HubSpot Portal Audit Checklist: What to Check and How to Fix It

A HubSpot portal audit checks whether the CRM still supports the decisions your team makes with it: who owns a lead, which stage a record is in, which reports leadership trusts. Start with HubSpot’s built-in Data Model Health Check, then test routing, lifecycle definitions, duplicates, workflows and integrations with real or test records. Fix findings in order of business impact.

This guide gives each audit area a decision it protects, a place to look in HubSpot, the evidence that passes and an owner. Examples use fictional records. Feature availability varies by subscription, so confirm each tool in your own portal before planning around it.

What should a HubSpot portal audit cover?

Audit the areas where a wrong value changes what someone does next. Settings that nobody acts on can wait.

Download the HubSpot portal audit checklist (CSV). It lists all 16 checks with where to look in HubSpot, the passing evidence, a test record or sample, an owner and the severity if a check fails. No sign-up is needed.

How do you run HubSpot’s Data Model Health Check?

HubSpot’s Data Model Health Check scans objects, properties, associations, pipelines and lifecycle stages, and HubSpot lists it for all products and plans. Open Data Management > Data Model and select the Health tab. A scan can take up to 20 minutes, a new run replaces the previous results, and the current version does not let you dismiss an issue. Export or screenshot the results before you change anything.

The check detects 16 issue types in five categories:

• Properties: low fill rate, critical properties with a low fill rate, low fill rate with high usage, inactive properties, similar or duplicate properties, and a high fill rate with no usage.
• Pipelines: non-standard usage, too many stages, a missing closed stage and a low record count per stage.
• Objects and associations: custom or toggleable objects with few records or no associations, and active associations with no associated records.
• Lifecycle stages: stages with low record counts.

Treat each finding as a question for the field’s owner. A critical property with a low fill rate may mean a broken form or integration. A similar or duplicate property often means two teams solved the same problem separately, and reports may be reading the wrong one.

Which properties should you archive, merge or keep?

Before changing a property, check where it is used: forms, workflows, lists, reports and integrations can all depend on it. Export the property list as a baseline first. Archive an unused property before deleting anything. HubSpot keeps archived properties in an Archived tab for 90 days before deleting them permanently, which leaves time to restore one if a hidden dependency appears.

For each critical property, record four things: its purpose, its allowed values, its owner and the system allowed to write it. A field written by a form, an import and an enrichment tool at once has no reliable value, whatever its fill rate says. The HubSpot CRM data model guide covers objects, properties and associations in more depth.

Are lifecycle stages and lead status defined and enforced?

Lifecycle stage describes the relationship with a contact or company. The contact Lead Status property tracks sales work within a stage, and deal stage tracks one opportunity. An audit checks that each has a written definition and that automation follows it.

Pick five recent records at each stage and ask whether the entry evidence is present. A contact marked Sales Qualified Lead with no accepted handoff, or a customer whose stage moved back without a recorded reason, points to a workflow or import setting the stage outside its definition. HubSpot tools move the default lifecycle stage forward only, so a backward move means someone cleared the value or changed it by hand. The lifecycle stages guide shows how to write stage definitions and test them.

Does every new lead reach the right owner?

Routing fails quietly. A lead with no owner or with the wrong one looks fine in a list, and nobody calls it. Create test leads for each routing rule: a new account, an existing customer, an account owned by another rep, a country outside your territories and a record with a missing field. Submit them through the real entry points, then check the owner, the task and the notification each one receives.

Recheck after every writer has run, including enrichment and sync jobs that may change the fields routing depends on. The HubSpot lead routing audit lists the test cases in detail.

What creates duplicates, and who prevents them?

Duplicates come from create paths: forms, imports, integrations, enrichment tools and manual entry. For each path, find out which identifier it checks before creating a record. HubSpot matches contacts by email, but companies created through the API, including sync apps, are not automatically deduplicated by domain.

Review existing pairs with HubSpot’s duplicate management where your subscription includes it, and merge only when the retained values and associations are agreed. Then close the create path that produced them. The HubSpot deduplication guide covers the review queue and merge checks.

Which workflows are dormant, conflicting or unowned?

List every active workflow with its owner, trigger, the properties it sets and the date it last enrolled a record. Look for three problems. Some workflows have enrolled nothing for months. Some set the same property as another workflow under different conditions. Some belong to someone who has left the team.

For each workflow that sets a property, check its enrollment and re-enrollment settings against the property’s definition. Two workflows that change lifecycle stage or owner under overlapping conditions will disagree eventually, and the record ends up with whichever ran last. Turn off a dormant workflow only after confirming that no list or report depends on its actions.

Are integrations writing only the fields they own?

List every connected app, private app and integration user, and the fields each one can change. Compare that list with the property owners from the data model step. Where two systems write the same field, decide which one owns it and make the other read only.

Check sync direction and conflict rules for two-way syncs, and look for integration users with broad permissions that nobody can explain. The Clay and HubSpot integration guide and the n8n HubSpot integration guide show how to design writes that a team can audit.

Do reports answer the decisions they are used for?

For each dashboard used in a pipeline or marketing review, write down the decision it supports, its data source, its filters and the date range. Check one number by hand against the underlying records. Attribution reports deserve extra care, because HubSpot and GA4 count sessions and sources differently; the GA4 vs HubSpot attribution guide explains why the numbers differ.

In what order should you fix the findings?

Rank each finding by what it costs the business while it stays open:

1. Leads without an owner or with the wrong owner, and create paths that keep adding duplicates.
2. Stage, status and owner fields written by more than one system or workflow.
3. Reports used for decisions that read the wrong property or miss records.
4. Unused properties, dormant workflows and empty objects.

Give each fix an owner and a test record that proves it worked, and re-run the health check after the first round of changes. A first audit of a small portal fits a 30-day plan: week one for evidence, weeks two and three for the top fixes, week four for retesting and documentation.

Frequently asked questions

For an audit scoped around your portal and sales process, see HubSpot development and RevOps or the systems readiness review.