How to Clean Up HubSpot Data: A Step-by-Step Guide
A step-by-step HubSpot data cleanup: audit, back up, fix properties, standardize formats, merge duplicates, clean lists and deals, and set guardrails so the mess doesn’t return.
By Krishnanshu Jaiswal, Revlyn ·
HubSpot data cleanup is the process of auditing your CRM, backing it up, fixing properties and formats, merging duplicates, filling gaps, tidying lists and deals, and putting rules in place so bad data stops coming in. Done in that order, it turns a CRM people work around into one they trust.
You probably know the symptoms. The same company appears three times with different spellings. Half your contacts have no owner or lifecycle stage. Reports disagree with each other, lead scores look random, and reps keep a private spreadsheet because they don’t trust the CRM. Almost every other RevOps project, from forecasting to attribution to automation, depends on getting this right.
This guide walks through a practical cleanup sequence, with the HubSpot tools that help at each step.
Before you start: define clean and protect yourself
“Clean” means different things to different teams, so decide what you’re fixing. Pick the ten or so fields that drive decisions, such as lifecycle stage, owner, lead source, industry, company size, country, email, deal amount, close date, and closed-lost reason. Then name one owner for the project and agree with sales and marketing on a quiet window, so people aren’t editing records while you merge them.
Also be careful with irreversible actions. In HubSpot, merged records cannot be reverted, so treat every merge as final. You can roll back rejected duplicate pairs for a limited time, but not merges. A backup is your safety net.
Step 1: Audit what you have
Start with HubSpot’s data quality tools, found under Data Management, then Data Quality. The overview page summarizes duplicate issues, formatting issues, and enrichment gaps, and the property insights show where properties are unused, empty, or duplicated. Set up the weekly data quality digest while you’re there, so changes show up in your inbox. Then build your own scorecard.
| Check | How to find it | Why it matters |
|---|---|---|
| Duplicate contacts and companies | Manage duplicates tool, plus exports | Splits history, inflates counts, breaks reports |
| Contacts with no owner or lifecycle stage | A saved filter or list | Leads go unworked; funnel reports are wrong |
| Contacts with no company association | A filter on associated company | Account-level reporting and scoring fail |
| Formatting problems | Data quality formatting issues | Breaks segmentation and personalization |
| Empty key fields | Property insights and filtered views | Lead scoring and routing lack inputs |
| Unused or duplicate properties | Property insights | Clutter and confusion |
| Stale deals | Open deals with past close dates | Inflates pipeline and distorts the forecast |
Step 2: Back up before you change anything
Export contacts, companies, deals, and any other objects you plan to touch, including the properties you’ll edit. Record the counts for each object. Export your lists and workflows’ configuration notes too, since you’ll want to know what depends on what. Store the files somewhere the team can find them, and keep them until the cleanup has been validated.
Step 3: Fix the structure first
Cleaning records before fixing the properties underneath them means doing the job twice. Review your properties for these issues:
- Free text where a dropdown belongs. “USA,” “U.S.,” and “United States” are the same value to a human and three values to a report.
- Duplicate or overlapping properties, such as two fields for lead source.
- Unused properties with no data and no dependencies.
- Unclear names and missing descriptions, which lead people to guess.
Before you delete or merge any property, check where it’s used in workflows, lists, forms, and reports. For guidance on when to create your own, see our explanation of custom vs. default properties.
Step 4: Standardize formats and values
The data quality tool flags formatting issues, such as a name stored as “LORELAI GILMORE” that should read “Lorelai Gilmore,” and you can accept or reject its suggestions individually or in bulk on eligible plans. Data Hub Professional and Enterprise also let you set automation so new records with the same issue are fixed automatically. One caveat from HubSpot’s documentation: if a record is later overwritten, for example by an API or integration, with the same bad format, the tool may not flag it again, so fix the source too.
For values that need consolidating, such as country names, industries, and job titles, use bulk edit or a cleaned import file to map the variants to one standard value. Then change the property to a dropdown so the problem doesn’t return.
Step 5: Deduplicate carefully
HubSpot’s duplicates manager compares record properties and surfaces likely pairs, with a daily scan. By default, contacts are compared on properties like name, email, phone, and company name, and companies on domain, name, country, and similar fields. Eligible plans can create custom rules with their own matching criteria. Work in this order:
- Companies first. Merging companies consolidates the contacts and deals attached to them, which makes later contact merges cleaner.
- Contacts next. Email is the usual unique identifier.
- Deals last, reviewing any that look like repeats of the same opportunity.
Test on a small batch and review associated records before you merge in bulk. Use the similarity score to prioritize, and when you reject a pair, remember that it won’t reappear unless one of the records changes. Decide your merge rule in advance, such as keeping the record with the most recent engagement or the oldest create date, so you keep the best data and the original source.
Step 6: Fill the gaps
- Owners: assign owners to unowned records, using routing rules where possible.
- Lifecycle stage: fix blanks and wrong values, using your agreed definitions. Our guide to HubSpot lifecycle stages explains the rules and the forward-only behavior to plan around.
- Company associations: associate orphaned contacts with companies, by domain where it’s reliable.
- Enrichment: the data quality tool shows enrichment coverage by property, which helps you see which fields enrichment can fill.
- Required fields: make the key fields required on forms and, where your plan supports it, on deals.
Step 7: Clean your lists, contacts, and deals
Lists. Audit your lists for stale static lists, duplicates, and unused ones. Consider whether a static list should be an active one that updates itself, and check dependencies before deleting. Our post on active vs. static lists covers when to use each.
Contacts. Review hard bounces, unsubscribes, and long-unengaged contacts. If you pay by marketing contacts, setting unengaged ones to non-marketing can reduce cost, and archiving or deleting old records should follow your retention and consent policy.
Deals. Find open deals with past close dates, no amount, or inactive owners, and close, update, or reassign them. Make closed-lost reason a required dropdown. See our guide to HubSpot deal stages for the stage rules that keep pipelines honest, and deal slippage for why stale close dates matter.
Step 8: Recheck scoring, routing, and reports
Cleaning changes the inputs to your automation. After the cleanup, review your lead scoring model, since fit scores depend on filled-in fields like industry and company size. Retest routing rules, and spot-check key reports against your pre-cleanup exports. If you’re also planning a platform move, finish the cleanup first, as our Salesforce to HubSpot migration checklist explains.
A before-and-after snapshot
Track a handful of measures so you can show the result. These numbers are illustrative.
| Measure | Before | After 30 days |
|---|---|---|
| Duplicate contact pairs | 4,100 | 150 |
| Contacts with no owner | 9,800 | 600 |
| Contacts with no lifecycle stage | 12,500 | 1,200 |
| Companies missing industry | 35% | 12% |
| Open deals with past close dates | 240 | 20 |
| Unused properties | 112 of 380 | 20 of 270 |
Step 9: Stop the mess from coming back
- Prevent at the source. Use dropdowns, required form fields, and validation instead of free text.
- Automate formatting on new records, where your plan allows.
- Control imports. Use a checklist: clean the file, match on a unique identifier, map fields deliberately, and test a small batch.
- Review integrations. Check which system writes each field and whether syncs create duplicates.
- Limit who can create properties and edit key fields.
- Keep a data dictionary with each property’s meaning, owner, and allowed values.
- Review quarterly. Use the data quality digest and add the check to your RevOps meeting cadence.
Mistakes to avoid
- Merging in bulk with no backup and no test batch
- Cleaning records before fixing properties and dropdowns
- Deleting properties or lists that workflows and reports still use
- Fixing formats in HubSpot while an integration keeps writing bad values
- Running the cleanup during active campaigns or pipeline reviews
- Treating it as a one-time project with no owner
- Deleting contacts without checking consent and retention rules
Summary
HubSpot data cleanup works best in a fixed order: define what clean means, audit with the data quality tools, back up, fix properties and dropdowns, standardize formats, deduplicate companies then contacts then deals, fill gaps in owners, lifecycle stages, and associations, clean lists, contacts, and deals, and then recheck scoring, routing, and reports. Because merges can’t be undone, test small batches and keep exports.
The cleanup only lasts if you prevent the mess at the source with dropdowns, required fields, automated formatting, import controls, integration reviews, and a data dictionary, then review the results every quarter. Track a few before-and-after measures so the effort is visible.
Frequently asked
What is HubSpot data cleanup?
It’s the process of auditing and fixing your CRM data: merging duplicates, standardizing formats, filling gaps, tidying lists and deals, and adding rules so errors don’t come back.
Can I undo a merge in HubSpot?
No. Merged records cannot be reverted. You can undo a rejected duplicate pair for a limited time, so back up and test on a small batch before merging in bulk.
Where are HubSpot’s data quality tools?
Under Data Management, then Data Quality. They cover duplicates, formatting issues, enrichment, and property insights. Some capabilities, such as custom duplicate rules and formatting automation, require Data Hub Professional or Enterprise.
How does HubSpot find duplicates?
It compares property values on a regular scan, such as email, name, and phone for contacts and domain, name, and country for companies, and assigns similarity scores. Eligible plans can create custom matching rules.
How often should we clean our HubSpot data?
Run a full review at least quarterly, and use the weekly data quality digest and automation to catch issues in between. After a migration, a large import, or a new integration, check sooner.
Should I delete or keep old contacts?
It depends on your retention and consent policies and applicable privacy rules. Many teams suppress or set unengaged contacts to non-marketing first, then delete according to a documented policy. Check with legal if unsure.
Do I need to clean data before migrating or adding AI tools?
Yes. Migrations, scoring, and AI features all inherit the quality of the data underneath them, so cleanup first is cheaper than fixing the same problems in a new system.
Who should own data quality?
Usually RevOps owns the standards, tooling, and reviews, while sales, marketing, and customer success own the accuracy of the data they enter. Name one accountable person.
Practical CRM notes, once a month.
HubSpot tips, RevOps playbooks, and what is working for growing revenue teams. Short, useful, and never salesy.
