Integrating Dimensional Fact Modeling (DFM) and Decision Model and Notation (DMN) for Complex Business Logic in BI Systems: A Case Study at Technogym

Conceptual models for data warehouses, such as the Dimensional Fact Model (DFM), describe what data is stored and how it is structured, but give no signal that some attributes are governed by volatile business decision rules. Conversely, the Decision Model and Notation (DMN) standard formalizes such rules but is agnostic to dimensional structure, grain and additivity. As a result, decision logic is typically embedded in procedural SQL, disconnected from the conceptual layer. This thesis proposes an integrated DFM-DMN methodology, formalized as the Decision-Enriched Fact Table (DEFT) framework, which extends the classical DFM fact schema with Decision Nodes mapping conformed dimensional attributes to measures. Following a Design Science Research approach, the methodology is applied through a non-linear three-stage cycle (forward requirements design, physical implementation, retrospective formalization) and illustrated on an Order-to-Service scenario inspired by SAP-driven enterprise environments. Two examples are developed: AccountingStatus, a multi-step decision chain reconciling line-level and order-level grain, and CreditStatus, a forward-designed decision with a Priority hit policy. An additivity matrix for multi-currency measures complements the conceptual schema. The thesis also outlines recommendations for adoption: a Dataform implementation blueprint, Data Catalog metadata integration, scaling of the DFM layer through reverse engineering, a phased roadmap grounded in diffusion-of-innovations theory, and a RACI governance model. The methodology has not been empirically validated on a production system; its benefits are argued analytically, and validation is left to future work.