How to Fix Bad Data in HubSpot: The Complete Guide (2026)

What is bad data in HubSpot?

Bad data in HubSpot is any record your team can't trust or act on: duplicates, invalid emails, missing company domains, inconsistent formatting, and stale information.

In practice, it comes in five flavors:

1. Duplicates. The same person or company exists two, three, or ten times — each version holding a fragment of the truth. Deals attach to one record while marketing emails go to another. Nobody gets a 360-degree view of the customer, because the customer is scattered across records.

2. Invalid data. Email addresses that bounce. Phone numbers that don't connect. Free-text fields full of "asdf" and "test@test.com."

3. Incomplete data. Companies without domains. Contacts without companies. Deals without amounts. Incomplete records can't be segmented, scored, routed — or deduplicated.

4. Outdated data. The contact changed jobs 14 months ago. The company rebranded. Job changes alone run at 15–20% per year, and they drive most B2B data decay.

5. Inconsistent data. "Germany," "germany," "DE," and "Deutschland" in the same country field. www.acme.com on one company record and acme.com on its duplicate.

Why does this matter so much? Because every one of these breaks something downstream: segmentation, lead routing, attribution, forecasting, and every report your leadership team looks at. If your team is unifying data across systems and still doesn't trust the numbers, this is usually why.

How does HubSpot deduplication actually work?

HubSpot automatically deduplicates contacts by email address and companies by company domain name — at creation and import only. It never uses company name, and it never deduplicates records created through the API or third-party integrations.

This is the most misunderstood part of HubSpot data management. It's also the reason the enrich-first order exists. Let's break it down.

Contacts: email address is the unique identifier

For contacts, HubSpot dedupes on email address.

Try to manually create a contact with an email that already exists? HubSpot blocks it. Form submission or import with an existing email? HubSpot updates the existing record instead of creating a new one. On forms, HubSpot also uses the browser cookie (usertoken) to recognize returning visitors.

This works well — as long as the email is present and valid. But a contact with no email, a typo'd email, or a second work address? Invisible to dedup.

Companies: domain name — never company name

For companies, the unique identifier is company domain name. On import, HubSpot matches on the primary domain (plus secondary domains).

Here's the part that surprises most teams: company name is never used for automatic deduplication. "Acme GmbH," "Acme," and "ACME Europe" are three separate companies to HubSpot forever — unless they share a domain.

Which means a company record without a domain can never be auto-deduplicated. That single fact is why enrichment comes before deduplication.

Record ID and custom unique properties

Two more matching options worth knowing: you can dedupe on Record ID during imports (export first, then re-import against the ID). And you can define up to 10 custom unique-value properties per object — an ERP customer number, a VAT or company registration number — which work as dedup keys on imports, but not on forms.

Why does HubSpot keep creating duplicates?

Because four gaps in the automatic deduplication let them through:

1. API and integration-created companies are not deduplicated by domain. ZoomInfo, a Salesforce sync, LinkedIn tools, Zapier — records from all of them bypass domain matching entirely. This is the #1 driver of duplicate companies.
2. Domain variations slip through. acme.com and www.acme.com can create separate records; only clean primary domains auto-match.
3. Missing identifiers. No email on the contact, no domain on the company — no dedup.
4. The auto-create setting. "Automatically create and associate companies with contacts" reliably associates — but the create half generates thin, near-duplicate company records from every new email domain it sees.

One more thing to be clear about: HubSpot never auto-merges suspected duplicates. It can block some duplicates at creation and flag likely pairs for review. But a human (or a third-party tool you've configured) always pulls the trigger.

Step 1: Enrich — make your data matchable

Enrichment means making sure every record carries a valid, consistently formatted unique identifier — a working email on contacts, a clean domain on companies — before you attempt any merging.

It sounds backwards. Resist the urge to start merging. You can't deduplicate what you can't match. HubSpot's merge logic keys on email and domain, so step one is making those two fields present, valid, and consistent across the database.

How to validate email addresses in HubSpot

Run your contacts through a validation service, store the result on a custom property, suppress the invalid addresses, and add real-time validation at the point of entry.

First, why bother? Every invalid email does double damage: it's a record that can't be deduplicated properly, and it's a future hard bounce eroding your sender reputation.

A validation service flags invalid syntax, non-existent mailboxes, disposable domains, catch-all servers, role-based addresses (info@, sales@), spam traps, and known abuse addresses. Here's how the main validators for HubSpot compare:

Now, the workflow:

1. Baseline. Export your contacts, run a sample through a validator, and record your current bounce rate and percentage of invalid addresses.
2. Bulk clean. Run the full list through NeverBounce or ZeroBounce. Import the results into a custom "Email validation status" property.
3. Segment. Build active lists for invalid, catch-all, role-based, and spam-trap addresses.
4. Act. Suppress invalid addresses from all sends via workflow. Review catch-alls manually. Move long-unengaged contacts into a sunset track.
5. Prevent at entry. Add real-time validation on creation and form submission — a native marketplace app, or (on Data Hub Professional) a custom-coded workflow action that calls the validation API and only routes valid emails into nurture.
6. Govern. Stamp a "last verified" date via workflow. Re-validate quarterly. Watch the bounce-rate KPI.

Pro tip: there are four ways to wire a validator into HubSpot — native marketplace app, list-based validation-and-routing workflow, Data Hub custom-coded workflow action, or Zapier. Native apps are the fastest start. But the custom-coded action is the only one that gates bad emails before they enter your nurture flows.

How to enrich company domains in HubSpot

Find every company record with an empty Company domain name field, then fill it with an enrichment tool — because a company without a domain can never be auto-deduplicated.

Your options:

Which one should you pick?

Breeze Intelligence — HubSpot's own enrichment layer, built on the Clearbit acquisition — is the default for most portals. It's native and cheap to start.

Dropcontact is the standout for European and Nordic teams with a strict GDPR posture: EU-hosted, CNIL-audited, and algorithmic rather than database-scraped. That distinction matters when your DPO asks where the data came from. It also does name-to-domain matching — exactly the gap you're filling here.

Pro tip: for contacts on free email providers (Gmail, Outlook.com), enrich the company from the contact's Website URL or company name instead. Otherwise those companies stay domain-less — and un-mergeable.

Standardize formatting so matches actually match

Enriched data still won't match if the formatting differs — so normalize domains, names, and locations with a Format data workflow before you dedupe.

https://www.acme.com, www.acme.com, and acme.com should all become acme.com.

Use a workflow with the Format data action (Data Hub Professional and up) to strip protocols and www. prefixes from domain fields, trim whitespace, and fix capitalization on names and cities. The Data Quality Command Center also suggests formatting fixes automatically — more on that below.

Once emails are validated, domains are filled, and formatting is standardized, your records are finally matchable. Now — and only now — deduplication works.

Step 2: Deduplicate — merge at scale

To deduplicate in HubSpot, surface duplicate pairs with the duplicate management tool (Data Management > Data Quality > Manage Duplicates), review them, and merge — up to 50 records at a time natively, or via a third-party tool for large backlogs.

With clean identifiers in place, this stage is surprisingly mechanical. Three native surfaces find duplicates for you:

• The duplicate management tool (Professional and Enterprise) — AI scans daily across email, name, phone, country, zip, and company name, and ranks pairs by match confidence
• Filtered views — companies sharing a domain, near-identical names
• Spreadsheet exports — for eyeballing at scale

How to merge duplicate companies in HubSpot (step by step)

This is the full enrich-match-merge sequence for companies — the object where duplicates hurt most, because deals and contacts hang off them:

1. Find the gaps. Create a view of companies where Company domain name is empty. That's your enrichment backlog.
2. Enrich. Run Breeze Intelligence (or Dropcontact for a GDPR-first stack) across the backlog. For companies attached only to free-mail contacts, enrich from Website URL.
3. Standardize. Strip www. and protocols with a Format data workflow so primary domains match exactly.
4. Surface duplicates. Open Data Management > Data Quality > Manage Duplicates. Supplement with filtered views on domain and near-duplicate names.
5. Merge. Review pairs, choose the primary record, merge. Natively that's up to 50 records at a time; Data Hub Professional bulk-manages up to 30,000 duplicate pairs (Enterprise goes higher).
6. Verify. Spot-check merged records while the brief undo window is open. Export your merge history for the audit trail.
7. Prevent. Audit which integrations create companies via API. Configure the auto-create/associate setting deliberately. Put recurring dedup on the calendar.

What does merging actually do?

When you merge two records, the primary record's property values win, all associations and timeline activity move to the surviving record, and the merge is effectively permanent.

The details: HubSpot standardized primary-wins behavior in January 2025. Contacts, deals, and tickets all re-attach to the survivor. The merged record gets a new Record ID (old IDs still resolve, so integrations don't break), and the merge is logged on the record.

Know the limits before you start clicking:

• Merges are effectively permanent. There's a brief undo window on the confirmation screen — after that, no native unmerge. Insycle or Emergency Unmerge can partially reconstruct records, but don't count on it.
• 250-merge cap per record chain.
• Native bulk merge = 50 records at a time. Fine for hundreds of duplicates. Painful for tens of thousands.
• Merging doesn't stop new duplicates. API-created companies still bypass domain dedup — fix the source too.

When should you use a third-party dedup tool?

Go third-party when you have a large backlog, need fuzzy matching or non-email/non-domain matching, want scheduled automated dedup, are fighting a Salesforce sync, or your database has passed roughly 100,000 records.

Native tooling is enough for small, manual cleanups. Beyond that:

Insycle is the broadest suite: any-field and fuzzy matching, bulk merge with preview, standardization, a 30-plus-issue Health Assessment, and scheduled automations. (One myth to retire: Insycle has been a HubSpot Certified App Partner since 2018 — it was not acquired by HubSpot.)

Koalify is the HubSpot-native option — merges run through CRM cards and workflow actions with custom match rules, and it openly markets the contact-tier savings angle.

Dedupely is the simple choice for occasional cleanup jobs.

Step 3: Govern — keep it clean forever

Governance means closing the doors bad data enters through (property rules, import discipline), removing data that's gone stale (sunset policies), and giving data quality a named owner with a recurring cadence and KPIs.

Here's the uncomfortable truth about steps 1 and 2: at 2.1% monthly decay, their effect has a half-life. Governance is what makes the cleanup stick — the least glamorous, highest-ROI stage of the three.

Property governance and validation rules

Most dirty data enters through properties that allow it. Close those doors:

• Picklists over free text wherever values are finite — country, industry, disqualification reasons. Free text is where "Germany/germany/DE/Deutschland" chaos is born.
• Validation rules on the free-text fields you keep: min/max length, regex patterns (phone format, VAT number), unique values where a field must never repeat.
• Required fields at the moments that matter — deal amount on deal creation, country on contact creation.
• Restrict property creation to admins. A portal where every team creates its own properties ends up with 400 properties, 300 of them near-empty. Use conditional logic so users only see fields relevant to the record in front of them.

Import discipline

Imports are the other flood door. The rules: always start from an import template. Map to Record ID or a custom unique property when updating existing records. VLOOKUP your file against a fresh export before importing, so you know what will match. Split files into update-vs-new before upload.

And never import contacts without email or companies without domain — you now know exactly why.

Sunset policies: the cleanup that lowers your HubSpot bill

Marketing Hub bills on marketing-contact tiers — so purging (or de-marketing) dead contacts directly cuts your subscription cost.

Every bounced, dead, or permanently unengaged contact you keep as a marketing contact is a line item on your invoice. That makes sunsetting one of the rare cleanups with direct, provable ROI.

The play: contacts with no opens, clicks, or site visits in 6–12 months get a re-engagement sequence. Non-responders get set to non-marketing — or deleted, where GDPR's data minimization principle actually supports removal (see the GDPR section below).

Pro tip: re-validate email addresses before every major send — not after the bounce report comes in.

RevOps ownership, cadences, and KPIs

Data quality without an owner decays like everything else. Make it a named RevOps responsibility with a standing cadence:

• Weekly: review new duplicate pairs, check import errors
• Monthly: run through the Data Quality Command Center, apply formatting fixes, prune unused properties
• Quarterly: refresh enrichment (monthly for high-velocity SaaS), run the sunset workflow, re-measure decay

Track a small KPI set: duplicate rate, fill rate on critical properties, email bounce rate, formatting-issue rate, required-field missing rate.

Round it out with freshness practices: progressive profiling on forms (ask for a little, often), workflow-stamped "last verified" dates on key fields, and quarterly enrichment refreshes so job changes get caught within a quarter — not within a sales call.

What do you get with HubSpot Data Hub in 2026?

Data Hub (formerly Operations Hub) is HubSpot's data quality product: the Data Quality Command Center, duplicate management, formatting automation, data sync, and programmable workflow actions — with bulk duplicate management unlocked at the Professional tier.

Quick naming note first, because search results are a mess on this: at INBOUND 2025, HubSpot renamed Operations Hub to Data Hub. Same product, same pricing — new name. Any article still selling you "Operations Hub data quality features" is describing Data Hub.

The centerpiece is the Data Quality Command Center (Data Management > Data Quality):

• Monitoring dashboards
• Property insights — unused properties, low fill rates, duplicate properties
• Record issues — formatting problems, invalid and temporary emails
• Data sync health
• Duplicate management

Know its rhythm: the scan runs roughly every two weeks, not in real time, with a weekly digest email. You need Super Admin or data-quality permissions to see it.

Around the Command Center sit the working tools: AI-suggested formatting fixes (applied to future records), the Format data and Copy property value workflow actions, import error monitoring, and the data model overview. Breeze adds AI-suggested merges, AI formatting rules, enrichment, and a credit-based Data Agent.

Here's what each tier gets you for data quality:

One tiering nuance that trips people up: any Hub at Professional or Enterprise level (Marketing, Sales, Service) lets you review individual duplicates, up to 10,000 pairs. It's Data Hub Professional that unlocks bulk management of up to 30,000 pairs — the difference between an afternoon of clicking and an actual cleanup operation.

(HubSpot Credits — which power Breeze enrichment and the Data Agent — run $0.010 each, or $9 per 1,000 on annual terms, and reset monthly.)

Is Data Hub Professional worth ~€620/month just for data quality? If you'll also use its programmable automation, datasets, and custom-coded actions — features a RevOps team touches weekly — yes. If you only need dedup at scale, compare it against Koalify at ~$10 per 10k records first.

How should EU teams clean HubSpot data under GDPR?

Under GDPR, EU teams should treat data cleanup as a compliance win: data minimization supports purging stale contacts, EU-based enrichment tools keep the stack defensible, and a documented cleanup cadence doubles as accountability evidence.

Most data-cleanup guides treat GDPR as a constraint. For European teams, it's closer to a tailwind. Three reasons:

1. Data minimization argues for deletion. GDPR Article 5(1)(c) and (e) say you should hold personal data only as long as it serves a stated purpose. A contact who hasn't engaged in two years, with no lawful basis for further processing, isn't just costing you contact-tier fees — keeping them is a compliance liability. Your sunset policy is a GDPR policy. Purge with confidence, or set contacts to non-marketing and suppress where you have a reason to retain (an active customer relationship, a legal retention duty).

2. Tool selection is a GDPR decision. Where enrichment data comes from — and where it's processed — matters. Dropcontact runs on EU servers, is CNIL-audited, and generates data algorithmically rather than reselling a scraped database: the cleanest answer when your DPO asks. Cognism screens against European do-not-call registries. ZeroBounce and Clearout both advertise GDPR alignment for validation. And if reverse-IP visitor identification is on your roadmap, EU-based providers like Dealfront and Albacross keep that stack defensible too.

3. Lawful basis shapes the cleanup itself. Before a re-engagement campaign, check what basis you're mailing under. Before deleting, check retention obligations. This is also why we recommend documenting your cadence — a written data quality process is exactly the kind of accountability evidence Article 5(2) asks for.

For the deeper compliance picture — Schrems II, data residency, and HubSpot's EU hosting — see our HubSpot GDPR and data privacy guide.

FAQ: fixing bad data in HubSpot

Does HubSpot automatically remove duplicates?

Partially. HubSpot auto-deduplicates at creation and import — by email address for contacts and company domain name for companies — but it never auto-merges suspected duplicates; a human always confirms the merge. Records created via API or third-party sync apps bypass domain deduplication entirely, which is why integrations are the biggest duplicate source.

How do I merge duplicate companies in HubSpot?

Open the company record and choose Actions > Merge, or go to Data Management > Data Quality > Manage Duplicates to review AI-flagged pairs. With Data Hub Professional or Enterprise you can bulk-merge up to 50 records at a time and manage up to 30,000 duplicate pairs.

Can you undo a merge in HubSpot?

Only within a brief undo window on the confirmation screen. After that, the merge is permanent. Third-party tools like Insycle or Emergency Unmerge can partially reconstruct pre-merge records, but treat every merge as final and preview before you commit.

Why does HubSpot keep creating duplicate companies?

Three causes: records created by integrations or the API (which skip domain dedup), domain variations like www.acme.com vs acme.com, and the "Automatically create and associate companies with contacts" setting generating thin near-duplicate records from new email domains.

What's the best email validator for HubSpot?

NeverBounce and ZeroBounce both integrate natively. ZeroBounce adds AI activity scoring, catch-all scoring, and a free monthly tier; NeverBounce offers a bounce-back guarantee and fits ZoomInfo stacks. Kickbox, Clearout, and Bouncer are solid lower-cost alternatives.

Which HubSpot tier do I need for duplicate management?

Any Hub at Professional or Enterprise level lets you review individual duplicate pairs (up to 10,000). For bulk duplicate management — up to 30,000 pairs — you need Data Hub Professional; Enterprise raises the limit further.

How fast does B2B data decay?

About 2.1% per month, or roughly 22.5% per year (MarketingSherpa, validated by HubSpot's Database Decay Simulation). Email fields decay faster — about 23% of email lists degrade annually per ZeroBounce — and job changes (15–20% of professionals per year) drive most of it.

Does cleaning duplicates save money?

Yes, directly. Marketing Hub bills on marketing-contact tiers, so purging duplicate, bounced, and dead contacts lowers your subscription cost — before you count the gains in deliverability, rep time, and report accuracy.

Is the "70% of B2B data goes bad every year" stat real?

Not verifiably. The viral "70.3% — Gartner" figure doesn't trace back to any locatable Gartner report. The defensible benchmarks are 2.1%/month (~22.5%/year) overall decay and ~23% annual email-list degradation — bad enough on their own.

Fix it once, keep it fixed

Bad data in HubSpot isn't a mystery. It's the predictable result of how deduplication works, colliding with how data actually enters a portal. And the fix follows from the same logic:

Enrich first, so every record carries a matchable email or domain. Deduplicate second, while the identifiers are clean. Govern forever, because 2.1% of your database will decay next month whether you're watching or not.

Run the loop and the payoff compounds: reports your leadership trusts, segmentation that reaches real people, automation and AI features working from truth instead of noise — and a smaller HubSpot bill.

Start by measuring where you stand. Get a free HubSpot audit to baseline your duplicate rate, bounce rate, and fill rates this week.

And if the backlog is bigger than your team's bandwidth, book a data cleanup with Superwork — as a HubSpot Gold Solutions Partner working with B2B SaaS and mid-market teams across the Nordics and Europe, cleaning portals and keeping them clean is a large part of what we do.

Last updated: August 2026. Prices verified against vendor pricing pages at time of writing; EUR figures converted at €1 ≈ $1.16 and marked approximate where vendors list a single currency.