Organisational Maturity Assessment during the Paradigm Shift from Monoliths to Data Mesh

Abstract

Incorporating big data into decision-making provides a substantial competitive advantage, leading organisations to increasingly adopt a data-driven strategy. However, the adoption by organisations often remains unsuccessful due to limitations associated with monolithic data architectures, such as data lakes and data warehouses. Data mesh is introduced as a decentralised socio-technical approach to alternatively manage data, aiming to overcome the limitations and gain the benefits of embracing a data-driven strategy. However, there is a lack of guidance on how to implement data mesh. The availability of generic and concrete data mesh implementation steps, including a maturity assessment, would be helpful for organisations. Consequently, this research proposes the design of a Data Mesh Maturity Assessment Model (DMMAM). In response to the main research question: ”How to assess the maturity of a data mesh implementation within an organisation?”, enabling the assessment of how mature a data mesh implementation is, by means of the DMMAM, would provide the guidance that is currently lacking for organisations. The qualitative Design Science Research Methodology is employed to structure the design process. Literature research, interviews, and cases are conducted to explore the contribution of, as well as design, demonstrate, and evaluate the DMMAM.

This research shows that the developed DMMAM evaluates data mesh based on four maturity levels, classified as Level 0: Non-Initiated, Level 1: Conceptual, Level 2: Defined, and Level 3: Achieved, and that data mesh is represented by five dimensions: A. Data Foundation & Organisational Change, B. Domain Oriented Decentralised Data Ownership & Architecture, C. Data as a Product, D. Self-Serve Data Infrastructure as a Platform, and E. Federated Computational Governance. These five dimensions are collectively represented by 54 characteristics. For each characteristic, labels for the People, Process, Technology (PPT) perspectives are assigned. Additionally, questions are formulated, and criteria and requirements are provided for all characteristics at each maturity level. This enables participants to self-assess their organisation’s maturity by individually rating 54 questions based on the current and target levels. Conducting the self-assessment yields various outcomes, including an overall data mesh maturity score, individual dimensional maturity scores, and maturity scores from PPT-perspectives. Moreover, the assessment helps to identify maturity gaps and allows benchmarking to compare results across organisations, providing organisations with guidance for improvement. The demonstration and evaluation of the DMMAM through maturity assessments for three organisations have demonstrated its applicability and usefulness. However, it is important to acknowledge that this research represents the first attempt to provide a comprehensive framework for assessing data mesh maturity in organisations and is not without limitations.

Future research is proposed to further refine and improve the DMMAM, supported by data mesh SME’s and practitioners, to ensure that the model remains up-to-date with the latest available research on data mesh. In addition, including additional guidance as an outcome of the maturity assessment would make the assessment more actionable and pragmatic. Furthermore, examining the optimal assessment structure will enhance the model’s reliability and validity. Moreover, expanding the benchmark functionality will enable statistical generalisations and comparisons for organisations within and across industries. At last, it is suggested to do further research about examining the overall contribution of data mesh as a strategy element towards becoming data-driven.