At this stage of your project, you have explored your data inside and out. You spent hours analyzing it, answering questions about it, and cleaning it up. You have now spent more time with your data as a whole than ever before. You signed off on a data map and you scheduled a data migration. You informed your users about what’s happening and have put a freeze on the old system. The data migration has taken place and you have probably answered even more questions about things that presented themselves during that process.
Now comes the fun part: data validation! You have to go through your new system and make sure that the data looks good. Confirm all records are in the new system and that fields on those records are all correct. It may seem daunting because there are thousands of records to check. But don’t fear, we have a few tips (which are valid whether you are doing a sample migration or a live migration in the production environment) to help make this process easier.
1. Don’t freak out if you find issues!
I put this one first to stress that this is a work in progress. No one knows your data quite like you. Your data specialist will do his/her best to follow the map and deliver a flawless system, but if you do find issues, know that a data specialist will work with you to fix those issues. It’s going to be a back and forth process, similar to the UAT phase where you tested out the functionality and reported any issues that you found. The same applies in the data validation stage.
Create a couple of reports that list out some of the more important pieces of data. That way you’ll be able to see all relevant fields in one place. For instance, create a Contact report that pulls in all demographic information. You’ll be able to sort and see which records have empty fields. If you know that all contacts need to have an emergency contact associated to them, then a report is a good way to see which ones are missing. You can also check your data using filters. Occasionally, after a migration you realize that some contact first names come over as “Joe and Jane.” You can filter these out by creating a filter where First Name contains “and” OR First Name contains “&.” From there, you can decide whether to update them manually or update them in bulk.
You can also use reports to view sums. This is a good way to do a quick check to see if totals of records and sums of amounts add up to what you expect. If you are expecting 100,000 donation records and there are only 80,000, then you know something is amiss. If you are expecting $1,989,349.50 in donations, and you only have $1,234,543.00, then you also know something is off. Be careful and make sure that your expectations are correct. Perhaps you only imported Active contacts that have donations in the last 5 years. If this is the case, the totals won’t match in both systems. Make sure that you are comparing apples to apples.
3. compare specific records
You can’t possibly check every single record to see if it all came over properly, but you can check a dozen records. You know your data best, so a suggestion is to pick 12 – 20 records in your old system that you know have a lot of activity on them. Have your old system open to that record on one side of your screen and that same record in Salesforce open on the other side to do a sight comparison. Make sure that the demographic and contact information is all in there. Confirm that the related records look good (Relationships, Donations, Affiliations, Attachments, Notes, etc). If you find something off in one of those records, chances are that the same problem exists with other records. Copy the URL of the record and send that to your data specialist, so he/she can dig deeper to find and fix the issue.
4. Use the data map
Your consultant worked diligently with you on that data map. Use it now! If you aren’t sure if something was supposed to come over in the migration, consult the data map. If you aren’t sure where something lives in the new system, consult the data map. It’s called a map for a reason. Use it to navigate around the new system.
It’s important to play a role in this process since you know your data better than your consultant. Consultants will fix issues, but spending time now finding them is going to ensure that you don’t encounter the same problems when the project is complete. It’s not magic (even though sometimes it would be nice if it was), so work with your consultant and make data validation a painless part of the project.
You may also want to read:
Creating a Data Map
Steps for a Successful Data Migration
Best Data Practices