Tuesday, April 12, 2005

"The Great Divide": Standardized Definitions in Enterprise Analytics Applications

One of the challenges in building an enterprise analytics application is the "great divide" that exists between the definitions of business terms. A 'billing fee' might be the incarnation of different measures to different people depending on the sub-stream they are working in, and it becomes the job of the analytics architects to decipher the appropriate definitions. The great divide becomes even deeper when the technical team of architects and developers, often brought on to work on the analytics project on a temporary basis, are too aligned with the technology rather than the client's business to know the subtle differences in the definitions. A well-laid metadata layer can alleviate the divide, although again it depends on the technical and functional teams to hammer out the exact definitions in the analysis and design phases. Often the confusion in definitions stems in the functional teams themselves, when two teams are in fact referring to the same entity having the same definition in slightly different ways. Thus it might be that two layers of analysis is required: the first between the technical team and each of the functional teams to identify the attributes and measures that need to be defined in the analytical application, an internal assessment by the technical team to analyze possible overlap in definitions, followed by a cross-team joint effort to seek clarifications and standardization of the definitions. It might be worthwhile to dedicate substantial time to the cross-team effort to iron out the inefficiencies due to conflicting and overlapping definitions at the project outset, rather than followed a siloed approach to the development effort as far as the definitions are concerned.