Great conversation. MDM is one of the most misunderstood first principle of data management and value by both business and IT. Recently the need amplified by AI context. No coincidence that SAP buys Reltio, SF buys Informatica… With all the positive talk about reference data during the conversation - table driven, cross walk, translation, standards, AI hard coding… Don’t recall the actual intentional sibling of MDM, “RDM” - “Reference Data Management”! The ownership/stewardship of this information in large organizations with many systems, silos, and regions is foundational! Basic concept, and a gift that keeps on giving. Not just paired values of codes and descriptions, but the different contexts and built in groupings - context and taxonomy. ID, Code, Short description, long description, system translations, mastering across systems.
100% RDM is often the forgotten step-child until an enterprise gets to standing up their semantic layer or feed context to a LLM and all of a sudden half their logic is trapped in hardcoded DAX and case statements segmented by siloed orgs.
I'd love to hear about how you are approaching this in your work!
Great conversation. MDM is one of the most misunderstood first principle of data management and value by both business and IT. Recently the need amplified by AI context. No coincidence that SAP buys Reltio, SF buys Informatica… With all the positive talk about reference data during the conversation - table driven, cross walk, translation, standards, AI hard coding… Don’t recall the actual intentional sibling of MDM, “RDM” - “Reference Data Management”! The ownership/stewardship of this information in large organizations with many systems, silos, and regions is foundational! Basic concept, and a gift that keeps on giving. Not just paired values of codes and descriptions, but the different contexts and built in groupings - context and taxonomy. ID, Code, Short description, long description, system translations, mastering across systems.
100% RDM is often the forgotten step-child until an enterprise gets to standing up their semantic layer or feed context to a LLM and all of a sudden half their logic is trapped in hardcoded DAX and case statements segmented by siloed orgs.
I'd love to hear about how you are approaching this in your work!