Implementation Thoughts – Data Migration

Ok, you’ve done your homework, gone through the software selection process, and selected the system that is right for you. Now the software contract negotiation phase begins, and one of the things you’ll need to review is the statement of work. One very important part of this is data migration. What, and how much data should I bring over to the new system? This should be spelled out in detail so all parties have the same expectations. Here are some factors to consider when making data migration decisions.

System History:
How long has the current application been in service? If you have the institutional knowledge, try to find out how that implementation went. Have there been any events that could have compromised your data, i.e. system mergers, the importing of large data sets, large operational changes, or anything else that may have had an adverse effect on your data’s quality? These factors can help you to determine a “data cutoff” date based on evolving business practices that made earlier data inaccurate or even just less valuable.

 
Existing Data Quality:
Remember the old adage; garbage in, garbage out? Well, that is especially true here. It is critical that your new system be implemented with good data. Whatever data you choose to bring over will probably require an extensive clean-up process. This is also a good time to perform simple maintenance on the data; for example, do you really need to bring over the 10,000 inactive vendors that you have not done business with for over 5 years?

 
Costs:
In general, the more data that you want to bring over, the more it will cost. There is a good chance that the data structures are different between the old and new applications. Data migration requires lots of planning, extensive field mapping, testing, etc. This all can escalate the implementation costs quickly.

 
Compliance/Legal Requirements:
This varies greatly based on domicile. Check with your legal team to reconcile what kind of data you are bringing over vs. how long you might want to or need to keep it. Keep in mind that there may be instances in which you choose specifically NOT to keep certain data.

 
Operational Needs:
With financials, you may choose to bring over only recent summary data into the new application, while keeping summary data in a data warehouse. For CRM applications, it is usually preferable to bring over as much clean and accurate historical data as possible.

 
Speed of Implementation:
The more data you choose to migrate equates to more time and effort to complete the implementation.

Remember that not all data is worth keeping, and the decision to migrate data you do want to keep involves performing a cost/benefit analysis to ensure that it is worth the effort.