Set the cutoff by object, not by date
People ask “how many years” as if one cutoff fits the whole CRM. It doesn’t. A contact from 2015 who renewed last month needs to move, while a lost deal from 2019 may not need to. Set the rule for each object, then check that the related records you keep still have their parents. An activity is useless if its contact stayed behind.
| Data | Typical rule | Why |
|---|---|---|
| Open deals, tickets, tasks | All of them | Work in progress can’t stop at cutover |
| Accounts and contacts | Active, plus anyone tied to an account you keep | Associations break if the parent is missing |
| Closed-won deals | As far back as your revenue and renewal reports go | Year-over-year reports need the prior years |
| Closed-lost deals | One or two sales cycles | Useful for win/loss analysis, little use after that |
| Emails, calls, notes, meetings | The same window as the records they belong to | They make up most of the volume, so set this cutoff with care |
| Attachments | Contracts and signed documents; skip the rest | File storage is often billed separately |
| Everything older | A searchable archive, not the new CRM | Kept for audits, not for daily use |
Retention rules apply to the archive, not to the new CRM
Finance and HR data often has a legal retention period. The US IRS, for example, says to keep records for 3 years in most cases, 6 or 7 years in specific cases, and employment tax records for at least 4 years. That obliges you to keep the data. It does not oblige you to load it into the new CRM. An archive with the original IDs usually meets the rule for much less effort.
What native tools do / where they stop
Native importers load whatever you give them. None of them helps you choose a cutoff. HubSpot, Salesforce and Zoho importers take files and match on keys, but they don’t check whether each activity’s parent record came across too. That check is the gap: if you filter the history by date and the accounts by “active”, you can end up with notes on contacts that were never imported.
Steps
- Count records by object and by year in the source.
- Agree on a rule for each object with sales, finance and support.
- Check that every record you keep has its parent records in the set you keep.
- Export the full source to an archive before cutover.
- Migrate the filtered set and reconcile the counts against the rules.
FlitStack AI can filter what moves and check parent links during the sample run. For ERP moves, most customers migrate 3–7 years of history, with a cutoff you can configure. Start from the CRM migration checklist or talk to us.