HubSpot Data Cleanup: The Complete Guide

Every HubSpot portal starts clean. New portals have zero records, consistent property definitions, and tidy pipelines. Then reality hits. Reps enter data in their own formats. Imports bring in records from other systems with different conventions. Forms capture data exactly as visitors type it. Integrations sync data from platforms that use their own naming standards.

Six months in, you have "United States" and "US" and "USA" and "U.S.A." in the same country field. You have phone numbers in eight different formats. You have company names spelled three different ways. Your automation breaks because a workflow checking for "Customer" does not match records labeled "customer" or "Existing Customer."

This is the guide for fixing it — and keeping it fixed.

To illustrate what a typical cleanup operation looks like, here is a country-name normalization pass:

Assessing the damage

Before you start cleaning, you need to know how dirty the data is. Randomly sampling records gives you a rough sense, but for a real assessment you need a systematic approach.

Property audit

Start with the properties that matter most to your business. These are typically:

Contact properties:

• Email (format, domain validity)
• Phone Number (format consistency)
• Lifecycle Stage (correct values, no blanks)
• Lead Status (matches your defined stages)
• Country/State (consistent naming)
• Job Title (standardized for segmentation)

Company properties:

• Company Name (consistent casing and formatting)
• Industry (standardized taxonomy)
• Website (valid URLs, correct domains)
• Number of Employees (populated, reasonable ranges)
• Country (same standards as contacts)

Deal properties:

• Deal Name (follows naming convention)
• Pipeline and Stage (valid values)
• Amount (populated for forecast-relevant deals)
• Close Date (populated, reasonable)

For each property, ask:

1. What percentage of records have this property populated?
2. How many distinct values exist where there should be a limited set?
3. Are the formats consistent?

Finding inconsistencies

HubSpot's list tool can help surface some issues. Create active lists filtered by:

• Country is any of "US" — then check: do you also have records with "USA," "United States," etc.?
• Phone Number contains "(" — this shows one format variant; compare against records without parentheses
• Industry is any of "Technology" — then look for "Tech," "Software," "IT," etc.

For text properties where you suspect fragmentation, you can export the property values and analyze distinct values in a spreadsheet. Sort unique values alphabetically and scan for variants that should be the same.

The cleanup process

Data cleanup has a logical order. Doing it out of sequence creates rework.

Phase 1: Remove duplicates

Start here because duplicates skew every subsequent analysis. If "John Smith" exists twice, any property analysis will count his data twice, and any bulk update might fix one record but miss the other.

HubSpot's built-in duplicate management tool (Contacts > Actions > Manage duplicates) identifies potential duplicates based on email address, first and last name, and company name. Review and merge them before moving to property cleanup.

For companies, HubSpot identifies duplicates by domain. Merge companies with the same website domain, then review remaining potential duplicates manually.

Deduplication is inherently a manual review process because the tool needs a human to decide which record to keep and how to merge conflicting property values. Budget time for this — it cannot be fully automated.

Phase 2: Standardize categorical values

These are the properties with a finite set of valid values where your data has fragmented into multiple representations of the same thing.

Country standardization is usually the biggest win. Typical mapping:

• "US," "USA," "U.S.A.," "United States of America" all become "United States"
• "UK," "U.K.," "Great Britain," "England" all become "United Kingdom" (or whichever standard your organization uses)

Run a find-and-replace for each variant. The operation is: search for "USA" in the Country property, replace with "United States." Repeat for each variant.

Industry standardization follows the same pattern. Decide on your canonical industry values, map all variants to those values, and execute the replacements.

State/Province fields are often a mess in international portals. "CA" could mean California or Canada. "NSW" might or might not have a period. Standardize based on your audience — if you are US-focused, use two-letter state codes. If international, use full names.

Phase 3: Fix formatting

Formatting issues are different from value standardization. The data is correct but inconsistent.

Phone numbers are the classic example. The number 5551234567 might appear as:

• (555) 123-4567
• 555-123-4567
• 555.123.4567
• +1 555 123 4567
• 15551234567

Choose a format and use regex find-and-replace to convert all variations. A regex pattern that captures digit groups and reassembles them works here.

Email domain capitalization is another common issue. "[email protected]" and "[email protected]" represent the same domain but look inconsistent in reports. Email addresses should generally be lowercased.

Name capitalization is trickier. "john smith" should probably be "John Smith," but programmatic capitalization has edge cases (McDonald, O'Brien, van der Berg). Regex can handle simple cases; edge cases may need manual review.

URL formatting — some records have "https://www.company.com," others have "company.com," others have "http://company.com/." Decide on a standard (most teams use the bare domain without protocol) and normalize.

Phase 4: Fill in missing data

After standardization and formatting, you will have a clearer picture of which records are genuinely missing data versus which appeared to be blank because of inconsistent values.

For truly missing data, your options are:

• Manual enrichment — Reps fill in data during their normal workflow
• Third-party enrichment — Tools like Clearbit, ZoomInfo, or Apollo append data from external databases
• Form progressive profiling — Collect missing fields through subsequent form submissions
• Workflow prompts — Create internal notifications when critical properties are blank

Missing data is a different problem than dirty data, but the cleanup process often reveals the true scope of it.

Phase 5: Prevent recurrence

This is where most cleanup projects fail. The team spends a week cleaning data, and six months later the same problems are back. Prevention requires structural changes:

Convert text fields to dropdowns. If a property should only contain values from a defined set, change it from a text field to a dropdown. This is a one-time HubSpot admin change that permanently prevents free-text variations.

Add property validation rules. HubSpot supports validation on some property types. Phone number fields can enforce a format. Number fields can enforce ranges. Use these where available.

Create import templates. If your team regularly imports data, create standardized CSV templates with the correct column headers, example values, and formatting notes. Better yet, use HubSpot's import with a default mapping template.

Train your team. Document your data standards and include them in onboarding. If reps know that Industry should be "Technology" (not "Tech" or "Software"), they will enter it correctly. If they do not know, they will guess.

Automate standardization. For high-volume properties that receive data from many sources, consider a HubSpot workflow that normalizes values on entry. A workflow triggered by "Country" property change can check the value against a list and set it to the standard form automatically.

Schedule recurring audits. Quarterly is a good cadence for most portals. Monthly for high-volume portals or those with many data sources. The audit does not need to be exhaustive — spot-check your critical properties and run quick find-and-replace operations to fix any drift.

Measuring data quality

Cleaning data is not a one-time event. You need metrics to track whether quality is improving, stable, or degrading.

Useful metrics:

• Fill rate — Percentage of records with a given property populated. Track this for your top 10 most important properties.
• Distinct values — For categorical properties, the number of distinct values relative to the expected set. If your Industry property should have 15 values but has 47, you have fragmentation.
• Duplicate rate — Percentage of records flagged as potential duplicates. This should trend downward after cleanup.
• Format compliance — For properties with expected formats (phone, email, URL), the percentage that match the expected pattern.

Track these monthly. A dashboard showing these metrics over time tells you whether your data quality practices are working.

The ROI of clean data

Data cleanup is often treated as a cost center — necessary but unglamorous work that does not directly generate revenue. That framing misses the compounding benefits:

• Automation reliability — Workflows that trigger on property values only work when those values are consistent
• Reporting accuracy — Segment-based reports are only as good as the segmentation data
• Sales productivity — Reps spend less time correcting data and more time selling
• Customer experience — Personalized emails with correctly formatted names and relevant content perform better than ones with obvious data errors
• Integration stability — Systems that sync data between HubSpot and other platforms break when data formats are inconsistent

Clean data is not a project. It is an ongoing discipline. The tools you use to maintain it — bulk find-and-replace, duplicate management, validation rules, and recurring audits — determine whether that discipline is sustainable or whether it is a Sisyphean task that your team abandons after the first attempt.