Changing till software frightens people for one reason: the fear of losing data. That fear is justified, and the answer fits in one sentence: you never migrate everything, you migrate what is useful, and in the right order. This guide sets out the full method. It is written to be useful to you even if, in the end, you decide to stay where you are.
The short answer
- Five things migrate: categories, products, customers, stock, and sales history. They do not migrate in any order, and not with equal ease.
- Sales history is the only real problem. It exports almost always, and re-imports almost never. Keep it as an archive rather than trying to recreate it.
- uniCenta stores everything in MySQL or MariaDB, so all of it is extractable with a SQL query. That is the most favourable case.
- With Aronium, go through the application’s exports and back up the database before anything else.
- One weekend is enough for a business with a thousand product lines, provided the files were prepared the week before.
The solutions compared
The three programs on this page are desktop applications that keep your data with you. That is what makes a migration possible without depending on anyone, and it is a shared point worth making before the differences.
uniCenta oPOS
A Java desktop till published under the GNU GPL v3 free licence. Multi-location, multi-terminal, multi-user, 17 languages, backed by a MySQL or MariaDB database. Version 5.4.0 added a WooCommerce integration. The project is distributed on SourceForge, where its last update is dated 10 August 2025.
Aronium
A native Windows till in two editions: Aronium Lite, free and with no registration, and Aronium Pro, which connects several registers to one database. Everything else is added through plugins, free or paid, billed per store.
BelloPOS
Our Windows till software, built in Morocco. It runs 100% offline: no sale depends on the connection. The interface and the receipts exist in French, Arabic and English. Three tiers: Lite free for life, Go and Pro bought once, never subscribed to.
The comparison, criterion by criterion
What each program lets you take with you, and by which route.
| The criterion | uniCenta oPOS | Aronium | BelloPOS |
|---|---|---|---|
| Where your data lives | MySQL or MariaDB database, on your server or machine. | The application’s local database, on your Windows machine. | Local database, on your Windows machine. |
| Catalogue export | SQL query on the product and category tables, or an application export. | Export from the application, to prepare before cancelling any plugin. | Catalogue import, and loading help included in Go and Pro. |
| Customer export | SQL query on the customer table. | Application export of customer records. | Customer import, with balances. |
| Stock export | Quantities extractable by query. | Export of quantities and movements. | Entry or import of quantities, to be done after a count. |
| Sales history | Fully extractable in SQL, but rarely re-importable elsewhere. | Exportable as reports, rarely re-importable elsewhere. | Keep the old history as an archive, do not try to re-inject it. |
| Backup before migrating | SQL backup of the database, done twice and on two media. | Backup of the application database. | Automatic backups included in Go and Pro. |
| Overall difficulty | Moderate if you can write a query, high if not. | Low to moderate: everything goes through the interface. | Low: file imports, plus loading help on Go and Pro. |
| Help provided | Community and documentation. | Online help centre. | 20 products and 10 categories loaded with Go, 1,000 products and 100 categories with Pro. |
| What is most often lost | Custom screen configurations. | Plugin-specific settings. | Nothing structural, but old history stays as an archive. |
| Realistic duration | A weekend, plus a week of preparation. | A weekend. | A weekend, less if the catalogue is small. |
The method, step by step
Here is the order that works. It is not negotiable: importing products before categories, or stock before products, guarantees starting over.
- Back up before touching anything. Two copies, on two different media, one of them off the machine. This step is not a precaution, it is the migration itself: everything else can be redone as long as it exists.
- Export in this order: categories, products, customers, stock. A category needs nothing, a product needs its category, a stock level needs its product. Produce each file separately and open it to check it is not empty.
- Clean the catalogue while it is in a spreadsheet. It is the only moment in a business’s life when you will see all your products on one screen. Delete duplicates and lines you no longer sell, and fill in missing purchase prices. A migration is a chance to tidy up, and many shopkeepers gain more from that cleanup than from the change of software.
- Import into the new software, and check twenty products. Not a thousand: take twenty familiar lines and check the selling price, the purchase price, the category and the barcode. If those twenty are right, the rest are.
- Count the physical stock on switchover day. Never carry over the old system’s theoretical quantities without checking them: you would be importing its accumulated errors. A count on a Sunday evening beats a wrong stock figure for six months.
- Run both systems for a week. The old one read-only, the new one live. It is the only real safety net, and it costs nothing since the old one is already installed.
Sales history: the question to settle early
This is where most migrations get stuck, so let us say it plainly: sales history practically never transfers from one program to another, and that is not a fault in the software. Each structures its tickets, lines, taxes and payment methods its own way, and an approximate import produces wrong numbers, which is far worse than no numbers.
The right practice is simple. Export the whole history to a readable file, CSV or PDF, file it with your accounts, and start fresh in the new software. Your retention obligations are met by the archive, your year-on-year comparisons run off that file, and your new till starts clean. Just plan to keep the old software installed, or a machine able to open it, for the legal retention period.
The typical weekend of a successful migration
Friday evening: full backup, two copies. Saturday morning: export categories, products and customers, then clean the catalogue in a spreadsheet. Saturday afternoon: import into the new software and check twenty products. Sunday: physical stock count and entry of the real quantities. Sunday evening: ten test sales, one return, one split payment, one receipt printed in Arabic. Monday morning: you open on the new system, with the old one still installed just in case. That schedule holds for a business with a thousand product lines.
How to choose: the twenty-minute method
The buying method stays the same, but a migration adds four checks worth making before committing.
- Count your real needs before you look at a single price. How many tills, how many users, do you invoice businesses, do you need stock, do you take delivery orders. Five answers, and half the offers rule themselves out. Our guide to choosing a till works through this questionnaire.
- Test the checkout, not the demo. Ask to ring up ten items, one return and one split payment, yourself, and time it. A well-rehearsed demo tells you nothing about what your cashier will face on a Saturday.
- Pull the internet cable and try to sell. That one test splits the whole market in two, and it takes thirty seconds. We go through it in detail in our offline or cloud POS comparison.
- Ask for an export of your own data before you pay anything. Products, customers, sales, in a file you can open. If the answer is vague, you already know how this ends.
- Add up five years, not one month. A subscription is compared with a one-off purchase over the time you will actually use it.
- Check that the receipt prints correctly in Arabic before you commit. On a thermal printer that is not a given, and it takes a minute to see.
- Ask who answers the phone on a Saturday at 8pm, and in which language. That is exactly when a till breaks.
- Ask for an export of your data before paying for the new software, to confirm getting out will be possible later too.
- Test the import with a twenty-line file before running the thousand.
- Keep the old software installed for a month. It costs nothing and it will spare you a panic.
- Never migrate on a busy day, nor the day before a stocktake or a tax deadline.
What each solution does better
uniCenta oPOS
Its licence. uniCenta oPOS is free software under the GNU GPL v3: you can install it on as many machines and in as many shops as you like, with no licence fee, and nobody can withdraw your right to use it. The database is yours, readable, on your own server. If independence from a software publisher matters to you more than anything else, this is the right answer and we are not going to pretend otherwise. Its multi-location support and its fifty reports are serious work, and rare at this price.
Aronium
Its plugin model, and its value for money. Aronium Lite is free, complete and asks for no registration: you download it and you sell. Aronium Pro at 100 dollars a year per store is honestly cheap for what it does, and the free plugins (stock control, expiry dates, named orders, takeaway, notes, dual currency) cover needs that others charge for. The publisher keeps a detailed changelog, which is a good sign about a piece of software’s health.
Mistakes to avoid
- Cancelling before exporting. It is the irreversible mistake, and it happens more often than you would think.
- Importing the old system’s theoretical quantities. You import its errors and treat them as truth for six months.
- Insisting on recovering the sales history. Archive it and move on: that is the right call in almost every case.
- Migrating on a month-end Friday. Choose the quietest weekend of your year.
- Not checking purchase prices after the import. A catalogue with no purchase prices gives you turnover with no margin.
Frequently asked questions
Can you migrate from one POS to another without losing anything?
You can migrate the catalogue, categories, customers and stock without loss. Sales history rarely transfers between systems: the right practice is to archive it in a readable file and start fresh in the new software.
In what order should the data be imported?
Categories, then products, then customers, then stock. Each level depends on the previous one: a product needs its category and a quantity needs its product. Importing out of order means starting over.
How do you get your data out of uniCenta oPOS?
uniCenta stores everything in a MySQL or MariaDB database, so all of it is extractable by SQL query, which is the most favourable case in this comparison. Start with a full database backup, then export table by table to CSV files.
How long does a migration take?
A weekend for a business with around a thousand product lines, provided the files were prepared and cleaned the week before. The physical stock count is the longest step, and it is the one you must not skip.
Should you keep the old software?
Yes, at least a month running read-only in parallel, and after that as a consultable backup for the whole legal retention period of your accounting data. It costs nothing and it keeps you from being stuck.
Does BelloPOS help with migration?
Yes. The Go tier includes loading your first twenty products and ten categories, and the Pro tier loading a thousand products and a hundred categories. That does not replace your catalogue cleanup, but it removes the most tedious part.
Our recommendation
A successful migration is decided beforehand, not during. The backup, the export in the right order and the catalogue cleanup are eighty per cent of the work, and they are done calmly the week before.
Archive the history, do not migrate it. It is the time-saving advice in this guide, and it holds whichever software you are moving to, including if it is not ours.
And if you stay where you are, this guide is still useful. At least do the backup and the catalogue cleanup: the first protects you, the second will make you money from the following month.
Sources
Every claim made here about a competing product comes from its publisher’s own public page, consulted on 27 July 2026. No figure is taken from a review aggregator or another comparison.
- uniCenta oPOS: https://unicenta.com/
- uniCenta oPOS: https://sourceforge.net/projects/unicentaopos/
- Aronium: https://www.aronium.com/
- Aronium: https://www.aronium.com/en/plugins
- BelloPOS: https://bellocommerce.com/fr/bellopos/
- BelloPOS: https://bellocommerce.com/fr/bellopos/pricing/
The trademarks named here belong to their respective owners. They are mentioned for the purpose of comparison, and this article implies no partnership with or endorsement by them.
Test the migration with no commitment
Install BelloPOS Lite free and import twenty products from your current catalogue. An hour will tell you whether the full switch is simple for you.
Read next
Other practical guides on the same subject: