Data migration is the least spectacular and most delicate phase of a CRM project. It is hard to see, hard to correct, and it decides whether your team trusts the new system. These are the points that should be settled before any import.
Before the export
- Quantify duplicates. How many contacts exist twice? Under 5 percent you clean up after the import; above that, before.
- Decide the required fields. Which fields must be filled for a record to be useful at all? Everything else is optional – and stays optional.
- Check ownership. Every record needs an owner. Records without one will be maintained by nobody after the migration.
- Check your legal basis. For which contacts do you have grounds for further processing? A migration is the right moment to retire old lists.
The order
Pipedrive links records by reference, so import order is not arbitrary:
- Organisations
- People, linked to organisations
- Open deals, linked to people and organisations
- Activities and notes
- Closed history – optional, and last
Run every stage with twenty test records first. Field-mapping mistakes then surface in minutes instead of after 8,000 rows.
What to leave behind
The reflex to "bring everything" is the most common mistake. As a rule, these should not move:
- Contacts with no interaction in the last three years
- Deals that have sat in the same stage for over a year
- Free-text fields whose content nobody can interpret any more
- Fields that contain only a single distinct value
After the import
Do not verify the total count – verify samples: ten records per object type, reviewed in full. And have the review done by the people who will work with the data later, not by the person who ran the import.
If you are unsure at this point, we are happy to talk through your data set beforehand. An hour of preparation quickly saves a day of rework.
