Data Ecosystem (DE) represents a new production model that is increasingly adopted by public and private stakeholders to share investment costs and maximize data sharing. While the benefits and costs associated with a DE are clearly perceived, its accurate definition both in terms of scope of activity, mechanisms and incentives for operation have not yet been sufficiently explored. This paper seeks to define a theoretical and conceptual framework within which DE can be properly analysed. In particular, the methodological approach to measure the information return of the data integration process also in economic terms is identified in the theory of the Value of Information (VoI) approach. This approach, based on Bayesian statistical inference, assumes the benefits of a decision-making have a direct relation with the quantity and the quality of available data. The paper outlines the possibility of extending the approach to the specific case of DE, highlighting its advantages and potential problems.