Best Practices with Any Data Loader

You may not want or need to migrate a whole set of data—choose which objects you wish to migrate. For example, you may want to migrate only the contact information from each account, or only migrate account information from a particular division.

Create one template for each object, for example in an Excel worksheet.

Identify the required fields for each object. In addition to the required fields for each standard object, there may be additional required fields such as those needed to follow business rules, or legacy ID fields. Use this guide or see the page layout definitions in the Salesforce user interface to find out which fields are required on standard objects.

You may wish to highlight the required fields in red for easier review of the data after you populate the templates.

You should also identify any ordering dependencies. Objects may have mandatory relationships, for example all accounts have an owner, and all opportunities are associated with an account. The dependencies in these relationships dictate the order of data migration. For Salesforce data, for example, you should load users first, then accounts, then opportunities.

To identify dependencies, review the related lists and lookup fields in the page layout of the given object, and IDs (foreign keys) in the database.

Clean your data before populating the template, and review the data in the templates.

Create custom fields to store legacy ID information. Optionally, give the custom field the External ID attribute so it will be indexed. This will help maintain relationships, and help you build custom reports for validation.

Load one record, check the results, then load all records.

Use all of these techniques to validate your migration:

• Create custom reports that validate record counts and provide an overall snapshot of migration.
• Spot check the data.
• Review exception reports to see what data was not migrated.