Can the data revolution drive good governance?

From many different quarters, incentives to harness the power of data for good governance are on the rise. The open governance movement, the Sustainable Development Goal process, and efforts to improve climate adaptation and disaster resilience are just some of the drivers behind this trend.

Of course the data revolution – the confluence of a rapid rise in the power of data technology, a dramatic fall in the cost of such technology, and a broad diffusion of relevant capabilities – helps give a sense of urgency and feasibility to the quest. Yet the full implications of designing data systems meant to drive good governance in the context of the data revolution have not fully sunk in. Here I focus on one important implication – the need to think of data systems from a portfolio perspective.

Before the data revolution, governance data systems tended to be designed in a way that applied the single best approach to measurement for a given data need. When data collection was comparatively expensive and technology change comparatively slow, such an approach served well. The methodology for counting population through a national census has remained robust and powerful for most of the last millennium.

Today, however, change is fast and costs are plummeting, so it doesn’t make sense to put all your data eggs in one instrument basket. Instead, designers of data systems need to think about what package of data collection strategies will produce the best value-for-cost combination. In much the same way that modern portfolio theory has shown how investors can improve risk-adjusted returns by combining distinctive asset categories, architects of data systems are discovering that they get better results by combining multiple approaches to data collection.

There are now worked examples across many different geographical settings and policy areas that have demonstrated the value of combining traditional data collection procedures with elements drawn from crowd sourcing, high-tech sensor technology, data mining, and other emerging approaches. Such approaches to mapping agricultural soils, for example, have enabled Ethiopia to rapidly improve its ability to characterize variation in soil conditions which in turn have permitted the country to offer place-specific fertilizer regimens. Yields have increased at the same time as ecosystem damage has been reduced.

The implications are profound. Those charged with designing data systems in the past had to know one thing well; today they must be comfortable with a broad array of technology and institutional choices. In the past there was a premium on consistency and stability; today the imperative is to adapt to change fast enough to take advantage of innovation but without triggering unhelpful disruption and distrust. As a result the task of designing a fit-for-purpose data system is increasingly a task that only a purposive community can take on, because no single individual or organization can plausibly have the right information.

In the private sector, most leading firms have responded to the rapidly changing data technology landscape by concentrating responsibility for strategic planning with respect to information systems in the position of a Chief Information Officer (CIO), as opposed to the head of Information Technology department. Within firms, CIOs are able to map organizational needs to the broad landscape of information technology in order to design data and analytic systems that add value. Where such strategies thrive they can go even further, as in the case of the innovations in information technology and informatics developed by Jack Levis at UPS which have been so successful that the core work is baked into the entire DNA of the firm.

Few countries or international organizations yet have Chief Information Officers, but all need to start building the kinds of capabilities associated with them. Successful data decisions today require carefully calibrating a portfolio of measurement solutions to meet decision-making needs.

The Summit on the Global Agenda 2015 takes place in Abu Dhabi from 25-27 October.

Author: Marc Levy, Deputy Director, Center for International Earth Science Information Network (CIESIN), Columbia University, a member of Global Agenda Council on Data-Driven Development.

Image: A scientist performs maintenance in the CERN LHC computing grid centre in Geneva, October 3, 2008. REUTERS/Valentin Flauraud