Internet, Politics, Policy Conference 2016: The Platform Society

Wally BostonThe fourth conference organized by the Policy and Internet Journal (PIJ, founded in 2009) and the Oxford Internet Institute (OII) was held at the Mathematics Institute at Oxford University in late September. I was pleased to attend and represent American Public University System. The following recaps some of the presentations.

“We move from one online platform to another as part of our daily lives,” said Professor Helen Margetts, OII chair. Dr. Margetts stated that online platforms have been able to create major shifts in labor utilization (think Uber, Lyft and the demand for taxi drivers), and academic research regarding online platforms expands on the theme of the last OII and PIJ conference, crowdsourcing. One of the policy implications of platforms includes the ability to offer micro-labor opportunities across all inhabited geographic regions, while concurrently eroding hard-won labor opportunities.

Government can conspire to become a platform. Among other benefits, governments and their departments offer scale that many commercial platforms may be unable to achieve. Large digital platforms provide opportunities for academics to understand human behavior in ways we have not understood before. The beauty of the conference and PIJ is about creating opportunities for multiple academic disciplines to focus their research on a single topic.

Eddie Copeland, Director of Government Innovation at Nesta, delivered the opening keynote. Nesta is a U.K.-based innovation charity. Years ago, Mr. Copeland’s first job was to integrate Information Technology (IT) for a petroleum company and its suppliers, while overcoming their collective IT communication and compatibility issues. His next job was with the U.K. government, with the same issues. He noted that one of the challenges of working with governments that outsource IT is that the government might have to pay its supplier to access its own data.

“All governments like to think their needs are unique,” stated Mr. Copeland. Far too many systems are custom-built and incompatible. In the U.K., those are referred to as bespoke systems. Government as a platform is an idea in which custom systems are replaced with small, commodity systems designed to get it right. Big IT is dead. It is far too easy to interconnect smaller systems and scale them when needs increase. Gov.UK is such a platform for publishing. Amazon, eBay and iTunes are effective and efficient platforms that scaled up with demand. What exists across the world is “Government as Lego.” If governments become a monopoly of self-suppliers, they cannot out-innovate platform suppliers to for-profit companies.

Most governments share the philosophy that, “if we just have the right technology, we can collect the right data and if that’s in place, we can do the right amount of work.” Bolting new technology onto an inefficient process is wasteful. According to Mr. Copeland, governments need to ask themselves the following questions: How do we want to work, what type of data do we need and what’s the technology to accomplish that?” Releasing data on a specialized application with just 200,000 followers, instead of millions, is inefficient. Getting the data out in different formats is difficult. There are legal barriers, reasonable and perceived.

Mike Flowers, founding director of the New York Mayor’s Office of Data Analytics, saw this in Iraq when tracking Improvised Explosive Devices. He started doing shared services intelligently. Annually in New York City, there are 18,000 property complaints, of which only 8% are legitimate. The first 25% of inspections generated 21% of reports on dangerous buildings, Flowers noted. Establishing an Open Data Portal (ODP) allows access by local authorities and the public sector. As public centers release the data, they feel like it’s a cost center that cannot be maintained. London is attempting to build an ODP so they can overlay it and see where services should be targeted.

Platform businesses are brilliant and disruptive because they have blown apart traditional business models. Uber owns no vehicles, Facebook creates no content, Alibaba has no inventory and Airbnb owns no real estate. GoodSam is an app that  lists  anyone who is medically-qualified and approved. Family by Family is an Australian app that provides families with mentors to help them work through difficult experiences. Governments need to ask the right questions and build apps that operate as effectively as commercial ones.

Michael Castelle from the University of Chicago presented a paper, “The Platform as Exchange: Financial Metaphors for the Regulation of Marketplace Platforms.” Mr. Castelle noted that debates about how to regulate stock exchanges were similar to debates about shared economies and platforms. Marketplace platform providers eBay, Uber and Airbnb produce markets for goods and services matching buyers and sellers which, in a fixed-role market, are distinct groups. In a switch-role market, in contrast, participants can play the role of buyer, seller or both (stock buyer/seller). Switch-role markets change competition and often cause fragmentation of financial markets. One example is Instinet, which registered as a broker but operated its own platform to match buyers/sellers. Another is Uber, whose drivers are customers, and customers, drivers. In both cases, there are low switching costs since there are no off-exchange restrictions.

David Nieborg from the University of Toronto and Anne Helmond from the University of Amsterdam presented their paper, “From Web to App(s): a PlatformStudies Approach to Facebook’s Political Economy.” They reviewed the evolution of Facebook from a web-based platform serving college students to a mobile-first enterprise platform serving all individuals, with 1.1 billion daily users. This evolution comprised its move to become a computational and business platform, beyond a social platform. Platforms like Facebook are the dominant computational, infrastructural and economic model of the social web. Part of the evolutionary process includes platformization – the decentralization of platform features and recentralization of platform-ready data.

To examine the computational model, review the changes in Facebook’s architecture. Tracking the business platform evolution simply requires following the money from users, businesses, developers, advertisers, publishers and marketing partners. The economics model evolves around platform envelope theory, in which an app in the platform becomes so important that it sucks up the app. Facebook failed in launching their mobile app so they decided to be the glue and provide the infrastructure in return for the data, which continues to drive model evolution.

Jean-Christophe Plantin and Alison Powell from the London School of Economics and Political Science presented their paper, “Open Maps, Closed Knowledge: what the Platformization of Maps Means for Citizenship and Society.” Google Maps, Open Street Map, the U.S. Geological Survey and the Ordnance Survey expanded maps from purely government to civilian data. Infrastructures and knowledge were combined, providing essential services, accessibility, exhaustive coverage and reliability. Infrastructure and platform are similar and often cohabitate, overlap or replace each other. What happens to knowledge and citizenship when special services are provided by platforms, and not infrastructure? There is the notion that a number of intersecting tensions exist; how do we get knowledge about the places where we live and work in a more intersecting way?

There are third-party platforms. For example, Citymapper recentralizes knowledge production by pulling open, as well as proprietary, data and makes it more accessible to users. Dashboards are a critical part of these types of platforms. It matters greatly what data streams turn up on your dashboard. The platform creates the logic for understanding the data. Transport for London is a product that no longer feels that it needs to make a new tool as long as other tools are available.

What results from the tension between decentralizing, and recentralizing knowledge? If platforms are installed as infrastructures, what is gained or lost for citizenship? Platforms may not be built for scale or sustainability. What can platforms gain from infrastructure? In his book, Seeing Like a State: How Certain Schemes to Improve the Human Condition Have Failed, James C. Scott wrote that the state renders the knowledge and makes the map. If the platform is the passage point, the tension point is more complex.

Conference attendance at Oxford is much smaller than at conferences in the U.S, with approximately 125 attendees at this year’s event. When I asked about the “low” attendance, I was informed that any such conference with more than 100 or so attendees does not provide the proper environment for conferring among participants. That’s certainly a different perspective from what we experience in America. Nonetheless, I was inspired by the variety of papers presented, abstracts of which are available on the conference website.

I am glad to see that institutions like OII are organizing conferences where researchers can gather to discuss topics that may not be in the mainstream of current academic research. Platforms like Facebook, Uber, Airbnb, Lyft, Google Maps, etc., are being utilized by millions, if not billions, of people. In many cases, governments are behind in studying the impact of platforms and determining if regulations are necessary to control their growth or behavior. Meanwhile, entrepreneurs and large companies alike are racing to build the next billion-person platform and not likely weighing all of the considerations important to users and governments. Usually, the OII conference publishes the papers presented in a special edition of PIJ. I recommend that anyone interested in this topic consider obtaining an electronic or print version of the journal when it becomes available.

Subjects of Interest

Artificial Intelligence/AI

EdTech

Higher Education

Independent Schools

K-12

Science

Student Persistence

Workforce