Shark Tank

28.12.2005
Big telecommunications company is building its first enterprise data warehouse, and things start out looking very, very good, reports a pilot fish who's on the project.

"The hardware vendor has provided the implementation team as part of the 'solution' agreement," fish says. "The prototype is completed and is received enthusiastically by the client IT department and also with rave reviews by the client's business users.

"Rigid protocol, however, insists on an independent quality-control review of what has been delivered so far before authorizing the next stage of the project. That is where I come in, as the person responsible for this independent QA review."

So fish begins drilling into the prototype. And what he finds isn't nearly so pretty.

"The heart of an enterprise data warehouse is its data model," says fish. "But one week before delivery, the project manager, in a fit of panic, discovered that they did not have one.

"So she and the lead technician, neither of whom had any significant expertise in either the telecommunications industry or data modeling, put together a data model -- in a single day!"