Plan your Odoo migration with owners, stable identifiers and acceptance tests. Validate relationships and balances before approving go-live.
Migration must prove that people can work
An import without technical errors does not guarantee useful data. This guide for management and process owners proposes validating migration before ERP go-live. Odoo 19 documentation explains how to reuse external identifiers to update records. The plan must fit the installed version, modules and each company’s data.
Decide what to transfer and who accepts it
Separate master data, open transactions and historical reference data. Assign owners to customers, products, suppliers and every necessary dataset. Define mandatory fields, duplicates, units and criteria for archiving inactive records. Do not transfer everything by habit: retain access to necessary history and document exclusions. Agree how to handle incomplete information before loading it.
Maintain a stable identity
Prepare a mapping table between source and target systems. Avoid identifying companies solely by name or products solely by description. Preserve identifiers throughout rehearsals and review relationships: contacts to companies, products to units and orders to their lines. Repeating a load should produce an expected, verifiable result.
Illustrative functional test
A fictional distributor migrates customers, products and open orders. The team selects a representative order and checks customer, address, unit, quantity and terms. It then follows the process in a test environment. Counting records is not enough: an order linked to the wrong customer may pass a quantity check and still fail operationally.
Rehearse cutover and define stopping conditions
Set an extraction date, who can modify the source and how later changes will be incorporated. Repeat the load using the planned go-live procedure. Compare counts, relationships and relevant totals with business-approved reports. Define issues that block progress and a fallback plan accounting for transactions created since cutover.
Approve with evidence
The acceptance record should include tests, outstanding differences, owners and the go-live decision. CADT can help prepare the migration and its controls. To assess scope, provide a data inventory, anonymised samples and the processes that must work on day one. This grounds the proposal in verifiable criteria.



