How do you deduplicate HubSpot?
Deduplicating HubSpot works in five steps: agree what counts as a duplicate, find the patterns that create them, choose the primary record by rule, merge in tested batches while protecting consent and ownership, then fix the sources so duplicates stop coming back. Merging alone is a one-off; prevention makes it last.
Data cleanup should not be a one-time spring clean. This is the system we use on our HubSpot cleanup projects: a clear definition of a duplicate, a safe merge process, and prevention rules that stop the same problems returning.
- 01 DefineMatching rules agreed in writing
- 02 Find patternsSnapshot and group likely matches
- 03 Choose the primaryBy rule, not case by case
- 04 Merge in batchesTested, logged, values copied first
- 05 PreventStable IDs, normalized entry, ownership
What counts as a duplicate in HubSpot?
A duplicate is two or more records that represent the same real person, company or deal. Agree the matching rules in writing before touching data: exact email for contacts, domain for companies, and a stable external ID wherever another system is involved. Also agree what is intentionally separate, such as subsidiaries.
- Contacts: the same email, or the same person with a personal and a work email.
- Companies: the same domain, or the same company entered with and without a domain.
- Deals: the same opportunity created twice, often by an integration and a rep.
Step 1: How do you find duplicate patterns before merging?
Start with a snapshot, not a merge. Export the records, keep the original record IDs, and group likely matches by email, domain, phone, normalized name and external ID. Then segment them by source, owner, lifecycle stage and last activity, so you can see which import, form or integration is creating them.
- Export a sample and preserve the original IDs.
- Group likely matches and give each a confidence level: certain, likely, or needs review.
- Classify records as duplicate, incomplete, conflicting, or intentionally separate.
- Record the source of each duplicate group; it tells you what to fix in step 5.
Step 2: Which record should be the primary?
Decide the primary record by rule, not case by case. Prefer the record with the correct owner, the most complete activity history and the right associations. When HubSpot merges, the primary record's values win for most properties, so make sure the primary holds the values you want to keep.
| Priority | Rule |
|---|---|
| 1 | Record linked to an open or closed-won deal |
| 2 | Record owned by the current account owner |
| 3 | Record with the most activity history |
| 4 | Record created by the system of record (for example, the billing system) |
| 5 | Oldest record |
Step 3: How do you merge safely?
Merge in small, tested batches. Before each batch, copy consent, lifecycle stage and owner values onto the primary record so nothing important is overwritten, and keep a log of every merge with both record IDs. A HubSpot merge cannot be undone, so the log is your only way to explain a change later.
- Set a system-of-record priority by object and field.
- Protect marketing consent and subscription status above everything else.
- Test the rules on a small batch, check the results, then scale.
- On Professional and Enterprise plans, work through HubSpot's duplicate suggestions pair by pair. Bulk review and custom matching rules need Data Hub Professional or Enterprise.
If duplicates keep coming back, the system creating them needs fixing.
Step 4: Why are company duplicates different?
Company duplicates usually come from missing or inconsistent domains, and they carry more weight than contact duplicates because deals, tickets and revenue reporting hang off them. Fix the domain first, merge companies before their contacts, and check that every deal and ticket ends up associated with the surviving company.
- Normalize domains (strip "www.", protocols and paths) before matching.
- Decide how parent and child companies are represented before merging subsidiaries.
- Re-check deal and ticket associations after each company merge.
Step 5: How do you stop duplicates coming back?
Prevention is where the return on a cleanup comes from. Make integrations update records on a stable external ID instead of creating new ones, normalize key fields on entry, control who can import and how, and review new suspected duplicates every week so a new source is caught early.
- Require integrations to upsert on a stable external ID where possible.
- Normalize phone numbers, countries, domains and key dropdowns.
- Allow imports only with a matching key, by trained users.
- Run a weekly exception report for new duplicates and failed syncs.
- Give users a simple way to flag a suspected duplicate.
The knock-on effects of skipping this step are covered in what duplicate data in HubSpot actually costs you.
Cleanup and deduplication checklist
Key takeaway: Do not measure cleanup by records merged. Measure the result: fewer routing errors, cleaner attribution, faster search, and reports people trust.
Frequently asked questions
Does HubSpot deduplicate records automatically?
Partly. HubSpot uses email address to match contacts, so a form submission or import with an existing email updates that contact instead of creating a new one. It can also match companies on domain. Records with different emails or no domain still create duplicates. On Professional and Enterprise plans, the duplicates tool suggests likely pairs to review; bulk review and custom matching rules need Data Hub Professional or Enterprise.
Can a HubSpot merge be undone?
No. A merge in HubSpot cannot be reversed, so choose the primary record carefully, protect consent and ownership values first, and test the rules on a small batch before merging at scale.
How long does a HubSpot data cleanup take?
On our projects, our fixed-price CRM Cleanup takes two weeks; larger portals take longer depending on record volume, the number of objects involved and how many systems keep creating new duplicates.
What causes most duplicates in HubSpot?
The usual sources are imports without a matching key, integrations that create records instead of updating them, people with several email addresses, companies without a domain, and manual entry that skips search. Find the source before merging, or the duplicates come back.
Sources
HubSpot documentation behind the feature and tier notes above, checked October 2026. HubSpot changes features and tiers often, so check your own portal too.