How to Fix Bad Data in HubSpot: A Practical 10-Step Process

A demo request arrives in HubSpot without a country. Your assignment workflow needs that field, so the lead never reaches the right salesperson. When someone finds it, the company record has no domain and the contact is associated with the wrong business.

Each issue looks small on its own. Together, they make a simple follow-up harder than it should be.

Bad CRM data costs your team time and makes it harder to trust the information in front of them. Sales misses context. Marketing leaves relevant contacts out of segments. Management sees a pipeline report that only tells part of the story.

I’ve been implementing HubSpot since 2015. At Superwork, the process I use starts with the work the data needs to support: identify what is wrong, correct and connect the records, fill useful gaps, and fix the sources of recurring errors.

Here is how to work through that process in your own portal.

Use the companion HubSpot Data Quality Checklist as you go. It includes the checks below, a baseline worksheet, priority fields, and a maintenance schedule.

1. Start with the process that is failing

A clean-looking database can still contain information your team cannot use. Look for six kinds of problems:

Start with a specific question. Are new leads reaching the right person? Does a marketing segment include everyone it should? Can a salesperson see the customer’s full history?

Choose one process, define the records in scope, and measure the problems before changing anything. Track missing owners, missing identifiers, confirmed duplicates, association errors, or the hours your team spends repairing records. Use the same definitions when you measure again.

Before bulk changes, export the affected records and relevant values, document dependencies, and test a small sample. An export gives you a reference; it is not a complete backup of your portal.

2. Check whether your contact emails are usable

An email address being present does not tell you whether it still works. People change jobs, domains change, and typing errors find their way into forms and imports.

For a list you intend to use, run your approved verification process. A service such as NeverBounce is one possible approach; use the tool and data-handling process approved by your business.

Store the verification result and date in HubSpot so the next person does not have to repeat the investigation. Separate invalid results from uncertain ones, and review uncertain addresses before using them.

Keep this information separate from subscription and permission records. Verification answers a technical question about an address. It does not establish permission to send marketing emails, and it does not guarantee inbox placement.

3. Give company records reliable identifiers

Acme, ACME, and Acme Ltd might be one business. They might also represent separate legal entities. The name alone is not enough to decide.

I look for a consistent company name, the correct domain, a LinkedIn company URL where useful, and a trusted business identifier such as a registration number or ERP customer ID.

Define what that identifier represents. A customer number may identify an account in your ERP rather than a legal entity. A registration number may need a country or registry prefix to be unique across your database.

HubSpot uses company domains for automatic deduplication in supported creation routes, but API-created companies are not automatically deduplicated by domain. Check the matching behaviour of each integration instead of assuming it behaves like a manual entry or import. (HubSpot’s deduplication documentation)

Reliable identifiers make later matching and enrichment easier. They also help you avoid merging separate subsidiaries simply because they share a website.

4. Clean up the properties your teams rely on

Suppose your portal contains Region, Sales region, and Region (old). Sales updates one. Marketing forms populate another. An integration writes to the third.

A report using Sales region will miss information stored elsewhere, even if every team believes it has completed the field.

Agree on one authoritative property for each shared purpose. Decide what it means, which values it accepts, who owns it, and which source wins when values conflict.

Then consolidate useful information and update the forms, workflows, integrations, segments, and reports that depend on those fields. Only retire the redundant properties after checking that the new setup works.

Treat empty or apparently unused properties as review candidates. A field may support a seasonal process or an integration your team has forgotten about.

5. Review duplicates, then repair the connections

Use trusted identifiers to find likely duplicates. A verified registration number or customer ID is a stronger starting point than a similar company name.

Where available, custom duplicate detection rules can help build the review queue. HubSpot currently supports up to two rules per object with Data Hub Professional or Enterprise. (Custom duplicate rules)

For each confirmed match, decide which record should remain and which values to keep. Check external IDs and connected systems before merging. Afterwards, verify the resulting timeline and associations.

HubSpot merges cannot be undone. Review ambiguous cases individually rather than letting a broad matching rule make the decision. (HubSpot merge guidance)

Also check records that are not duplicates. A unique contact can still be attached to the wrong company, and an otherwise complete deal can lack the contacts involved in the purchase. Fixing these connections makes the information useful across teams.

6. Standardise the values behind your reports

A country segment filtered for Germany may miss records stored as DE or Deutschland. Everyone has entered something meaningful, but the system treats the values differently.

Choose a standard for important fields, map existing values to it, and use controlled options where free text creates unnecessary variation. Document the rules so imports and integrations use them too.

HubSpot’s Data Quality tools can suggest supported formatting fixes. Formatting automation is available with Data Hub Professional or Enterprise. Use it for the issues it covers, alongside your own business rules for fields such as region and customer type. (HubSpot data quality tools)

Correcting today’s values is only part of the work. If an integration writes the old values back tomorrow, fix the mapping at the source.

7. Enrich the information you actually need

A segment for SaaS companies with 50–200 employees depends on industry and employee count being populated correctly. Missing values can exclude otherwise relevant companies.

Define the information the process needs before enriching records. Check the identifiers used to match them, review overwrite settings, and inspect a sample of the results.

Handle the existing database deliberately. Switching on automatic enrichment does not backfill every record already in your portal; existing records need to be selected for enrichment. Review continuous-enrichment settings separately if you want supported information refreshed over time. (How HubSpot enrichment works)

Some gaps will remain. Put those in a research queue rather than filling them with guesses. A blank field is easier to investigate than a confident but incorrect value.

8. Let AI help without giving it unchecked control

Useful information often sits in call transcripts, emails, and notes. AI can help turn that material into structured fields, but the source needs to support the answer.

For example, create a dedicated smart property for the CRM a prospect currently uses. Select the appropriate activity or transcript source, then test a prompt along these lines:

Identify the CRM the company currently uses from the selected source. Do not treat a product being evaluated, a former system, or our own recommendation as its current CRM. If the evidence is missing or conflicting, return Unknown.

Review a varied sample before filling a larger segment. Check straightforward records, ambiguous conversations, and records with no relevant evidence. Keep AI-generated results separate from operational fields until you have a reliable review process.

HubSpot smart properties consume credits when run, including when no value is filled. Set a budget and review automatic-fill settings before expanding usage. (HubSpot smart-property documentation, and our guide to HubSpot smart properties)

9. Automate rules your teams have agreed on

Lifecycle stage, lead status, and buying role help teams organise follow-up and handoffs. Automation works best when those definitions are clear first.

Document the event that should trigger each change, the required information, and the exceptions. Decide which source wins when an integration, workflow, and salesperson can all update the same field.

A lead-assignment workflow, for example, needs a route for records with a missing country. Send them to an exception queue with an owner instead of allowing them to disappear between teams.

Test the ordinary case and the difficult ones: an existing customer submitting a new form, an incomplete record, a re-enrolment, or a manually confirmed value. Confirm the downstream report and handoff still behave as intended.

10. Make data quality part of regular maintenance

Forms, imports, integrations, and changes in your customers’ businesses continue to introduce new information. Your cleanup needs an owner and a review routine.

Start with this cadence and adapt it to your volume and response commitments:

For every recurring problem, record where it came from and who can fix that source. Otherwise, the team will keep repairing the same symptoms.

The outcome to aim for is practical: leads reach the right person, segments include the right records, sales can see the customer history, and reports reflect the process your business actually uses.

Work through your own portal

Use the HubSpot Data Quality Checklist to record your baseline, review the ten steps, assign your next three actions, and set a maintenance schedule.

If you want help finding the causes and deciding what to fix first, explore a HubSpot audit with Superwork.

Get the free HubSpot Data Quality Checklist →

Feature details checked on 2 November 2026. Subscription, permissions, and rollout affect the tools available in your portal.