Le rapport formule 36 recommandations politiques sur le financement et les modèles économiques pour rendre le modèle de données FAIR durable. Il apporte des éléments aux décideurs sur la mise en place d'actions à court et à long terme, pertinentes, pour la mise en œuvre des principes FAIR.

Cost-Benefit analysis for FAIR research data – Policy Recommendations

FAIR research data encompasses the way to create, store and publish research data in a way that they are findable, accessible, interoperable and reusable. In order to be FAIR, research data published should meet certain criteria described by the FAIR principles. Despite this, many research performing organisations and infrastructures are still reluctant to apply the FAIR principles and share their datasets due to real or perceived costs, including time investment and money. To answer such concerns, this report formulates 36 policy recommendations on cost-effective funding and business models to make the model of FAIR data sustainable. It provides evidence to decision makers on setting up short and long-term actions pertinent to the practical implementation of FAIR principles.

 

Executive Summary

To drive the implementation of the FAIR principles in Europe, the European Commission together with a number of pioneering European research stakeholders is taking measures to raise the awareness about costs and benefits of FAIR data, and is encouraging funding bodies to set guidelines or support the development of an infrastructure for publishing FAIR data. In a study which preceded this report, the cost of not having FAIR data for the EU-28 has been estimated at EUR 10,2 bn per year, and this is bound to grow unless action is taken.

Despite this, many research performing organisations and infrastructures are still reluctant to apply the FAIR principles and share their datasets due to real or perceived costs, including time investment and money. To answer such concerns, this report formulates 36 policy recommendations on cost-effective funding and business models to make the model of FAIR data sustainable. It provides evidence to decision makers on setting up short and long-term actions pertinent to the practical implementation of FAIR principles.

The recommendations presented in this report are organised in two groups:

  • The first covers recommendations for covering the initial costs for FAIR research data implementations in Europe, while
  • The second covers recommendations for covering the sustainability of FAIR research data implementations in Europe.

 

Recommendations for covering the initial costs for FAIR research data implementations in Europe:

(a) Work out the business case for FAIR at the national level

Rec. 1. Apply the cost-of-not-having FAIR methodology in every EU member state

Rec. 2. Apply the cost benefit mechanism for the strategic research centres in EU member states, such as data-intensive research labs (e.g. genomics), and data infrastructures (e.g. ELIXIR, CLARIN)

Rec. 3. Think beyond of organisations and disciplines. Cross-disciplinary FAIR data use cases have the potential to create positive externalities, spill-over effects and innovation

Rec. 4. Integrate the outcomes of the national FAIR cost benefit assessments at the European level to identify and quantify positive spillovers and externalities

(b) Prioritise investments in the national FAIR implementation roadmap

Rec. 5. Build a solid FAIR baseline across Europe by prioritising high-impact and high-feasibility activities to maximise ROI. Start with activities related to Findability and Accessibility, such as common data management policies and practices, metadata standards, persistent identifiers and common research data infrastructures

Rec. 6. ROI will come only if the current working behaviours around data management and sharing change. Invest early enough in culture change and skills development

Rec. 7. Establish a working group under EOSC which will be mandated to decide on FAIR investment priorities, evaluate current progress and prepare future development roadmaps

Rec. 8. Move towards shared national and European cross-discipline Cloud-based data infrastructures which can significantly reduce the data storage and compute costs, and can drive time and cost efficiencies in data access, sharing and collaboration

Rec. 9. Use emerging technology, such as artificial intelligence and robotic process automation for automating and industrialising repetitive, standardised and time-consuming activities, such as data transformation, data classification or assignment of identifiers, to reduce operational costs linked to FAIR implementation

Rec. 10. Working in iterations and increasing the maturity of FAIR research data implementations in well-defind cycles helps to align investments with progress towards achieving the policy objectives

Rec. 11. Engage in the iterations at least one discipline or country which has not yet started their FAIR implementation or is lagging significantly behind to achieve buy-in, encourage them and share with them lessons learnt, actionable advice and reusable outcomes of others

Rec. 12. Opt for demand-driven provisioning of FAIR data, within and across research disciplines, to optimise investment expenditure and maximise ROI. Sustainable growth in the maturity of the FAIR implementations as well as in the number of FAIR data available will lead to network effects

Rec. 13. Provide financial incentives, such as grants and funding, for making legacy data FAIR on a demand-driven basis

(c) Measure progress and impact of FAIR implementation

Rec. 14. Endorse and provide financial support to a working group under EOSC which coordinates and monitors FAIR implementation at the European level to ensure the alignment between investments and spending compared to the level of achievement of the FAIR policy objectives

Rec. 15. Create a European mechanism for measuring progress, for example based on earned value management

Rec. 16. Create a European FAIR implementation maturity model which will define the activities required to achieve a specific level, the associated costs and the expected benefits.

Rec. 17. Provide the expertise and financial assistance for helping countries apply the maturity level and making the transition from one level to the next one.

Rec. 18. Define templates for service-level agreements with which trusted FAIR research data infrastructures will need to comply, for establishing a European baseline for service quality

(d) Share and reuse knowledge and solutions within and across countries and disciplines

Rec. 19. Benefit from significant efficiency savings in the total cost of ownership of FAIR data implementation by reusing solutions, technical assets, practices and experiences between FAIR and other data-related policy implementation initiatives, such as those of the revised PSI directive, GDPR, INSPIRE and CEF Telecom

Rec. 20. Mutualise FAIR implementation resources and investments across countries and disciplines by co-investing in common frameworks, solutions, technical assets and shared services

Rec. 21. Explore business models for FAIR research data infrastructures and services based on shared service provision, e.g. following the example of the CEF building blocks

Rec. 22. Provide financial support for developing and customising FAIR-compliant open source in collaboration with the European open sourcecommunity, and multiplicate savings by sharing across the EU research community

 

Recommendations for the sustainability of FAIR research data implementations in Europe

(e) Explore mixed business models for FAIR research data infrastructures

Rec. 23. For the sustainability of FAIR research data implementation, research data infrastructures and research performing organisations must shift the focus towards data monetisation and value-added data services.

Rec. 24. Several alternatives exist for FAIR data and data services pricing models, from profit maximisation and cost recovery through to charging only for marginal costs (hence coming closer to the open data paradigm). Several parameters have to be considered for the selection of the right one, including the way data creation was funded, applicable IP or patents, data management costs, and value added.

Rec. 25. FAIR research data infrastructures must be encouraged and supported via fiscal incentives, such as seed funding, tax breaks or deductions, and policy interventions, including legislation, to explore mixed business models, which combine a healthy balance between public funding and other revenue streams.

Rec. 26. Fiscal incentives, e.g. tax breaks or deductions, will encourage industry to form partnerships, collaborate, sponsor, fund or buy data/services from FAIR research data infrastructures, to broaden the market for FAIR data.

(f) Secure public funding for implementing and sustaining FAIR research data implementation

Rec. 27. Funding FAIR data implementation has to remain available not only at the European level, e.g. as part of Horizon 2020 and Horizon Europe, but also as part of the national research and innovation programmes. FAIR applies to all publicly funded research in Europe

Rec. 28. FAIR-related costs, e.g. for data stewardship and management, or data infrastructure operational costs must be made eligible for specific cases and only if repored at a granular level respecting transparent cost accounting practices

Rec. 29. Culture change related costs, including training and awareness raising activities, must be made eligible based on transparent cost accounting practices

Rec. 30. FAIR-by-default policies and mandatory FAIR compliance must be included in the award criteria of research grants.

Rec. 31. Incentivise research data infrastructures and research performing organisations to reinvest savings made as a result of FAIR in the sustainability of FAIR implementations

(g) Develop a community and an ecosystem around FAIR data

Rec. 32. Provide financial support for communication, knowledge sharing, community building and marketing projects and activities for example via continuing with coordination and support actions through European and national research programmes. Full costs to be made eligible for all types of participants

Rec. 33. Provide financial support for organising mutual learning exercises, in the context of which member states, third countries, FAIR data practitioners and experts work together on a topic of common interest, such as FAIR data management on the open science cloud, FAIR data management in AI applications, or the costs and revenues for preparing a data infrastructure for joining the open science cloud using the cost benefit mechanism developed by the current study

Rec. 34. Take measures and make the means available for encouraging innovation through cross-disciplinary projects and applications. Such means may include promoting the use of FAIR research data in projects funded under the current focus areas of existing European and national research programmes, or defining new innovation stream with a focus on research, life sciences and/or SMEs under future European and national research programmes, such as Horizon Europe

Rec. 35. Consider the establishment of a public-private partnership focusing on creating societal and economic value from FAIR data

Rec. 36. Place universities at the heart of the European FAIR data community of practice. They are the most important type of research performing organisations and are also the ones preparing the workforce of tomorrow who will be able to support the implementation of FAIR principles in Europe. To this end, support and incentives should be provided to them for reviewing their curricula and current data management practices in the light of FAIR.

 

© European Union, 2018