Data Migration Plans
If you are migrating from database system to another database system, it is a delicate operation. You need to study the record layout of the current system, you need to study the record layout of the new system and identify the gaps between the formats. Often we need to build external tables that have a record layout that accommodates fields from both systems.
Now comes the fun part… setting up tags and record locators in harmony with new system, yet preserving the data from the old system!
Before importing the data in the new system we run the data through a battery of cleansing routine and update the data with now current information. We remove people who passed away, we change addresses of people who requested a change of address with US Postal Service, we tag people as ‘do not call’ if they registered with Do Not Call registries and do other deep data cleansing work depending on what is appropriate.
The biggest issue is data truncation. Unnoticed data truncation cause irreversible data loss.
The delicate part is doing all this in one swift movement because the business is going on while the data is being processed. The entire docking and undocking is done on a pre-determined schedule and you have only one shot at doing it right.
Things don’t always go as plan. Often complications develop and project takes life of its own. Once we dig deeper we run in to things that we did not take in to consideration or data problems that were present in the old system.
We have a process we follow. First we draw up old system and new system on mind maps. Then we create a workflow and a task order for each step. We take sample data and go through a dry run of migration. We take detailed notes during the rehearsal, discuss complications and adjust the workflow to accommodate the complications.
When the real migration is scheduled we have two people work on the project together, one person just watching activities of the other person. We use our process diagram as a check list. The process buddy watches the job being processed while the data processor runs through the process. Parthiv builds his own QC checks during the dry run and he reviews the work at each step to make sure everything pass through his QC process. When something does not pass QC we roll back and start over.
The process can be challenging, frustrating and nerve racking. We ask for your attention, cooperation, assistance and patience during this process. Our goal is to see the process go through without a hitch and do it right the first time so we can deliver on time and on budget.
We are change management experts and we are here to assist you through the transition. The journey might be bumpy, but you will be happy to arrive in a brand new data world.
