Data hygiene

Clean data requires a working system that runs beyond a cleanup day. Start with the account record, resolve ambiguity before it reaches a list, and run a repeatable check between every new record and every outbound send. Suppression is a sending control, so it can stop a send while the underlying marketing list stays dirty. Keep both under control. A target account list built on duplicated or inconsistent account records is unreliable, and the CRM layer needs attention before the list does.¹ Suppression lists remain useful; marketing lists still need cleaning.²

Run the hygiene loop

1

Set the standard

2

Reconcile records

3

Refresh

4

Suppress

5

Prevent

1. Set the standard
2. Reconcile records
3. Refresh
4. Suppress
5. Prevent

Run hygiene in a fixed order. Resolve the record, test its freshness, decide whether it should be contacted, then add a check that catches the same problem next time.

Stage	What you are trying to learn	Example question
Set the standard	What a usable record must contain and how fields should be written	Which fields must be present before this record can enter outbound?
Reconcile records	Whether several records describe the same account or contact	Does this record describe an existing account or contact?
Refresh	Whether the details are still current	What should be rechecked before this record enters outbound?
Suppress	Whether the record is eligible for contact	"Is the record already in your CRM or suppression list?"3
Prevent	Which process catches the same issue next time	Who owns the next review?

Set the standard

Start with the CRM structure that feeds your lists. Unclear rules turn cleanup into personal judgment, and the records drift again.

A mailing list is only as clean as the CRM supporting it.⁴ Documented data rules help maintain long-term accuracy across marketing, sales, and operations.⁵ Write down what counts as a complete record, which values are allowed, and what happens when a field is unknown.

Use the standard at intake and before export. A rep should be able to understand the account, identify the contact, and see why the record is eligible for outreach without checking several unrelated records. Another person should be able to apply the rule and reach the same answer.

Reconcile records

Deduplication comes before segmentation and sending. Work at the account and contact level, and preserve the record with the most useful history.

Deduplicate files with several kinds of matching logic.⁶ Remove duplicate or old records as part of CRM data hygiene.⁷ Compare the fields that identify the company, the contact, and the activity history before merging anything. Keep a clear record of which entry survives so the team can trace the decision later.

Check company relationships separately. Company relationships can be missing, incorrect, or out of date.⁸ Look for parent, subsidiary, and duplicate account records that could send the same message to several versions of one business. The pass is complete when each account has one working record, its relevant contacts sit in the right account, and the history has not been split across duplicates.

Refresh the records

A clean record can become stale after its first review. Refreshing means testing the fields that can change and setting a repeatable interval for records you continue to use.

Enrich the source of truth with accurate and fresh data.⁹ Treat purchased data as provisional because it becomes more inaccurate as time passes.¹⁰ For accounts in active outbound, use a bi-weekly bounce-rate review and a quarterly list refresh.¹¹

Keep phone records current as well. An updated contact list reduces the chances of calling numbers that may report calls as spam.¹² Bounce activity, failed calls, changed roles, and missing contact details all signal that a record needs review. Finish with a recent verification decision and a clear next review date for each record.

Apply suppression

Suppression sits between a usable record and an outbound action. Use it to prevent contact when the record fails your eligibility rules, while keeping the source list clean for the next campaign.

Create a dynamic list for cold prospects, apply the additional criteria, and mark matching prospects as "do not email".¹³ Put dirty, duplicate, non-responsive, or non-consensual records into a review path before they reach a send.¹⁴ Keep the suppression decision visible on the record so another rep does not re-add it from an old export.

Check the CRM and suppression list before every campaign. Look for conflicts between an active segment and a prior suppression decision. Send only when the record has one contact status, one campaign eligibility decision, and no unresolved reason to withhold outreach.

Prevent the next cleanup

A cleanup pass clears the backlog. Prevention keeps new records from recreating it, so assign ownership to the steps that happen before a record enters a list.

Cleanup is necessary, and prevention maintains a clean CRM.¹⁵ Data quality work requires correcting anomalies at the source and carrying out periodic maintenance afterward.¹⁶ Put processes in place that make sure the data stays good.¹⁷

Hold sales and marketing teams accountable for keeping and maintaining the database.¹⁸ Sales reps should keep opportunity stages, amounts, and close dates updated, while sales operations should remove duplicate or stale deals.¹⁹ Give each intake, merge, refresh, and suppression decision an owner. When the same error appears more than once, review the rules and fix the entry point instead of adding another manual patch.

What not to do

These mistakes leave the list looking usable while its underlying records keep weakening.

• Do not send a campaign before manually cleaning the list.²⁰
• Do not accept a procured list without checking that duplicates and obsolete data have been removed.²¹
• Do not treat CRM data quality as a one-time project; it requires ongoing standards, measurement, tooling, and organizational commitment.²²

Use the loop before the next export and give every exception a visible decision. You can then see which records need correction, which need suppression, and which are ready for a controlled outbound send.