ERP data migration is moving data from the old system into the new ERP correctly, completely and in a usable form. Often postponed with a "we'll deal with it later", it is one of the most common causes of delays. However well the new system is designed, it is only as good as the data that goes into it.
1. Define the scope
Moving everything is neither possible nor necessary. A typical scope looks like this:
- Master data: customers, suppliers, materials, bills of materials, prices, chart of accounts
- Open items: open orders, open receivables and payables, stock quantities and values
- Balances: G/L balances, fixed assets
- History: usually not migrated; kept read-only in the old system or archived
2. Cleanse early
Duplicate customers, unused materials, missing addresses, inconsistent units… Cleansing should start months before migration. Ask the business "which records are really active?" and document the decisions. In S/4HANA projects this is also a prerequisite for the Business Partner migration.
3. Write the mapping rules
A mapping document defines where every old field goes in the new system and how it is transformed: code lists (payment terms, units, tax codes), numbering and default values for mandatory fields. This document is the contract between the technical team and the business.
4. Run mock loads
Plan at least two, ideally three, mock loads. Each one measures the error rate, duration and reconciliation results. The last one should be a cutover rehearsal run on the real schedule. You cannot build a cutover plan without knowing how long the loads take.
5. Reconcile and sign off
Migrated data is validated in two ways:
- Numeric reconciliation: record counts, total stock value and total receivable and payable balances are compared with the old system.
- Sample checks: the business checks randomly selected records field by field.
Each data object needs written sign-off from its business owner. Data without sign-off does not go live.
Common mistakes
- Leaving cleansing to IT, when only the business knows which record is correct
- Loading straight into production without mock loads
- Not designing how open orders are migrated and re-keying them after go-live
- Not planning for transactions between the final load and go-live
Data migration is the invisible backbone of an ERP project. In ERP consulting engagements, giving the data package an owner and a budget from day one is the strongest safeguard for go-live.
