Insights / Checklist / Enterprise software implementation
ChecklistData migration readiness checklist
Poor data shows up on day one of a new system. Starting migration work early, and rehearsing it, turns a major go-live risk into a routine.
When a new system goes live, the first thing users notice is not the new screens. It is the data. A customer with three duplicate records, a product with the wrong unit of measure, an opening balance that does not match the old ledger: each one erodes confidence in the system on the day it matters most.
Poor data shows up on day one, but it is created long before. Starting early turns a risk into a routine. Use this checklist from the design phase onwards, and review it at each mock migration and before the final cut-over.
Decide what to move
Moving everything is rarely necessary and often harmful. Old, incomplete or obsolete records slow the migration and clutter the new system. Migrate current balances and open items, and archive history you rarely use.
- Each data object is listed: customers, suppliers, products, prices, open orders, balances, assets and so on.
- For each object, the scope is agreed: active records only, a date cut-off or a full history.
- Open transactions are distinguished from closed history, with a plan for each.
- Historical data that is not migrated has an accessible archive, with retention rules confirmed.
- Legal and audit requirements for record retention have been checked with finance and compliance.
Assign ownership
Data quality is a business responsibility, not an IT task. The people who use customer, product or supplier data every day are the ones who can tell whether it is correct.
- Each data object has a named business owner who signs off its quality.
- Owners understand what fields the new system needs and what each one means.
- Time is allocated in owners’ plans for cleansing and validation, not squeezed in alongside normal work.
- A migration lead coordinates owners, technical team and the cut-over plan.
Clean at the source
Clean data in the current system or in controlled extracts before it is loaded, not in the new system afterwards. Merge duplicates, fix formats and fill missing fields before loading. Cleansing after go-live happens under pressure, while users are trying to work.
- Duplicate customers, suppliers and products are identified and merged, with a rule for which record survives.
- Formats are standardised: addresses, tax numbers, phone numbers, units of measure, dates.
- Mandatory fields for the new system are populated, such as payment terms, tax codes and item categories.
- Inactive records are marked or excluded according to the agreed scope.
- Data quality is measured with simple counts: duplicates found, fields missing, errors remaining.
- A freeze date is set after which master data changes are controlled until go-live.
Rehearse the cut-over
The first full migration attempt almost always reveals problems: mappings that do not work, fields that are too long, records that fail validation. That is why it should never be the real one. Run mock migrations until the result is predictable and reconciled.
- At least two or three full mock migrations are planned, using recent production extracts.
- Each mock is timed end to end, so the cut-over window is realistic.
- Load errors are logged, fixed at the source or in the mapping, and retested.
- Business owners check samples of migrated records in the new system, not just counts.
- The final mock runs cleanly, within the planned window, with no unexplained differences.
Reconcile and sign off
Every migrated balance and every key count should be proven against the source. Reconciliation is what gives finance, auditors and users confidence that the new system starts from the right position.
- Record counts match between source and target for every object.
- Financial balances reconcile by account, customer and supplier, with differences explained.
- Stock quantities and values reconcile by location and item.
- Open orders, invoices and payments reconcile in number and value.
- Each business owner formally signs off their data before go-live is approved.
Plan for the gap between extract and go-live
Transactions keep happening while the final migration runs. Agree how orders, receipts and payments made during the cut-over window will be captured and entered, and who will check them.
Getting started
Run a quick data profile of your main master data now, even before the new system is chosen: count duplicates, missing fields and inactive records. The results will show how much cleansing lies ahead and give business owners an early, concrete reason to start.
Source: Intelliversal analysis.
