Data migrations occur for many reasons - acquisitions, mergers, demergers, implementing new systems, data cleansing activities, and for them to be successful they must be well thought out, thoroughly tested, executed with precision timing and supported by the highest level of management.
Moving data from one system to another is conceptually quite straightforward, which is one of the main reasons things go wrong. Having participated in dozens I have experienced first-hand those that are successful and some appalling disasters. The worst I saw was the combination of modifying the existing data architecture and at the same time updating the data to make it fit the new structures.
I was not actually involved in that data migration as my involvement had been the reengineering of a collection of data cubes and reports, which I had thoroughly tested against the proposed architecture.
A successful data migration also needs to be considerate of the source and target data types, the limitations of the target system and each and every step of the transformation needs to be logged.
Having been in this industry since the time of mainframe processing, when every single change was meticulously tested, I continue to follow these practices ensuring nothing is overlooked. Having worked on many migrations I have only ever achieved success in my involvement.
If you are considering undertaking a migration we can certainly help you make it a successful one.