A CRM upload is the point where small inconsistencies become operational. A duplicate creates another account. A malformed value can disrupt routing. An inconsistent category can make segmentation unreliable.
1. Confirm the entity
Decide whether the row is a company, contact, location or another object. Define how subsidiaries, branches and duplicates should be handled.
2. Check company names and domains
Normalize names using the project's canonical rules. Review domains for obvious errors, redirects and cases where the current website differs from the original source.
3. Apply duplicate logic
Exact text matching is rarely enough. Use identity keys such as normalized company name, domain, geography and contact identity when appropriate.
4. Validate required fields
- Company identity and website
- Contact identity and role where applicable
- Geography
- Project-specific classification and status fields
5. Normalize values
Industries, functions, countries and statuses should follow controlled values when the workflow depends on filtering or automation.
6. Review contact fields
Check email formatting, phone formatting, job-title consistency and duplicate contacts. Use the project's agreed country-aware phone convention.
Run the same sequence every time
Identity → duplicates → required fields → formatting → normalization → contact checks → status → final spot review.
Working with a large B2B dataset?
DBI supports validation, cleansing, duplicate handling, normalization and research-led QA.
Review Data Validation →