Migrating to Odoo without stopping trading.
Nobody switches systems because they want to. They switch because the current one has become the constraint — on capacity, on reporting, or on the work they are allowed to bid for. This page sets out what moves across, how long it takes, and the method we use so that the old system stays available until the numbers agree.
What are you moving off?
From Sage
Nominal codes, VAT history, year-end timing and the bolt-ons around it.
How it mapsFrom QuickBooks or Xero
Where the books are fine but the five tools around them are the problem.
How it mapsFrom spreadsheets
When the business runs on a workbook only one person fully understands.
How it mapsRun both systems until they agree, then switch.
A big-bang cutover is how migrations become famous inside a company. This takes longer and costs less in disruption.
Document it, then extract while the old system is still live
You complete the discovery documentation, we verify it on site, then we take the messy export rather than a tidied one. What is wrong with the data is information about the process.
Load into a staging database
Deduplicated, mapped and loaded where nothing is at risk. Usually three or four passes before it is clean.
Reconcile against the old system
Trial balance, stock valuation, debtors and creditors, matched line for line. Your accountant signs the opening balances before go-live, not after.
Your team runs a real week in parallel
Same orders entered in both. Their objections change the build — this is the week that decides whether go-live is calm.
Final delta load, then switch
Only what changed since the last load. The old system stays readable for 12 months, and someone from us is on hand for the first week.
First month-end together
The month after go-live is the one that matters. We close the first one with you.
What comes across.
Always
Customers, suppliers, products, price lists, stock positions, open orders, open invoices, and opening balances.
Usually
Two to three years of transaction history, so comparatives and customer records still make sense in year one.
On request
Full ledger history. It costs more to reconcile than most businesses need, so we ask before assuming.
Rarely worth it
Dead stock codes, lapsed contacts and duplicate accounts. A migration is the one chance to leave them behind.
Where migrations go wrong.
Four failures, none of them technical. Worth reading before you appoint anyone, including us.
Nobody could settle a process argument
Implementations stall on unanswered questions, not on code. One person has to be able to decide in the room.
The old process was rebuilt exactly
Paying to reproduce a workaround in new software. If a step only exists because the old system forced it, it should not survive the move.
Go-live landed on year-end
Or on the peak trading month. The calendar is part of the scope, and it is usually the cheapest thing to get right.
Training was a demo, not a rehearsal
Watching someone else use the system teaches nobody. Your team has to run their own work through it before the switch.
Common questions.
How long does a migration take?
Typically 8–14 weeks from discovery to go-live, of which data work is 1–4 weeks. A single-entity business off one accounting package sits at the short end.
Will we have to stop trading?
No. The parallel run means both systems are live together, and the final switch happens on a quiet day of your choosing.
Can we keep our accountant?
Yes, and we would rather they were involved. They sign off the opening balances and we give them access to the staging database during reconciliation.
What happens to the old system?
Keep it readable for 12 months for audit and comparatives. You can usually drop to a read-only or archived licence rather than paying full subscription.
What does it cost?
Migration is part of a deployment rather than a separate product. The cost page breaks down all three components.
When is the best time to start?
Work back from your year-end and your peak season. Going live at the start of a quarter, away from both, makes the first close considerably easier.