Data migrations often fail because they are attempted with traditional methods
Custom programs and packaged ETL solutions are the most common methods used to migrate large volumes of data. Both of these efforts require some degree of development. In the case of a custom program, the entire solution is developed from the ground up. In the case of an ETL solution, the various interfaces must be created and then transformation logic is added to them. Both of these activities are typically completed using traditional software development methodologies. But traditional methods often fail when they are applied to data migrations. Why?