting autonomous driving vehicles to take over the wheel, or trust robots to perform surgery on us. But we are increasingly doing just that in a growing number of real-world areas. Broadly, we seem comfortable trusting machine-made decisions in domains in which the risks are acceptable: mistakes are relatively infrequent and their consequences do not exceed a reasonable tolerance threshold.1

The question is, will future investors trust FinTech platforms to the degree that previous generations have trusted traditional banks? Conversely, what will it take for FinTech platforms to be trusted sufficiently by future generations?

Back to Top

FinTech and Platforms

To answer this type of question, we first need to define what we mean by "FinTech" and "platform," and identify the "platform opportunities" in finance. We define FinTech as: "Financial sector innovations involving technology-enabled business models that can facilitate disintermediation; revolutionize how existing firms create and deliver products and services; address privacy, regulatory and law-enforcement challenges; provide new gateways for entrepreneurship; and seed opportunities for inclusive growth.a

We define a platform as an entity that provides "A nexus of operational and business rules; integrated technology architecture and engines; and channel access that facilitates exchange between two or more interdependent groups, usually consumers and producers."3,5

Internet businesses are increasingly structured as platforms, and in most markets, a few such platforms tend to dominate.

Platforms have a number of important properties, but today's "complete" Internet platforms always have three essential components:

  1. They are "open," allowing easy participation;
  2. They implement key business and operational processes, some of which typically exhibit network effects that increase in value as participation increases; and
  3. They implement these business processes automatically using enabling technology (which may also capture and generate vast amounts of data that enhances the value of the platform).6

Internet businesses are increasingly structured as platforms, and in most markets, a few such platforms tend to dominate. Facebook has dominated social networking with 1.8 billion users;b Amazon has a 65% market share in online books; Google's global search market share is 77%;c Uber claims more than 80% of U.S. market share in its market.d And so on. Complete platforms such as these become increasingly difficult to dislodge once they are established. As a result, the allure of market dominance provides a significant incentive to establish a complete platform.

Back to Top

FinTech Platforms

Because of historical factors and regulation, platforms in finance have tended to be "incomplete" in that they lack one of more of the three essential components.2 Exchanges are examples of some of the earliest "members only" incomplete platforms—the Royal Exchange of London (opened in 1571),e and the Osaka Rice Exchange (Dōjima kome ichiba, established in 1697) are early instances of these.4 Modern exchanges have replaced manual process with machines and added process sophistication, but are still generally accessible only to members in order to minimize risk to the platform. Similarly, clearinghouses and payment platforms with retricted accesss are incomplete.

Competition among platforms often arises from differentiation of some component. For example, electronic exchanges, the evolutionary successors to the Royal Exchange of London and the Osaka Rice Exchange, are complete platforms, allowing a much broader set of investors to trade assets such as foreign exchange, equities, or bonds directly with each other and without human brokers. Dozens of e-exchanges have arisen in the last two decades, each designed to meet some kind of specialized need, such as the ability to transact large sizes or to provide incentives to specific types of liquidity takers or makers, and so on. Liquidnet is an example of a complete platform, well suited for institutional trading involving execution of large blocks of assets that are often illiquid. Historically these were handled through brokers to minimize market impact. Now, robots match market participants wishing to transact, without third-party intermediation. In so doing they increase liquidity at a much larger scale and with more efficiency.

Lack of open access is not the only way in which a finance platform may be incomplete. Yahoo Finance, for example, is incomplete because it lacks key business processes to complete transactions. Ratings agencies are incomplete because most of their key value-adding business processes require human expertise and are not amenable to codification in IT systems.

The Venn diagram shown in the accompanying figure provides a graphical view of our framework, which combines the three platform components and also provides some examples. The intersection at the center defines complete platforms; all of the other regions denote incomplete ones. The Venn diagram representation is intuitive. The businesses on the periphery of the diagram tend to be those that play supporting roles, rather than central ones, in the life cycle of financial transactions. The platform framework can be useful for analyzing incumbent and new entrant business models in a number of ways. To start, the framework identifies which capabilities are required for each type of platform to attain completeness.

Conversely, the framework suggests which business functions are vulnerable to different types of disruptions due to their incompleteness. It also provides a way to think about transitions from partial to complete platforms, the implications of these, and the opportunities that would motivate such transitions. Finally, the framework permits us to examine new technologies in terms of the businesses they are likely to impact. Technologies that facilitate platform completion along a specific dimension will likely affect businesses that are incomplete along that dimension.

Complete platforms may also be created, by design, de novo. Amazon and PayPal are early examples in retail and payments. Peer-to-peer lending and robo-advisor platforms are recent instances of complete platforms in finance. Lending Tree, for example, is "open," and provides the technology infrastructure and processing required to connect lenders and borrowers directly. Robo-advisors are openly accessible to retail investors and aim to do much of what human advisors have traditionally done—screening investments and providing standard analytics like portfolio optimization—through technology.

Figure. The three core components of a complete platform with examples of platform businesses exhibiting various levels of completeness.

For businesses that are currently based on incomplete platforms, the diagram also suggests approaches to platform completion. One strategy for transitioning from a legacy incomplete platform to a complete platform is to add a missing component. This platform completion strategy can be pictured schematically by considering arrows originating at peripheral points on the diagram and terminating in the center. An example of this form of platform completion from a non-finance domain is the expansion of Angie's Lists from a home services rating system to home services booking system, through the addition of a process for conducting transactions. In finance, an equivalent example might be if Google or Yahoo Finance were to introduce capabilities like trading, investment advice, asset management (or even investment services) on their platforms.

An alternative strategy for FinTech platform completion is component replacement, aimed at introducing functionality that is cheaper, faster, or safer, relative to legacy platform components. Such replacement strategies most often seek to leverage the increasing speed, connectivity, and safety of technology components. A compelling FinTech example is the potential for Blockchain technology to replace legacy post-trade processes that currently require trusted third parties such as clearinghouses and depositories to manage and administer the clearing, settlement, and custody associated with trades and payments. Properly implemented, Blockchain-based systems could significantly reduce the need for specialized custodial institutions since, in principle, market participants would be able to exchange and record transactions securely in the cloud. This new technology, which essentially uses crowdsourcing to implement distributed ledgers, is currently too slow and expensive to replace existing practices. However, the expectation is that speed and cost improvements will displace platform components that address post-transaction processes. Indeed, Blockchain could serve as the core technology for other future complete platforms that require verification and trust.

The broad implications for FinTech platforms are clear.

Emerging AI technologies could be similarly disruptive in replacing existing business processes that involve basic human perception and reasoning about both structured and unstructured data from text, images, and sound. As the volume of unstructured data continues to increase, the artificial intelligence functionality for interpreting and acting on it automatically will likely replace a growing number of human-intensive processes.

Dominant AI platforms are already beginning to emerge. They are fueled by their access to "big data" (searches, purchases, and posts from participants using their technology) and are well positioned to become critical components of complete FinTech platforms of the future, where the platform owners can exploit their expertise in cloud computing, secure transaction processing, search and optimization, without incurring the up-front costs of formally entering the finance industry as full-blown competitors.

In the current regulatory landscape, platform completion and component replacement in FinTech seem likely to occur mostly through platform partnerships, driven by the forces of regulation and economics. Customer acquisition and regulatory compliance activities in finance can be very expensive for newcomers, making it difficult to dislodge incumbents. At the same time, incumbents tend to resist ceding access to their customers, and this aversion will likely induce them to pursue strategies centered on acquiring innovators for platform completion or component replacement while exploiting the existing levels of trust embodied in their brands. Finally—and perhaps most significantly—even though they are not natural participants today, once market participants begin to trust the data-handling and AI capabilities of the large Internet platforms, they could become core components of FinTech platforms for the same reasons they are trusted in the social, retail, and device domains.

Indeed, partnerships between technology platforms and financial services franchises are already emerging. H&R Block's partnership with IBM's Watson in the tax arena is a recent case in retail finance. Just as H&R Block chose not to build its own version of Watson from scratch, other financial players are unlikely to choose to build Google's or Amazon's artificial intelligence and machine-learning capabilities from scratch.

The broad implications for FinTech platforms are clear. If future investors and consumers of financial services begin to trust FinTech platforms in the ways they have for retail and travel, then financial advisory and intermediation activity may well go the way of manufacturing and brick-and-mortar retail stores. In such a connected world of data-intensive platforms, in which products and services can be micro-tailored to specific clients, and in which trusted platforms provide a broad spectrum of these products and services to these clients, will future generations really care if they buy stock through Fidelity, Google, or Amazon ... or through some combination?

Back to Top


1. Dhar, V. When to trust robots with decisions and when not to. Harvard Business Review (May 2016);

2. Dhar, V. and Stein R.M. FinTech platforms and strategy. MIT Sloan Research Paper No. 5183-16, 2017; SSRN link:

3. Dhar, V. and Sundararajan, A. Information technologies in business: A blueprint for education and research. Information Systems Research 18, 2 (June 2007);

4. Moss, D. and Kintgen, E. The Dojima Rice Market and the Origins of Futures Trading. HBR Case 9-709-044. Harvard Business School, 2010.

5. Parker, G. and Van Alstyne, M. Platform strategy. In The Palgrave Encyclopedia of Strategic Management, Macmillan, Hampshire, U.K., 2015.

6. Van Alstyne, M., Parker, G., and Choudary, S. Pipelines, platforms, and the new rules of strategy. Harvard Business Review (Apr. 2016);

Back to Top


Vasant Dhar ( is a professor at the Stern School of Business and the Center for Data Science at New York University.

Roger M. Stein ( is an adjunct professor of Finance at NYU's Stern School of Business and a research affiliate at the MIT Laboratory for Financial Engineering.

Back to Top


a. See

b. See

c. See

d. See

e. See

Copyright held by author.
Request permission to (re)publish from the owner/author

The Digital Library is published by the Association for Computing Machinery. Copyright © 2017 ACM, Inc.