Skip to main content

Kennedy School Review

Topic / Science, Technology and Data

Government as a Platform: How Policy Makers Should Think about the Foundations of Digital Public Infrastructure

For all the promise of digital government, it has done very little in the Western world to fundamentally alter the structure or processes of the state. Through much of its history, the primary goal of modernization and digitization has been to automate repetitive work or back-office tasks. The paper form became an electronic survey; the accounting book became a spreadsheet. Even new headline-grabbing groups such as the United States Digital Service have mainly focused on improving user experiences or putting out fires—and have not changed the underlying structure of government.

Because automation and digitization are driven at the agency level, a fundamental rethink of how government functions is difficult to bring about. The siloed nature of government means that the purchasing and building of technology occurred in parallel, even when the requirements or business operations were similar or even identical. Instead of creating a digital government, modernization efforts created many digital governments, each with its own security flaws, inflated costs, and distinct user experiences. The result is a maze of databases and applications scattered across government departments that are difficult, or nearly impossible, to integrate and function together.

To understand the impact of this arrangement, it’s helpful to think of the national highway system as a metaphor. Imagine an agency has been tasked to create a new road to move traffic between downtown Boston and Manhattan. They draw a design, plan the route, and pour concrete. But then they get an additional task: create another route from Boston to Brooklyn. One planner suggests an off-ramp from the planned highway but is told this won’t work; the way that the highway was built makes an off-ramp impossible. It will be necessary to build a new route all the way from Boston to Brooklyn.

They’ve been asked to do a job that is clearly duplicative. But in the world of digital infrastructure, when the government builds databases and digital services, this is often how it works. Systems are built with a specific task in mind, and so leveraging existing digital infrastructure to deliver new services is often difficult or impossible.

In the private sector, large organizations face the same fundamental challenge. One reason legacy companies struggle is because they, too, developed technology from scratch within siloed departments. The new leaders like Amazon, in contrast, build common databases and applications that can be leveraged across the organization. This reduces transaction costs and allows new services to re-use infrastructure others have built. Like Amazon, modern government now provides a large number of services that are partly or fully online, like passport renewal, vehicle registration, tax filing, and health-record retrieval. But government is managed like legacy companies: systems are built at inflated cost, managed within departmental silos, and unable to connect effectively. Companies like Amazon use a platform approach to scale up services using central tools to meet the complex needs of diverse users. A parallel approach could help the public sector achieve similar goals.

Platform government moves the public sector away from duplicative, siloed structures and toward a shared, interoperable infrastructure and canonical datasets. It organizes a technology foundation that allows governments, civic institutions, and businesses to deliver radically better services to the public more safely, efficiently, and accountably. And in the long term, it could transform the way that the government works with technology—and relates to citizens.

This essay introduces the early work being done around the world to build public-sector digital platforms at scale. We begin by discussing the evolving ideas and ambitions of digital government platforms—and articulate a definition that policy makers can use. We then discuss potential risks and governance challenges with a digital-platform approach in the public sector before providing recommendations for how policy makers should think about these issues.

* * *

Platforms are not a new concept in the public sector. The highway system in the United States—and the German Autobahn on which it is based—provide useful comparisons. A shared, national network of large, multi-lane roads links to a broader network of smaller local streets. The system works extremely well for the transportation of bulk goods across hundreds and thousands of miles; it also provides huge value for ordinary users, who only need to comply with a simple regulatory framework (e.g., lanes, engine types, vehicle width) to ensure accessibility across the entire network. Just as important, the system works for “producers” at the state and city levels, who can connect smaller-scale roads to the larger network based on a simple set of interoperability rules (e.g., turn ratios). But until relatively recently, this style of “platform thinking” was not applied to government technology—meaning that, to date, there has not been a clear, shared definition of platform government that policy makers outside of the technology world can use.

In 2010, internet advocate and technologist Tim O’Reilly kickstarted the conversation, arguing that companies like Apple and Google—moderators of a robust and competitive marketplace rather than original creators of content—offer a powerful model for reimagining government. Rather than recreate services endlessly for each context, “a platform provider builds essential infrastructure, creates core applications that demonstrate the power of the platform and inspire outside developers to push the platform even further, and enforces ‘rules of the road’ that ensure that applications work well together.”[1] Since O’Reilly’s essay, much attention has been paid to how designing government as a platform removes silos and allows for coproduction of policy and services, which generates better services for less money and with greater accountability.[2]

The core requirement of a platform-government model is that government provides access to a set of canonical databases and shared applications that other parts of government, civic society, and the private sector can leverage to solve public problems. The clearest current example of such a system is in Estonia, where a national individual identification layer sits alongside communication infrastructure called X-Road that moves information between agency data registries.[3],[4],[5] If an agency or private company wants to create a new service, it must comply with some basic data standards to build on top of the X-Road stack. In practice this means that, instead of creating a database and application layer, they need only create the “last mile” to the user. This allows new services to be created more quickly and efficiently, and with more citizen input. Practitioners in Estonia have already found that lower-sunk costs mean that there is far less opposition when it comes time to sunset a tool or service.[6]

Building on government’s shared infrastructure removes silos and allows for coproduction and collaboration across sectors. As co-founder of the United Kingdom’s Government Digital Service Mike Bracken argues: “Siloed approaches to transformation don’t work. Reinventing the wheel every single time we build a service has led to far too much duplication and waste.”[7]

The most interesting work in platform government currently happens in Estonia and India where there is less legacy software and fewer established business processes to impede a new model. But more mature markets are looking to explore a platform approach. For example, the Obama administration’s 2012 Digital Government Strategy directed federal agencies to “look first to shared solutions and existing infrastructure when developing new projects, rather than procuring new infrastructure and systems for each new project.”[8] The new Canadian federal Directive on Information Technology directs agencies across the government to adhere to detailed, mandatory procedures for open application program interface (API) structures. In the short term, these directives push agencies to do work in similar ways. In the longer term, it’s not yet clear whether these edicts are truly changing behavior within the silos.

Breaking down silos between agencies enables government and other sectors to collaboratively develop policy and services. Civic-society organizations and the private sector already play critical roles in the distribution of public services (that is why they are sometimes referred to as the “wholesale” side of government).[9] When government facilitates access to businesses, it can “encourage wider digital innovations that internal public service employees might never dream of” if they were creating services alone.[10]

* * *

Government as a platform reorganizes the work of government around shared APIs, open standards, and canonical datasets so that civil servants, businesses, and others can deliver radically better services to the public more safely, efficiently, and accountably. But several critical challenges still face the rapidly evolving world of platforms. We may need to rethink many core structures in government and design new institutions to meet the needs of a platform-enabled world. There are governance questions that we simply haven’t yet been able to address. The answers we find could dramatically impact the relationship between citizens and the state. Before we race into the future, a great deal of critical thought is still needed about the risks associated with top-down-platform mandates and what happens when central governments arbitrarily decide what works best.

To Make Platforms Work Effectively, We May Need New Institutions that Can Work and Adapt in the Digital Age

Former British Cabinet minister Francis Maude argues that: “[if] you were to create government today, you would not build it around large, free-standing Departments of State. Instead of a series of siloed hierarchies, you would structure it as a platform responding to the needs of the end user.”[11] Right now, platforms are located where it was politically or operationally convenient for them to be built. India platforms are owned by quasi-autonomous non-government organizations that can access external IT talent. In the United States, 18F was tasked with innovation, so it owns login.gov. In the UK, Verify and Notify sit in the Government Digital Service in the Cabinet because that’s what was politically expedient. The location of these services are artifacts of where it was easy to create them, not where it would be appropriate to govern them. The challenge is understanding what kinds of governance models, and what kinds of institutions, should be at the center of these platforms.

A Platform World May Violate Many of the Norms and Expectations between Government and Citizens

The ability to share data that can then be leveraged by agencies across the public sector is exciting because of the applications we can imagine: better linking health outcomes to various services to improve how we treat patients or allowing more people to enter and participate in the formal economy. But it also poses complex questions about citizen privacy and trust in their government. When we share data with one agency, we might be able to imagine how it is used. When that data is being fed into a cross-government platform, we lose out on our ability to see, understand, or question what happens to our information. An increasingly complex data infrastructure limits citizen visibility into what is happening behind the scenes—making a focus on legibility and understandability critical if platforms will cohere with democratic values.

The Agencies that Control Servers and Own Services Will Wield Huge Influence over Government

Opting in to a shared-platform approach implies giving up some authority over how data are managed and how services are structured. And if policy makers aren’t careful, they will invest early platform owners with control over access to datasets with enormous value. Turf wars between agencies could become catalysts to lock down resources and constrain development. Having public tools that are core to so many services and scaled so broadly across government creates thorny issues that have yet to be addressed by most governments pursuing a platform approach. These issues must become a priority if platforms are to succeed.

Centralized Planning Could Focus Too Much on Efficiency and Standardization and Lose Sight of the Real Upside Offered by Platforms: Radically Better Services for the Public

What platforms have in common with government is their capacity for scale. Married together, they are both compelling and scary. In Seeing like a State, historian James Scott found that unbounded optimism in technology-driven progress, combined with a disconnect between central planners and actual users, led many of the last century’s most ambitious public-sector plans to dismal failure.[12] There is a positive side to this: if one has finite resources and expertise in something like user-centered design, they can concentrate those staff and scale their work across government to maximize their value and contribution. The United Kingdom’s success in rolling out a series of platform services, like Verify, Pay, and Notify, serves as a great example.[13] But platforms are only about scale. On their own they won’t improve or democratize the choices we make, only radically grow the impact of those choices on the lives of human beings.

As platform government moves from a theoretical possibility to reality—as it already has in nations like India and Estonia—policy makers should be both optimistic about possibilities and clear-eyed about the challenges and risks that remain. Platform infrastructure for the digital age could transform the way that government designs, builds, and delivers services to the public. The degree to which these tools can truly improve life for citizens will depend on how well policy makers can address challenges in institutions, democratic values, and governance.

Edited by: David Hicks

Photo by: aotaro, Wikimedia


[1] Tim O’Reilly, “Government as a Platform,” MIT Press Journal 6, no. 1 (2010): 13–40.

[2] Richard Pope and Malika Mehrotra, “Definitions: quotes and themes on ‘government-as-a-platform,’” Platform Land, 27 November 2018, https://medium.com/platform-land/definitions-quotes-and-themes-on-government-as-a-platform-b73d287ca9a1.

[3] More information about the E-ID platform can be found at “e-identity,” e-estonia, https://e-estonia.com/solutions/e-identity/.

[4] More information about the X-Road platform can be found at “x-road,” e-estonia, https://e-estonia.com/solutions/interoperability-services/x-road/.

[5] Estonian citizens can use their e-Identity to access a variety of services, including travel, health insurance, bank access, digital signatures, voting, medical records, taxes, and prescriptions.

[6] David Eaves and Ben McGuire, “The Future of Governance for Digital Platforms,” in David Eaves and Ben McGuire, 2018 State of Digital Transformation (Belfer Center for Science and International Affairs, Harvard Kennedy School, October 2018).

[7] Mike Bracken, “Government as a Platform: the next phase of digital transformation.” Government Digital Service (blog), 29 March 2015, https://gds.blog.gov.uk/2015/03/29/government-as-a-platform-the-next-phase-of-digital-transformation/.

[8] CIO Council, “Digital Government: Building A 21st Century Platform to Better Serve the American People,” https://obamawhitehouse.archives.gov/sites/default/files/omb/egov/digital-government/digital-government.html.

[9] Aneesh Chopra and Nick Sinai, “Wholesale Government: Open Data and APIs,” Shorenstein Center for Media, Politics and Public Policy, 9 April 2015, https://medium.com/@ShorensteinCtr/wholesale-government-open-data-and-apis-7d5502f9e2be.

[10] Sarah Barns, “Smart cities and urban data platforms: Designing interfaces for smart governance,” City, Culture and Society 12 (2018): 5-12.

[11] Francis Maude, “Foreward,” in Jonathan Dupont, The Smart State: Redesigning government in the era of intelligent services (Policy Exchange, 2018) [PDF file].

[12] James C. Scott, Seeing like a State: How Certain Schemes to Improve the Human Condition Have Failed (New Haven, CT: Yale University Press, 1998).

[13] Respectively, UK government websites for verifying individual identification to access services, setting up payment sites for government fees, and creating automated alerts from agencies to citizens.