Running an Import Wizard load is a guided sequence; the steps below cover the standard path plus the preparation that prevents the most common failures.
- Prepare the CSV
Export current data if updating, or build the CSV from your source system. Confirm column headers align with Salesforce field labels for easier mapping.
- Test on a sandbox
Run the import in a sandbox first with a small subset of records. Confirm the field mapping works and the validation rules pass.
- Launch the wizard
Setup > Data > Data Import Wizard > Launch Wizard. Choose the object type (Accounts and Contacts, Leads, custom object).
- Configure the import
Pick the operation (Add, Update, Add or Update), upload the CSV, choose the duplicate handling rule. The wizard guides through this step-by-step.
- Map fields
Confirm or override the field mapping. Required fields must be mapped; unmapped non-required fields default to blank or existing value.
- Review and start the import
The wizard shows a preview and record count. Confirm and start. Imports usually complete within minutes for sub-50K runs.
- Review results and handle errors
After completion, check the success count and download the error CSV if any. Fix the errored records, re-import as a second run.
Insert new records only. The simplest operation for first-time loads.
Update existing records identified by Salesforce ID or external ID. Source records must already exist.
Update if matching, insert if not. Requires a matching key (external ID or email).
Wizard-specific matching for Accounts/Contacts and Leads. Distinct from org-level Duplicate Rules.
Map source columns to Salesforce fields. Guesses based on header similarity; override as needed.
- 50,000 record cap per run. Larger imports need to be split or moved to Data Loader.
- Duplicate Rules fire during wizard imports. If a Block rule applies, the wizard records the failure; plan rule activation around the load.
- Mappings are not saved between runs in most contexts. Recurring imports benefit from Data Loader with saved mappings.
- The wizard runs synchronously. Browser closing or session timeout during a long import can cancel the load; for safety, run on a stable network.
- Not all objects are supported. Cases, Opportunities, and many advanced objects need Data Loader or API. Check the supported list before starting.