Migrating to HubSpot without importing your mess

A CRM migration is the one moment when every record passes through your hands. It's also the moment most teams copy their old problems into a new tool: the same duplicates, the same empty fields, now with a fresh create date. Cleaning before the import costs a few days. Cleaning after it costs months.

Decide what is worth moving

Not everything in the old CRM deserves a place in HubSpot. Before mapping a single column, sort the data into three groups:

• Move as is: customers, open deals, active prospects, and the companies they belong to.
• Move after cleanup: records that are useful but incomplete or duplicated.
• Leave behind, or archive outside the CRM: contacts nobody has touched in years, test records, bounced addresses, leads from events you no longer run.

Every record you leave behind is one you'll never have to deduplicate, enrich or explain in a report.

Clean the export, not the new portal

Work on the exported files, where bulk edits are easy and nothing is synced yet.

1. Merge duplicates first. Group contacts by email and companies by website or domain, and keep one row per person and per company. A duplicate imported twice becomes two HubSpot records with their own history.
2. Give every company a domain. HubSpot deduplicates companies on the company domain name, and uses it to associate contacts by their email domain (HubSpot: automatically create and associate companies with contacts). A company without a domain matches nothing.
3. Normalise the values of every dropdown. For dropdowns and other enumerations, the imported value must match an existing option's label or internal value. "France", "FR" and "france" need to become one value before the import, not after.
4. Format dates and phone numbers. HubSpot accepts day-month-year, month-day-year or year-month-day dates, and phone numbers in the +[country code][number] format (HubSpot: set up your import file).
5. Map owners. Every record should land with an owner who still works with you. Decide now who inherits the records of people who left.

Import in the right order, with identifiers

HubSpot needs a unique identifier to update a record or link it to another one: the record ID, an email for contacts, a company domain name for companies. Without one, HubSpot states that the import creates duplicate records instead of associating each row with the same record. A few rules follow from the same import guide:

• Contacts need an Email column; companies need at least a Name or a Company domain name; new deals need a Deal name, a Pipeline and a Deal stage.
• To import associated objects, either put the identifiers of both records on the same row, or use two files connected by a common column.
• Blank cells don't clear existing values, so a second pass can fill gaps without erasing what is already there.

When contacts and companies are imported together, HubSpot doesn't let its automatic association setting override the associations from your file. What you import is what you get, so check those links in the file.

Protect the new portal from day one

The migration is also the right time to set the rules you lacked in the old CRM:

• Validation rules on text, number, date and phone properties: no leading or trailing spaces, a minimum length, numbers only, dates in a range. HubSpot applies them when data is edited or imported, though not in workflows or legacy forms (HubSpot: set validation rules for properties).
• Required properties per deal stage, so a deal can't move forward without its close date or amount (HubSpot: set up and customize pipelines).
• A short data dictionary: which property means what, who fills it, and when.

Audit right after the first sync

Even a careful migration leaves gaps. Run an audit as soon as the data is in, before the team starts working on it: duplicates that slipped through, companies without contacts, open deals without a close date. The HubSpot data audit checklist lists what to look for.

Forgeo does this automatically: connect HubSpot, and the first audit runs once the initial sync is done, with 40 checks and a data quality score for contacts, companies and deals.