Skip to content
Monday, November 16, 2026Newsletter
Advertisement
Odoo ERP Solutions

ERP Data Migration Testing Checklist Before Go-Live

A phased checklist for validating your ERP data migration before go-live, covering row counts, business logic, user acceptance, and cutover dry runs.

By

Founder

6 min read

ShareLinkedInXEmail
Junious Digital Labs Insights title card with the headline "ERP Data Migration Testing Checklist Before Go-Live" and the Odoo ERP Solutions label

Moving your business data into a new ERP system is one of the riskiest transitions you will make. A solid erp data migration testing checklist is what separates a smooth go-live from one that breaks invoices, scrambles inventory, or delays payroll. If you are a small business owner or operations manager preparing for an ERP cutover, this is for you.

By the end, you will have a phased checklist covering data validation, dry-run cutovers, and user-acceptance testing. Work through it before go-live and you will catch errors while they are still cheap to fix, not after they reach a customer or a paycheck.

Why Data Errors Hide Until Go-Live

Most ERP migrations look clean in a test environment because test environments are quiet. You load a sample of records, run a few transactions, and nothing breaks. The problem is that real business data is messier than any sample you hand-picked.

Field mismatches are a common culprit. Your old system stored a customer name in one long field. The new system splits it into first name, last name, and company. During migration, the split logic works on simple names and fails silently on names with suffixes, abbreviations, or punctuation. You will not see the failure until a sales rep pulls up a record and the name looks wrong, or an invoice prints incorrectly.

Truncated records follow the same pattern. A legacy text field held 500 characters. The new field holds 255. The migration tool cuts the data without an error message, and no one notices until someone reads the truncated note six weeks later.

Advertisement

Broken relationships are the hardest to catch. Orders linked to customers, line items linked to products, contacts linked to accounts: if any parent record migrated with a different identifier than the child record expects, the link silently breaks. Those orphaned records sit invisible until a process tries to touch them.

Use This ERP Data Migration Testing Checklist

Run your testing in three phases. Each phase catches a different class of problem, and skipping ahead wastes time because later tests depend on earlier ones passing first.

Advertisement

Phase one is data validation. Confirm that every record landed in the right table, that required fields are populated, and that identifiers match across related tables. Focus on the parent-child links described above. Run row counts against your source system. Flag any difference, even a small one.

Phase two is business logic. This is where you run real transactions: create an order, receive a payment, generate an invoice, process a return. Use values that reflect your actual business, not the sample data the vendor loaded. Watch for rounding errors, tax calculations, and unit-of-measure conversions. These rarely break in simple tests and almost always surface in real ones.

Phase three is user acceptance. Put the people who will use the system daily in front of it. Give them a script of their most common tasks. They will find workflow gaps and labeling confusion that no technical tester will catch.

Document every failure. A failure that gets fixed without a written record tends to reappear.

Phase One Row Counts and Field Mapping Spot-Checks

Phase one is a simple sanity check, but it catches more problems than most teams expect.

Advertisement

Start by pulling record counts from your source system for every major object: customers, vendors, products, invoices, purchase orders, employees. Then pull the same counts from your new ERP after the migration load. The numbers should match. If they do not, stop and find out why before you go further. A missing five hundred customer records is easier to explain now than after go-live.

Next, do field mapping spot-checks. Pick a sample of records from each object type and compare them side by side. Look at every field that matters to daily operations: names, addresses, account numbers, unit prices, tax codes, payment terms.

Watch for truncation, where a long value got cut off silently. Watch for encoding errors, where special characters became garbled text. Watch for blank fields that should hold data.

Pay close attention to fields that drive calculations. A wrong tax code or a missing unit of measure will cause transaction errors the moment someone tries to use the record.

Advertisement

Log every mismatch in your issue tracker, even small ones. A pattern of small errors usually points to a mapping rule that needs to be corrected at the source.

Phase Two Business Logic and Open Transaction Validation

Phase two goes deeper than row counts. You are checking whether the data behaves correctly inside the new system, not just whether it arrived.

Start with open invoices. Pull every unpaid invoice from the source system and match it line by line against the new ERP. Check the invoice date, due date, amount, customer or vendor, and payment terms. A wrong payment term can break your aging report and trigger incorrect late notices on day one.

Next, check inventory on-hand balances. Compare the quantity and value at the item level, not just the warehouse total. Costing method mismatches, such as FIFO versus average cost, produce totals that look close but are wrong in ways that compound over time.

Advertisement

Then validate open payroll records and any accrued liabilities. These are business assets in the same category as equipment and inventory (SBA). Errors here affect compliance reporting, not just your books.

Finally, check every calculated field: running totals, credit limits, reorder points. These fields are rarely migrated directly; they are usually rebuilt from source data. A flawed mapping rule can produce a wrong result even when the underlying numbers transferred correctly.

Phase Three User Acceptance Testing With Real Transactions

Phase three is where real users replace the testing team. Pull three to five employees who work in the system daily: an AR clerk, a buyer, a warehouse lead. Give each one a script of transactions they already processed in the legacy system last week.

Have them enter the same transactions in the new system. Then compare outputs side by side. Look for differences in totals, tax calculations, inventory adjustments, and printed documents like invoices or purchase orders. Users will catch formatting problems and missing dropdowns that no automated script would flag.

Run at least one full order-to-cash cycle and one procure-to-pay cycle before you sign off. If your business carries inventory, run a cycle that triggers a reorder. Edge cases surface during UAT that never appear in controlled test data.

Advertisement

Before you close out testing, confirm your backup posture. Maintain offline, encrypted backups of critical data, and regularly test the availability and integrity of backups in a disaster recovery scenario (CISA). A go-live that fails without a tested restore path is a crisis instead of a setback.

Document every defect users find, assign an owner, and retest that specific transaction before cutover.

Advertisement

Cutover Dry Run and Go No-Go Criteria

Schedule a full cutover rehearsal the week before go-live. Run every step of your actual cutover plan, timed, with the same people who will execute it for real.

During the dry run, migrate a recent data snapshot into a clean environment. Then run your phase one and phase two checks again. If row counts are off or open transactions are broken, you have found a process gap, not just a data gap.

Define your go/no-go criteria before the dry run, not after. Write down three things: the maximum number of open defects you will accept at cutover, the specific transaction types that must pass with zero errors. The person with authority to call it off. Without those thresholds on paper, the pressure of a cutover date will override good judgment.

Set a rollback trigger point. Decide in advance how far into cutover you can go before reverting becomes harder than pushing forward. Know which systems stay live during migration and for how many business days. If the dry run exposes a critical gap, delay. A one-week slip hurts far less than corrupted financial records after go-live.

Before go-live, run your full dry-run checklist one final time and document every result. If any critical validation fails, treat it as a stop signal, fix the root cause, and rerun the test rather than carrying the problem into production.

For a related guide, see How Odoo Enterprise Per-User Pricing Works and What It Costs. If you want help with this work, see our Odoo ERP services and Odoo development and implementation pages.

If you want help with this, book a free 30-minute call with Joshua and talk it through.

Frequently asked questions

Why do data errors not show up until after go-live?
Test environments use hand-picked sample data that is cleaner than real business data, so problems like silent truncation, field mismatches, and broken parent-child record links only surface when real transactions touch the migrated records.
Who should be involved in user acceptance testing?
Employees who work in the system daily, such as an accounts receivable clerk, a buyer, and a warehouse lead, should run the same transactions they processed in the legacy system and compare outputs side by side.
What should you do if the cutover dry run reveals a critical problem?
You should delay the go-live, fix the root cause, and rerun the test rather than carrying the problem into production, because a short delay causes far less harm than corrupted records after cutover.

Sources

  1. These are business assets in the same category as equipment and inventory ( SBA ). sba.gov (accessed 2026-11-15)
  2. Maintain offline, encrypted backups of critical data, and regularly test the availability and integrity of backups in a disaster recovery scenario ( CISA ). cisa.gov (accessed 2026-11-15)
Advertisement

Outgrowing spreadsheets and disconnected apps?

Book a free 30-minute call. We will map how your business runs today and whether one connected system makes sense for you.

Book a free Odoo call

Related service: Odoo ERP Solutions

Junious Digital Labs logo

Joshua Junious

Founder of Junious Digital Labs, a veteran-owned IT consultancy. Army National Guard veteran and former IT director. About Joshua

Reach readers across the Junious Media Network. Advertise with us →