Data migration is more than just moving records from one system to another. It is, in fact, a challenging and time-consuming task that demands ample planning along with a good understanding of the previous system. Poor data mapping can lead to the eventual loss of important data and critical functionality. There can also be other, more unexpected effects of a careless migration on existing data and rules and all of this can become even more troublesome when you are migrating into a running instance that is already in use as opposed to a new one. Therefore, even when importing thousands of records, you have to be sure that each and every one of your records have been transferred.
The factors that need to be considered during data migration may vary depending on the complexity of your source data. However, for any data set, you can avoid having to constantly go back and forth to make sure all data is imported by following the standard protocol of planning, migrating and testing. While these steps are absolutely essential, it is equally important to avoid the following mistakes:
1 – Planning without involving stakeholders.
You might know how to go about a data migration but stakeholders are the ones who are fully familiar with the structure and flow of the data. Only they can tell exactly what kind of issues the data may run into during migration. You might need to make modifications in the fields or validation rules or even your overall approach accordingly.
2 – Importing related Custom Objects in Random Order.
You might be used to always importing accounts before contacts or opportunities as a standard that is reinforced everywhere, but taking care of the ordering of custom objects in accordance with the relationships they have with each other is also equally important. Along with the ordering of objects, you will also need to include the fields of each object into other objects multiple times to make sure that all the lookup relationships are set up.
3 – Importing large data in the Production Environment in the first go.
Even if you have planned the migration well and cleansed data beforehand, importing all data in the production environment the very first time is never recommended. It is preferable to migrate a small subset of your data in a sandbox so that all issues are exposed at an early stage. This will allow you to make any required fixes easily. You might have to repeat this step several times until all issues have been resolved.
4 – Mapping irrelevant data fields
While it might be necessary to add fields such as Legacy ID or User ID in your source data, getting rid of unnecessary fields that are used by only very few records is an important step in the mapping process.
5 – Forgetting to review active workflows and triggers.
Prior to a data import, you must check if you need to disable any active workflows or triggers that affect the objects in migration. It is possible that you end up sending hundreds or thousands of unwanted emails to customers as your data is uploaded if workflows or triggers are still active. You should also check to see if any validation rules need to be added, modified or deleted.
6 – Not testing in the end.
The successful execution of the data loader without errors is not a guarantee that all your data has migrated properly. A good way to test is by running reports or using the developer console to check the record counts. You can apply filters or perform soql queries to check the number of records for various objects.
We are Salesforce Consulting Partners with a team of highly professional CRM Experts that specializes in providing Salesforce Customization, Integration, Migration and Support services to other Partners. Get in touch for a Free Business Analysis.