In short
A good CRM data migration starts at the front of the project. You decide what moves, count duplicates and empty records, deduplicate before the import and import in a fixed order so relationships survive. Then you test with a dry run and compare counts before and after. A CRM implementation with a migration takes 2 to 4 weeks with us for one pipeline and clean data. With several pipelines or a lot of history it takes longer, and that duration is in the quote. We quote a migration after we have seen the source system. Within a CRM implementation, from €2,950, it is one of the phases.
Source: Sales Surge, own implementation and migration projectsAs of: 16 November 2026
On this page
- Why dirty data breaks more than reports
- Scope the migration before you touch anything
- Count before you clean
- In which order do you import objects?
- How do you map fields from HubSpot, Zoho, Salesforce or Excel to Pipedrive?
- What happens to emails, attachments and activity history?
- Deduplicate before the import, never after
- Import in a way you can undo
- How long does a CRM migration take?
- Should you do the migration yourself or outsource it?
- What do you check for GDPR during a migration?
- Keeping it clean after go live
- Frequently asked questions
Migration looks technical, so it gets scheduled near the end, when the calendar is tight. Then the export holds three versions of your biggest customer, and the whole project is suddenly about data. Trust in the new system is won or lost here. So plan it first, and plan it in the order below.
Why dirty data breaks more than reports
Because the first wrong number someone sees decides whether they trust the system. The obvious cost of bad data is a wrong dashboard. The expensive part is what follows. Someone spots the error, the team quietly decides the CRM cannot be trusted, and from then on people keep their own list on the side. Now you have two sources of truth and you are paying for one of them.
Cleaning the data afterwards fixes the numbers but not the habit. That is why we treat migration as an adoption problem with a technical side, and why it belongs at the start of the project.
Scope the migration before you touch anything
Migrate what you will act on, archive what you might need to look up, and delete what you cannot justify keeping. Most teams default to moving everything because it feels safe. It is not. Every old record you carry across is one nobody can verify anymore, and it makes search worse for the whole team.
| Data type | Decision | Why |
|---|---|---|
| Open deals with their full history | Migrate | The operational core. Someone will act on it tomorrow. |
| Active customers and their contacts | Migrate, after deduplication | You work with them every week, so they have to be findable once. |
| Won deals | Migrate outcome, value and close date | Needed for reporting. Old notes usually stay in the archive. |
| Lost deals older than two years | Archive | Unless you actively approach them again. |
| Contacts with no activity in three years and no email address | Archive or delete | They are a GDPR risk, not an asset. |
| Custom fields nobody filled in for a year | Delete, do not recreate | An empty field in the old CRM stays empty in the new one. |
| Email threads and attachments | Migrate for open deals and active customers only | See the section on history below. |
Count before you clean
Export everything to a spreadsheet and count, before you decide anything else. This takes an afternoon and it changes every other decision in the project. The numbers are usually worse than the room expects, which is exactly why you want them early.
- How many organisations share a normalised name, ignoring legal suffixes and capitals?
- How many organisations share a domain but have different names?
- How many contacts share an email address, and how many have none at all?
- How many records have no owner, or an owner who has left?
- How many deals sit in a stage nobody has used in six months?
In which order do you import objects?
Import from the outside in: first the objects others depend on, then the objects that point to them. Every record keeps its old ID from the source system in a custom field, such as Source ID. That ID is the key that ties a deal to the right person and organisation in the next step. Without it you end up matching on names, and that is where relationships break.
- 1. Users and owners. Map every owner in the source to a user in the new CRM by email address. People who left get a fixed replacement owner.
- 2. Pipelines and stages. Create them first and map every old stage to a new one, so deals land in the right place.
- 3. Organisations, with their Source ID.
- 4. Persons, linked to their organisation through the organisation's Source ID.
- 5. Deals, linked to person and organisation through Source IDs, with pipeline, stage, value, currency and owner.
- 6. Activities: calls, meetings and tasks, linked to the deal and person, with their original date and done status.
- 7. Notes, linked to the deal, person or organisation they belong to.
- 8. Files and attachments, last, and only where they are needed.
How do you map fields from HubSpot, Zoho, Salesforce or Excel to Pipedrive?
Write a mapping sheet before you import anything: for every source field the target field and the rule that converts it. The table shows eight fields we map in almost every migration. Field names differ per system and per setup, so check your own export.
Two rules save you most of the trouble. Never map a free text field into a dropdown without first cleaning the values. And if a source field has no clear owner or purpose, leave it out. You can always add it later, removing it once the team uses it is much harder.
| Field | Typical name in the source | Pipedrive field | Conversion rule |
|---|---|---|---|
| Company | Company name (HubSpot), Account Name (Zoho, Salesforce), a company column (Excel) | Organisation name | Strip legal suffix for matching, keep the full name as the value |
| Domain | Company domain (HubSpot), Website (Zoho, Salesforce) | Custom field Website or domain | Lowercase, remove https and www, used as the duplicate key |
| Email (all systems) | Person email | Lowercase and trim, one primary address per person | |
| Phone | Phone (all systems) | Person phone | International format with country code |
| Deal name | Deal name (HubSpot, Zoho), Opportunity Name (Salesforce) | Deal title | Keep as is, add the company when it is empty |
| Stage | Deal stage (HubSpot), Stage (Zoho, Salesforce) | Pipeline and stage | Mapping sheet from every old stage to one new stage |
| Amount | Amount (all systems) | Deal value and currency | Decimal point, separate currency field, no text in the number |
| Owner | Owner (all systems) | Deal and record owner | Match on email address, replacement owner for people who left |
What happens to emails, attachments and activity history?
Records, activities and notes move well. Email history and some system data do not, so decide upfront what you need.
- Activities move with their date, type and done status, so the timeline of a deal stays readable.
- Notes move as notes. Long notes with formatting sometimes lose their layout, so check a few.
- Email threads usually cannot be rebuilt as real email conversations in the new CRM. The common choice: save the key emails on open deals as notes, and connect mailboxes in the new CRM so new mail links automatically.
- Attachments can be uploaded again and linked to the deal or person. That takes time per file, so limit it to open deals and active customers.
- Original creation dates cannot always be set on new records. Keep them in a separate field, so reports on age and cycle time still work.
- Change history and audit logs from the old system do not move. Keep the source export for that.
Deduplicate before the import, never after
Clean in the export and import once. Merging duplicates in a live CRM is painful and sometimes irreversible, because by then your team has attached activity to both versions.
Matching on company name alone is the classic mistake. Names differ by legal form, by spelling and by whether someone typed the trade name. Domain is a much stronger signal, with name as a second check.
- Match organisations on email domain first, then on a normalised name.
- Match people on email address, and treat name matches as suggestions a person confirms.
- Decide in advance which record wins a merge: usually the one with the most recent real activity.
- Never merge automatically on a fuzzy match. Two similar names can be two real companies.
- Keep a log of what was merged. Within a month someone will ask about one specific customer.
Import in a way you can undo
With a dry run on real data, judged against criteria you set beforehand. Never treat the first import as the real one, especially when the data looks clean.
The rehearsal is not about whether the import runs. It is about looking at real records and noticing what no validation rule catches. Our tool Surge Shift is built around that: mapping, staging and a dry run before anything touches the live CRM.
- Run the import into a test environment or a staging area first.
- Pick ten records you know personally, from different owners and pipelines, and check them field by field.
- Compare counts per object before and after. They must match the counts you decided to move, minus what you deliberately left out.
- Check relationships: is each person on the right organisation, each deal on the right person?
- Check the boring things: currency, date format, decimal separator, country codes, phone formats.
- Agree who signs off. Only after that sign-off does the real import run.
- Keep the source export untouched and findable for at least six months.
How long does a CRM migration take?
The import itself takes minutes. The preparation takes the time. A CRM implementation with a migration takes 2 to 4 weeks with us for one pipeline and clean data. With several pipelines or a lot of history it takes longer, and that duration is in the quote. The data migration within that takes 1 to 4 weeks, depending on the age of the source system and the number of sources.
Start the export on day one, not in week four. Our article on how long a CRM implementation takes breaks down the other phases.
Should you do the migration yourself or outsource it?
You can do it yourself when there is one source, a few thousand records, and you only need open deals and active contacts. Pipedrive's own import handles spreadsheets and maps columns to fields, and our HubSpot to Pipedrive checklist walks through the steps.
Outsourcing makes sense when there are several sources, activity history has to move with its links intact, custom objects need to be translated, or the team cannot afford a week of errors. We quote a migration after looking at the source system, because the work depends on what is in it. As part of a CRM implementation, which starts at €2,950, it is one of the phases. See our CRM migration service and HubSpot to Pipedrive migration.
What do you check for GDPR during a migration?
A migration is the moment to stop holding what you should not hold anymore. Three things to check:
- Retention: compare your retention policy with the export. Contacts without a relationship or activity for years stay behind.
- Processing agreement: whoever handles the export, including a partner like us, needs a data processing agreement with you.
- Data location and copies: know where staging copies live, do not send exports by email, and delete working copies once the migration is signed off.
Keeping it clean after go live
With two rules: a duplicate check every month, and one person who can say no to new fields. A migration gives you one clean moment. Without those rules the data starts degrading in week one, mostly through a rep creating a new organisation because search did not find the existing one, repeated a few hundred times a year.
- Run the duplicate check monthly, not yearly. Small batches stay reviewable.
- Give one person the authority to say no to new fields. Field sprawl is data decay in slow motion.
- Fix the intake, not only the database. Most duplicates enter through forms and imports, not by hand.
- Make search good enough that finding the existing record is easier than creating a new one.
Frequently asked questions
What does a CRM data migration cost?
That depends on the number of sources, the amount of history and how dirty the data is, so we quote it after looking at the source system. As part of a CRM implementation, which starts at €2,950, the migration is one of the phases. A simple move of open deals and active contacts from one source costs far less than a rebuild with full activity history.
Can I migrate from HubSpot to Pipedrive myself?
Yes, if it is one account, a few thousand records and you mainly need open deals and active contacts. Export per object, clean in a spreadsheet, map stages and owners, and import in the order organisations, persons, deals, activities. Test on ten known records first. With years of activity history or custom objects, it usually pays to outsource it.
Should we migrate all our history?
Almost never. Open deals need their full history because someone will act on them tomorrow. Won deals mostly need the outcome, value and date for reporting. Everything else can live in an archived export that you keep findable. Migrating everything feels safer and mainly imports noise that makes search worse for the whole team.
Can we clean the data after we go live?
You can, and it costs several times more. Once the team logs calls and emails on duplicate records, merging is no longer a spreadsheet job. You are deciding which activity history survives. Clean before the import, in the export, where undo means closing the file without saving.
How do we deduplicate reliably?
Match organisations on email domain and people on email address, then use name similarity as a second pass that a person confirms. Decide beforehand which record wins, usually the one with the most recent real activity. Log every merge, because within a month someone will ask what happened to one specific account.
What about GDPR when migrating?
A migration is a good moment to remove what you should not still hold. Compare your retention policy with the export and leave behind what you cannot justify keeping. Make sure whoever handles the data has a processing agreement with you, know where the staging copies live, and delete working copies once the migration is signed off.
How long does a migration take?
The import itself takes minutes. The preparation takes the time and grows with the age of the source system. A CRM implementation with a migration takes 2 to 4 weeks with us for one pipeline and clean data. With several pipelines or a lot of history it takes longer, and that duration is in the quote. The data migration within that takes 1 to 4 weeks, depending on the age of the source system and the number of sources. Start the export on day one of the project.
