Loading
In order. The order is what stops the expensive surprises.
| # | Stage | Task | Why | Done |
|---|---|---|---|---|
| Before | Agree the stage names in the new system | Renaming stages after import means re-mapping every deal | ||
| Before | Export a full copy of the old system and keep it | The rollback plan is a file, not an intention | ||
| Before | List custom fields and decide which survive | Half of them are somebody’s abandoned experiment | ||
| Before | Decide how much history comes across | Three years is enough for almost everyone | ||
| Export | Companies and contacts, with the link between them | Contacts imported without their company are the classic loss | ||
| Export | Open deals, with owner, stage, value and close date | |||
| Export | Closed-won and closed-lost, with dates | Without these the forecast has no baseline for months | ||
| Export | Activity and notes, at least for open deals | This is the second thing that always gets dropped | ||
| Import | Load into a sandbox first, never straight to live | |||
| Import | Check five records by hand against the old system | Row counts match while fields are silently empty | ||
| Import | Re-point integrations and forms one at a time | |||
| Switch | Run both systems for one full week | |||
| Switch | Make the old system read-only rather than deleting it | |||
| Switch | Tell everyone the date the old one goes away |
Tasks 7 and 8 are the ones people skip and then miss for a year.