The main risks are incomplete or mis-mapped master data, open transactions that cross the cutover date, business logic hidden in NAV or Business Central customisations, and operational steps that staff have not rehearsed. Each risk can be reduced with a field-level data map, a dated cutover plan, reconciliation checks and a parallel run before go-live.
Key takeaways
- Master data mapping errors cause more post-go-live problems than any other single step.
- Open purchase, sales and inventory transactions must be closed or re-entered with a documented rule.
- Custom code and extensions must be inventoried, because their business rules are often undocumented.
- A reconciled parallel run and a rollback plan turn cutover from a single bet into a controlled process.
Why does a NAV or Business Central migration carry data risk?
A migration carries risk because the source system has usually been shaped by years of configuration, manual habits and local workarounds. Microsoft Dynamics NAV and Dynamics 365 Business Central can each hold posting logic, number series and reporting rules that nobody wrote down. Moving the visible data does not automatically move that logic.
Odoo organises the same business objects differently, so a field in one system may split into two records in another. Treat the move as a redesign of the data model, not a copy operation. For a broader view of how integrated systems fit together, see our guide on maximising business efficiency through ERP and CRM integration.
Which data risks should you check first?
The first data risks to check are master data, open transactions, history and customisations. Each one fails in a different way, so each needs its own test. Start with the data that drives daily operations, then work outward to archives.
Master data: customers, suppliers, items and accounts
Master data errors are the most common source of trouble after go-live. Duplicate customers, inconsistent item units and unmapped general ledger accounts all flow into invoices and stock valuations. Build a mapping table for every master object and sign it off with the finance and operations owners.
Open transactions and balances
Open sales orders, purchase orders, unposted journals and partially shipped deliveries need an explicit rule. You can close them in the old system, re-enter them, or carry them over with their history. Each option changes audit trails, so agree on the rule before extraction begins.
Historical data and archives
Moving full history is slower and riskier than most teams expect. Many organisations migrate only the opening balances and keep the old system in read-only mode for reference. Decide which years must be queryable in Odoo and which can stay in an archive.
Customisations and extensions
Custom objects, reports and add-ons often encode rules such as approval thresholds or pricing exceptions. Make an inventory of every extension, record what business rule it implements, and decide whether Odoo already covers it. Undocumented logic is the riskiest category because nobody notices it is missing until a process fails.
Which process risks appear at cutover?
Process risks appear when people, timing and systems disagree about who does what and when. A data migration can be technically correct and still stop invoicing for a day. The cutover plan should therefore cover staff roles, freeze windows and integrations as well as the data.
Integrations are a frequent weak point. Connections to banks, e-commerce platforms, warehouse tools and payroll must be repointed or rebuilt, and each one needs its own test. For a structured approach, our system integration service covers this type of connection work.
Freeze windows matter because transactions keep arriving during the migration. Decide when the source system stops accepting new postings and who approves exceptions. Without this rule, the final extract will not match the balances you reconciled the week before.
Which migration approach fits your situation?
A big-bang cutover is fastest but leaves no fallback, while a phased approach reduces risk and extends the period of double running. The right choice depends on how many entities, sites and integrations you must move at once. The table below compares the options with qualitative criteria.
| Approach | Main advantage | Main weakness | Best suited to |
|---|---|---|---|
| Big-bang cutover | Single switch date and simple governance | Little room to recover if reconciliation fails | Small, well-documented data sets with few integrations |
| Phased by module | Limits the scope of each failure | Two systems must stay synchronised for longer | Organisations with separate finance and operations timelines |
| Phased by entity or site | Lets the team learn from the first wave | Consolidated reporting is harder during the transition | Groups with several legal entities or warehouses |
Each approach fails in predictable ways. Big-bang fails when reconciliation is late and there is no tested rollback. Phased approaches fail when the interim bridge between systems is poorly maintained and balances drift apart.
How does GDPR affect moving personal data?
GDPR applies to the personal data inside customer, supplier and employee records, so the migration itself is a processing activity. The regulation requires that personal data be processed lawfully, fairly and for specified purposes, and that it be kept accurate. The official text is available on EUR-Lex (Regulation (EU) 2016/679). This article is general information and not legal or tax advice.
In practice, limit the data you copy to what the target system needs, restrict access to extracts, and delete temporary files after use. Record the migration in your processing register and agree retention periods for the archived source system. Rui Codex provides technical implementation support for these steps, and the legal assessment remains with your organisation and its advisers.
Hypothetical examples of cutover failures
Here are two hypothetical scenarios: in the first, a distributor maps two customer accounts with the same tax number into one Odoo partner, and credit limits from both accounts merge unexpectedly. The fault appears only when the sales team reviews open orders in the first week. In the second, a manufacturer forgets an approval rule that lived in a custom NAV codeunit, so purchase orders above a threshold post without sign-off until a monthly audit finds them.
Both scenarios could have been caught earlier. The first needed a duplicate check on tax identifiers before import. The second needed a written inventory of custom logic and a test that submits an order above the threshold in the target system.
How do you plan a safe cutover?
A safe cutover starts with a documented plan and ends with signed reconciliation. Work through the steps below in order, and do not schedule go-live until each one is closed.
- Inventory every master object, open document, report, integration and custom extension in the source system.
- Write a field-level mapping for each object and agree the rules for open transactions.
- Run at least two trial loads and reconcile balances, open items and stock values after each one.
- Rehearse the freeze window, the final extract and the rollback decision with the staff who will use the system.
- Run a parallel period for critical processes, such as invoicing and purchasing, and record every difference.
For projects where the data model or integrations are complex, a ERP and CRM integration review can help identify these dependencies before the plan is fixed. Rui Codex is led by a founder with 200+ projects delivered, and published project work is listed on the case studies page.
Frequently Asked Questions
Can we move all historical transactions from Business Central to Odoo?
Yes, but it is rarely the best first step. Moving full history increases load time and the number of records that must be checked. Most teams migrate opening balances and keep the old system in read-only mode for audit queries, which keeps the cutover smaller and easier to reconcile.
Why do open purchase orders cause problems during migration?
Open purchase orders hold committed supplier obligations and goods that may be partly received. If they are re-entered incorrectly, receipts and invoices will not match. Agree a rule for each open document type and test receipts and supplier invoices against it.
Do customisations in NAV need to be rebuilt in Odoo?
Customisations need to be assessed, not automatically rebuilt. For each one, decide whether standard Odoo functionality covers the business rule, whether configuration is enough, or whether custom development is justified. Document the decision so the reasoning survives staff changes.
How do we keep personal data safe during a migration?
Limit extracts to the fields the target system needs and restrict who can access them. Delete temporary files after each load and record the activity in your processing register. The GDPR text on EUR-Lex sets the obligations, and your data protection lead should confirm how they apply to your organisation.
How long should we run both systems in parallel?
The parallel period depends on the complexity of your processes and how many integrations are involved. Keep it long enough to complete at least one full month-end and reconcile the results. Define exit criteria in advance so the decision to stop running the old system is based on evidence.
Sources
- Regulation (EU) 2016/679 (General Data Protection Regulation) - EUR-Lex, Publications Office of the European Union
- Odoo Documentation - Odoo S.A.
- Dynamics 365 Business Central documentation - Microsoft Learn