Attribution data becomes unreliable when different people use different names for the same source. Cleanup starts with a small controlled vocabulary and a record of how each value is assigned.
Inventory current values
Export or inspect the source field and group spelling, capitalization, and naming variants. Do not overwrite raw evidence before preserving a copy. Mark values that are genuinely unknown instead of forcing them into a convenient category.
Define the vocabulary
Use names that reflect decisions the team can make, such as organic search, referral, partner, outbound, or event. Document the rule for assigning each value. Keep campaign detail in a separate field when it has a different owner.
Clean carefully
Map obvious variants to the approved value, then review ambiguous records manually. Check that the cleanup does not alter lifecycle stage or ownership. The marketing analytics setup guide can help align source fields with reporting needs.
Add a prevention step
Make the approved options available in forms and CRM views. Review new unknown values weekly. Clean data stays clean when the process prevents new variants instead of relying on another large repair.
