How to Dedupe and Standardize Your Prospect List
TL;DR: Prospecting teams should preserve the original file, identify duplicates with reliable identifiers, reconcile matches against the CRM, standardize fields without losing meaning, and check outreach eligibility before activation. Keep uncertain matches for review. Import matching, email verification, and suppression solve different problems; none proves that every remaining row is ready to contact.
What is the safest order for cleaning a prospect list?
Separate identity, formatting, and outreach eligibility before editing records. Identity asks whether rows refer to the same person or company. Formatting makes values usable. Eligibility determines whether the intended outreach should happen, given ownership, customer status, contact preferences, and current activity.
| Work | Output | Hold when |
|---|---|---|
| Preserve the input | Original file and source-row references | Source or intended use is unclear |
| Resolve duplicates | Reviewed identity matches and exceptions | Identifiers conflict or refer to shared contacts |
| Reconcile the CRM | Existing record, owner, and relationship context | Several records match or ownership is unresolved |
| Standardize fields | Normalized values alongside originals | A change would erase meaningful distinctions |
| Validate activation | Accepted contact data and applicable exclusions | Identity, contactability, or permission remains unresolved |
Use a working copy and keep a source-row identifier so you can explain every removal or change. The enriched CSV review workflow helps preserve original values while comparing proposed replacements. Do not make the first cleanup action an irreversible overwrite of the file or CRM.
How do you remove duplicates within the list?
Start with reliable identifiers and review conflicts before collapsing rows. An exact email match can identify records for comparison, but shared inboxes, aliases, and stale addresses require care. A display name alone is not enough to merge people, and a company domain does not distinguish employees at that company.
Keep the raw identifier alongside any normalized version. Trim unintended surrounding whitespace and apply a documented domain-format rule, but do not strip punctuation or email tags indiscriminately. A formatting change should not silently create an identity match that did not exist in the source.
When identifiers disagree, preserve both rows and record the reason for review. A matching name and company can suggest a possible duplicate; it should not automatically merge two people. Likewise, a job change can require an update to employment context without proving that every old contact value still belongs to the person.
- Collapse an exact repeated row only after checking that no meaningful fields differ
- Keep conflicting email, company, role, or ownership values visible
- Choose the surviving value using a stated source and verification rule
- Retain source references and the reason for each merge decision
- Hold ambiguous identity matches rather than increasing the apparent cleanup rate
How should you reconcile the list against your CRM?
Look for existing records and relationship context before creating new ones. Check the person’s identity, associated company, account owner, customer status, open opportunities, and active conversations. A duplicate check answers only part of the question: an existing record may need a different follow-up from a new prospect.
In Unify, HubSpot contact matching uses email and company matching uses domain. The integration does not merge existing HubSpot duplicates. If several records match, one is selected for an update; resolve the ambiguity in your CRM process rather than assuming the connector fixes it.
The Salesforce integration also checks for existing matches without merging existing duplicates. Its lead or contact creation behavior depends on the configured Write people as setting and sync rule. Check which object the team expects before importing the same person through another route.
Keep account ownership and permission to engage separate. Engagement policies can govern claiming and contacting independently, using a configured CRM ownership field. A clean contact record does not give every rep permission to work the account or change its owner.
How do you standardize job titles and company names?
Preserve the original value and add a normalized field for the specific decision you need. Keep job title, seniority, and function separate. Do not turn different roles into the same title just because they belong in a similar prospecting segment.
For company records, distinguish a display name from the domain and the legal or CRM entity. Review subsidiaries, brands, and regional entities before grouping them. If multiple records should remain distinct for ownership or contracting, a cleaner-looking company label should not erase that distinction.
Write the normalization rule, keep an exception list, and check the output before using it in personalization. A segment label may be appropriate for filtering but awkward or inaccurate in an email. Test the actual value that will reach the recipient rather than assuming a standardized column is ready for every use.
In Unify, review and refine a table in chat before saving a list. Inspect supporting evidence when an enriched cell provides it. Keep the accepted rows separate from unresolved rows so a cleanup request does not become an implicit instruction to activate everything.
What should you check before uploading the cleaned file?
Check matching and write behavior at the import boundary, not only in the spreadsheet. Unify’s CSV import matches companies by domain and people by email, and mapped values can update an existing record. Review the preview, destination fields, enrichment choices, and warnings before starting the upload.
People need a valid email, either supplied or found through enrichment; company rows need a valid domain. Required custom fields also affect whether a record can be added. Providing inputs for enrichment does not guarantee a result, so compare the completed upload with the expected accepted records before outreach.
Review CRM field mappings separately. They control read and write behavior for mapped CRM syncs, including choices that preserve or replace existing values. Those settings should not be treated as a blanket guarantee that every CSV import leaves populated Unify fields untouched.
For a recurring integration, the Data API supports querying, creating, and modifying records. Define the matching, update, and exception policy for that integration explicitly. An automated import is only as controlled as the rules applied to each incoming record.
How do verification and suppression differ?
Verification assesses the contact data; suppression determines whether outreach should proceed. A deliverable address can still belong to someone who opted out, an existing customer who needs a different workflow, or an account another rep is already working. Recheck the applicable restrictions at activation.
Unify’s email verification settings block invalid addresses and let you configure catch-all, unverified, and role-based addresses. Do not treat every role address or catch-all domain as permanently unusable. Use a documented policy for the specific motion and review the decision to verify, hold, or exclude catch-all addresses.
Use sequence exclusions for the groups the prospecting motion should not contact. Customer exclusions can apply to new-business sequences while remaining disabled for appropriate customer follow-up. Applicable exclusions are checked at enrollment and continue for active enrollments, but a CRM change must reach Unify before it can affect the check.
For users assigned to an engagement policy, its exclusions replace sequence-ruleset exclusions; other policy rules still apply. Review the user’s policy as well as the sequence. Do not assume a familiar ruleset name proves every rep is governed by the same restrictions.
How do you keep the next import from recreating the problem?
Make the cleanup decisions repeatable without hiding the exceptions. Record source, identity key, normalization rule, accepted field changes, exclusion state, and review owner. Route uncertain records into a review process rather than forcing a match to keep an automation moving.
- Compare expected accepted records with the actual saved list
- Investigate skips, conflicting matches, and unexpected updates
- Confirm that account ownership and restrictions survived the handoff
- Keep rejected or held rows with their reasons
- Review recurring source defects before accepting the next batch
Measure quality using accepted identities and correct updates, not a target percentage of rows removed. A list with little duplication may need few removals, while a heavily repeated source may need substantial consolidation. Neither result establishes email performance or pipeline impact by itself.
To inspect, refine, and import a reviewed prospect list, try Unify.
Frequently asked questions
What key should I use to dedupe people?
Start with reliable identifiers and inspect conflicts. Exact email matches can support comparison, but shared or stale addresses need review. Do not merge on display name or company domain alone.
Does Unify automatically merge CRM duplicates?
No. Its CRM integrations check for existing matches when writing supported records, but they do not merge existing duplicate CRM records.
Should I replace original job titles with normalized titles?
Preserve the original and add separate normalized fields for segmentation. Avoid erasing differences in function, seniority, or actual role.
Can a CSV upload change an existing Unify record?
Yes. CSV matching uses company domain or person email, and mapped values can update an existing record. Review the preview and field mappings before uploading.
Should I delete all catch-all or role-based addresses?
Do not make that a universal rule. Define the permitted address categories for the outreach, verify what can be verified, and hold uncertain records when necessary.
Is a clean list ready to send immediately?
Not necessarily. Confirm identity, contact data, current ownership, active conversations, and applicable outreach restrictions before activation.
Glossary
- Deduplication: Resolving records that represent the same entity
- Normalization: Applying a consistent representation without losing meaning
- Identity key: A value used to match a person or company
- Suppression: A restriction that prevents the intended outreach
- Catch-all domain: A domain that accepts addresses that may not correspond to a real inbox

