The value of data has become quite evident over the past decade, and while there has been a growing appreciation of the societal value of public data, there has been much less success in opening up ostensibly private data sources to aid in the delivery of societal benefits.
Many of the early efforts to do just this have revolved around data collaboratives, and a new report by The GovLab explores just how prominent such efforts have been, and what lessons can be learned from early movers.
The report examines over 150 case studies of data collaboratives to explore the collaboration behaviors they exhibit, and the models they work under. In total, six distinct models for data collaboration are identified in the paper:
- Public Interfaces: Companies provide open access to certain data assets, enabling independent uses of the data by external parties.
- Trusted Intermediary: Third-party actors support collaboration between private-sector data providers and data users from the public sector, civil society, or academia.
- Data Pooling: Companies and other data holders agree to create a unified presentation of datasets as a collection accessible by multiple parties.
- Research and Analysis Partnership: Companies engage directly with public-sector partners and share certain proprietary data assets to generate new knowledge with public value.
- Prizes and Challenges: Companies make data available to participants who compete to develop apps; answer problem statements; test hypotheses and premises; or pioneer innovative uses of data for the public interest and to provide business value.
- Intelligence Generation: Companies internally develop data-driven analyses, tools, and other resources, and release those insights to the broader public.
Each of the models contains a number of real-world examples to illustrate how things function, ranging from BBVA’s Urban Discovery project to Orange Telecom’s Data for Development Challenge.
Choosing the right approach
It perhaps goes without saying that with so many different approaches to choose from, the matter is far from straight forward, and there is no real best or worst approach to take. Nonetheless, the authors provide three recommendations to help guide action for organizations taking the path towards data collaboratives.
Firstly, they recommend ensuring you have the ability to assess the variables in the data collaborative so that you can assess the merits of the approaches outlined above. They then advocate the creation and empowerment of data steward roles to devise new ways of establishing public value from cross party collaboration. Last, but not least, they argue for new intermediaries to lower transaction costs among data suppliers and users.
“The analysis presented in this paper is a start, an informational basis describing current practice and providing clarity on the variables that determine a data collaborative,” the report concludes. “But moving forward, we need more “data about the use of data,” and we need more analysis to achieve actionable intelligence on how to structure data collaboratives given certain requirements and needs.”
