The demo always goes well. The salesperson taps a few products, splits a bill, prints a neat receipt, and the numbers on the dashboard update in real time. That is the point of a demo. It is a controlled environment where nothing goes wrong. Choosing a pos system based on that thirty-minute performance is how retailers end up locked into tools that fail on the exact days that matter most.
I have helped several Indonesian retailers untangle themselves from a point-of-sale decision made in the meeting room. The pattern is almost always the same. The features that sold the system were never the problem. The problem was everything the demo quietly skipped: what happens when the internet drops, whether you can actually get your own data out, and which fees only show up on the fourth invoice.
So before you sign, run the system through the questions a demo will never volunteer. Choosing a pos system well is less about the shiny screen and more about the boring failure modes.
What Happens When the Internet Drops
In a lot of Indonesian retail locations, the connection is not a guarantee. A cafe in a ruko, a stall in a mall basement, a store during a PLN outage running on a small UPS. The question is simple: can you still ring up a sale when the network is gone?
Ask the vendor to unplug the router during your trial. Then check:
- Can you complete a cash transaction fully offline?
- Do offline sales sync automatically once the connection returns, without duplicating or losing anything?
- What is the offline limit, in transactions or in hours, before the system refuses to work?
Some cloud-only systems freeze the moment they lose the server. For a business doing two hundred transactions a day, a two-hour outage is not an inconvenience, it is real lost revenue and a queue of annoyed customers. A proper offline mode with reliable sync is worth more than any loyalty-points feature.
Can You Actually Export Your Own Data
Your sales history is one of the most valuable assets your business generates. You need to be able to leave with it. Test the export before you commit, not after.
A trustworthy system lets you pull:
- Full transaction-level data as CSV or Excel, not just a summary PDF.
- Product, stock, and customer lists in a standard format.
- Reports covering any date range you choose, not only fixed monthly windows.
If the only way to see your numbers is through their dashboard, you do not own your data, you are renting a view of it. That matters when you want to run your own analysis, hand figures to your accountant, or migrate to a different tool in two years. This is the same ownership logic I apply in Build vs Buy Software: A Decision Framework for Owners. Whatever you build on top of, you should be able to walk away with your data intact.
The Hardware Lock-In Trap
Point-of-sale hardware breaks. Receipt printers jam, cash drawers wear out, card readers die. The real question is what a replacement costs and whether you are free to buy it anywhere.
Watch for these traps:
- Proprietary hardware. If the receipt printer only works with their brand, a replacement can cost three to four times a standard thermal printer. A generic 58mm or 80mm printer runs around Rp 400.000 to Rp 900.000. A locked-in unit can be Rp 2.500.000 or more.
- Bundled leases. Some deals bundle hardware into a monthly fee that never ends, so you keep paying long after the device has paid for itself.
- Locked payment terminals. If the card reader is tied to one acquirer, you lose the leverage to negotiate transaction fees later.
Prefer systems that work with standard, off-the-shelf hardware. Freedom to source your own equipment locally is worth more than a slightly prettier all-in-one unit.
The QRIS and Reconciliation Reality
For Indonesian merchants, QRIS is now a default expectation, not a bonus. But accepting QRIS is only half the job. The half that causes headaches is settlement and reconciliation.
Ask hard questions here:
- Settlement timing. When does QRIS money actually land in your account? Same day, next day, or T+2? Cash flow depends on the answer.
- Reconciliation. Does the system match each QRIS payment against its transaction automatically, or does someone on your team reconcile by hand every night?
- Fees. The QRIS MDR for most merchants sits at a set rate. Confirm exactly what percentage is deducted and whether the POS provider adds anything on top.
A POS that records a QRIS sale but leaves you to reconcile settlements manually against a mutasi rekening is quietly handing your staff an hour of tedious work every day. Good reconciliation is invisible when it works and expensive when it does not.
The Fees That Appear in Month Four
The first invoice is always clean. The costs that hurt tend to surface later:
| Fee type | When it usually appears | Typical range |
|---|---|---|
| Per-transaction fee | After a free-tier volume cap | 0.5% to 1% per sale |
| Extra outlet or user | When you add a second location | Rp 150.000 to Rp 500.000/month |
| Premium report access | When you need real analytics | Added tier, often 2x base |
| Support beyond basic | First real emergency | Per-incident or higher plan |
Before signing, ask for the full price list in writing, including every add-on and every tier. Then model your cost at your expected volume twelve months out, not at today's numbers. A plan that looks cheap at fifty transactions a day can become uncomfortable at three hundred.
The Practical Takeaway
Choosing a pos system is a multi-year commitment dressed up as a software purchase. Do not let the demo make the decision for you. Instead, run your own trial with the failures built in.
- Pull the internet and confirm you can still sell.
- Export a full month of transaction data yourself before you buy.
- Price a replacement receipt printer and check the hardware is standard.
- Trace exactly how a QRIS payment settles and reconciles.
- Get every fee in writing and model your real twelve-month cost.
A point-of-sale system touches every sale you make, so a careful choice pays back quietly every single day. If you want a second opinion before committing to a retail tech decision like this, that is exactly the kind of thing I help with through a technical partnership. The best time to ask the hard questions is before the contract, not during the outage.