Top 3 Things to Keep in Mind During Data Migration
Data migration typically accounts for more than 60% of the total hours in a database development project, yet it often receives the least attention from the client. Clients often think of data migration as being done during the night by "the man behind the curtain", while in reality, a good data migration requires a lot of input and engagement from the client. Because we are very much in the business of proper expectation setting, we find that with data migration, it pays to know what's going on behind that curtain.
To that end, we've created a list of the Top 3 Things to Keep in Mind during Data Migration.
#1: Physical movement of data
The first step in data migration is the physical movement of the data from one database to another (done by your consulting partner). The thing to note here is that it is the first step in the migration.
#2: Perfecting the migration
Perfecting the migration; elimination of any data migration defects. This phase consists of moving, reformulating and mass editing of the newly migrated data within the new database. This work is done by your consulting partner, at the direction of the client project team.
Data migration should not be complete until the client project team is satisfied with the results of #2. The purpose of #2 is NOT to critique #1, which WILL produce migration defects. Defects noted after the initial data migration should be anticipated, and the client project staff should have the expectation of working with their consultant to identify and correct the imperfections of #1. Once the client project team is satisfied, #2 is complete and general users are allowed into the new system.
#3: Ongoing adjustments
Ongoing adjustments of data validation rules and data translations. Again, it is imperative that users' expectations be set properly. When users begin using the new database, unanticipated issues will occur. Some will be part of the iterative design process: "we changed our minds about how we want the new system to handle X", as well as cases where data validation rules must be reconciled between new and old data. #3 is not a critique of #1 and #2, but it is necessary to a good migration.
Keep this list in mind and you'll be on the road to a successful data migration.
