Design and evaluation of a reference model for data strategy development

 The value of data to organisations, whether public or private, is undisputable yet public sector entities,  particularly those of local government, have yet to everage this value. Instead data is often perceived  as a by-product of transactional processing which little inherent value. This is compounded by the  ever-increasing volume of data, and the complexity of managing this data within the diverse and siloed  delivery models required by local authorities. Statutory requirements and legislative imperatives, such  as GDPR, place a significant responsibility on the public sector in the management of data with most  giving little attention to the strategic management of this valuable asset as they scramble to deliver  services with limited resources and changing landscapes.

The overall objective of the research is to design and evaluate ways in which local government data  management can be improved within a Design Science Research (DSR) context, stated as follows: (1)  Identify the components of a data strategy development framework using a taxonomy-building  approach; (2) conduct a case study to evaluate the ensuing artifact; (3) evolve and complete  construction of the artifact as a reference model for data strategy development; and (4) complete an  ex-post evaluation of the artifact that demonstrates its utility. 

 DSR entails the development of an artifact, which can be a model, method, or instantiation, that is  theoretically underpinned and rigorously evaluated. DSR falls within the pragmatist philosophical  paradigm where a real-world emphasis and practical manifestation are to the fore. For the current  study, the emphasis is on creating a useful artifact within an established domain, that of public sector  data, but for which there is a dearth of solutions to improve how this data is managed.

 Though the data management domain has much available by way of technology and tools and there  is a body of research available in the IT and data governance fields, yet there is little research to guide  the strategic management of data. Enterprise Architecture (EA) speaks to the alignment between  business and technology in an organisation and existing EA frameworks and tools are adopted in the  development of the solution for the current study. This EA-based approach guides the design of the  solution artifact, a Reference Model for Data Strategy Development (RM-DSD) and follows the rules  of the ArchiMate Modelling specification. 

 There are two key contributions of the study: that of the theoretically underpinned Reference Model  for Data Strategy Development (RM-DSD) guided by Tallon et al (2013) An Emerging Theory of  Information Governance, and, the development of a qualitative research method for reference model  build (QRM-RMB) underpinned by Nickerson et al. (2013) A Method for Taxonomy Development and  its Application in Information Systems. The noted primary research contribution is the RM-DSD,  evolved from a rigorous taxonomy development method, represented as a conceptual framework  which in turn serves as the viewpoint for the development of the model. The model, as represented  through the ArchiMate (The Open Group 2019) modelling specification, is presented over three layers:  the motivation, strategy, and business layer as a prescriptive tool to guide data strategy development.  A rigorous evaluation follows the Pries-Heje et al. (2008) strategies for ex-ante and ex-post evaluation.  The research methods include case study, focus groups and interviews.

 Findings from the study confirm the need for a strategic approach to data management and the utility  of the derived artifact as a reference model for data strategy development. The value of the qualitative  research method has also been evident in its use throughout the current study. 

 A clear next step in the research will be to instantiate the model in the development of an actual data  strategy in a local government scenario and subsequently within the broader public sector. This will  provide an opportunity for further refinement of the model and the potential to develop the model  as a formalised toolset.