Five key factors for Data Migration Success
Authored by Ashvinder Rana, Data Migration Lead, Utopia, Inc.
Many of us have been a part of various data migration projects at some point or the other. We all know data is important. But can we name at least five key factors that will ensure a successful go live? Well, here is my two cents (although I believe it’s worth a lot more ). Though, I know there are more than just these five key factors the following are my top five.
- Understanding Data – it is absolutely imperative for the data team to understand the legacy data and how it exists in the legacy systems. Along with how the data is structured in the legacy systems.
- Resource Scheduling – now of course, you want the right resources with the right skill sets and knowledge to do your data migration project! Imagine, calling a plumber to fix your AC at home!! I’m sure we would never think to do that. Just the same, we should ensure the data team houses resources with appropriate data skill sets as required for the project.
- Scoping the Requirements Accurately and On Time – not too early, not too late. You know how the saying goes: “there’s a time and place for everything!” You don’t want to scope out the data requirements before the business blueprint requirements are signed-off on. If you do, then just be prepared for those “change orders” to start rearing their ugly heads well into the beginning of the development phase. Unfortunately we’ve all run into these way to frequently!!
- Data Quality Framework – ensuring that a data quality framework is in place is equally important among other things that will lead to a successful data conversion. This implies not just having means/tools to perform data profiling and analysis, but also a plan to identify data bottlenecks and recommended solutions and/or data cleansing plan of actions as well.
- Data Validation vs. Data Testing Strategy – projects need to ensure an effective data migration testing strategy is in place and is part of the data migration project. The clear demarcation should be also be made between data validation and data migration testing. Wherein, data validation can comprise random sampling methods to ensure that the data is converted accurately as per the data mapping rules. However, a data migration testing strategy should comprise a series of iterative “mock conversion runs” for all objects in scope where the converted data is utilized by the business process / functional teams to thoroughly test the integration points/transactions as well. In addition, these iterative “mock conversion runs” also allow for validation or conversion programs, conversion error analysis and fixes that will eventually lead to a “zero-error” data migration!