Short answer. To run a business in Algeria from France in 2027 you need neither more calls nor more cameras: you need four things recorded where they happen, every sale at the till, every stock delivery, every expense with its photo, and the cash count at closing, and you need them to reach a server in your name on their own, even after a power or network outage. From there, one screen on your phone every evening: five numbers and the alerts that matter for your type of business. On a business that is already trading, this takes six to ten weeks to set up, without closing a single day, and it only holds if the person running the business gains from it too.
This guide is for owners who live in France and whose business is in Algeria, and extends launching a tech startup in Algeria from France. It does not cover company law or powers of attorney: it covers what you see every evening, and what makes what you see reliable.
Why the photo of the notebook is no longer enough
The photo of the notebook shows you the total someone wrote, not the sales that happened. The camera shows you who was there, not what was sold or what is missing. The evening call gives you an impression and wears the relationship down: you check, the other person feels suspected, and nobody has a number both sides accept.
The problem is not trust, it is the lack of shared numbers. A cancelled sale, a discount given to a neighbour, a delivery paid in cash without a receipt, a till opened to make change: none of this is necessarily dishonest, and none of it shows on a photo. From a distance, the gap builds up for months before it becomes visible, and when it does, it cannot be explained.
Two changes make 2027 a good moment. Card payments at the counter are growing fast: in the first half of 2026, GIE Monétique counted about 8 million transactions on card terminals, up 81 %, and every electronic payment leaves a trace. And the 2027 finance bill provides a tax reduction for traders whose receipts are at least half electronic: if it is passed as drafted, recording every payment also becomes a tax argument. What remains is cash, which is exactly what the notebook does not keep.
«The problem is not trust, it is the lack of numbers both sides accept.»
From the counter to your phone: four layers
The system fits in four layers. Only the last one is in France.
On site, four actions are recorded: the sale at the till, on a tablet or a card terminal with a till app, with a receipt for every sale; the stock delivery, with the supplier's note photographed; the expense, with its amount and a photo of the receipt; and the cash count at closing. Sync works offline first: the till keeps running during an outage, keeps everything locally and sends it when the network returns, with the exact time it happened. A till that stops when the network drops sends the team back to the notebook at the first outage, for good.
The server is in your name: the account, the domain, the daily backup and the history, not in the manager's or the contractor's name, for the reasons we described in my app is stuck with a developer who vanished. The detail that changes everything is traceability: every line says who entered it and when, and a paid sale is never edited, it is cancelled, and the cancellation stays visible. Your phone, finally, shows an evening screen and a few alerts, in Arabic or French, with separate rights: you see everything, the manager sees the business, the accountant sees the exports.
What you look at every evening, by type of business
Five numbers are enough, but they are not the same for a restaurant and for a rental agency. Here is what we put on the evening screen, and the alert that deserves a notification.
In a restaurant, covers, average ticket and the cash gap say almost everything; the alert that matters is a ticket cancelled after payment, as we detailed for a restaurant till, from the floor to the kitchen. In a convenience store, it is margin by aisle and the gap between counted and expected stock, the subject of stock management in Algeria. In a shop, sales by seller and discounts above the ceiling. In car or apartment rental, occupancy, deposits, and any car out or booking without a recorded payment. In wholesale, receivables and customer credit above its limit. With several outlets, the same numbers side by side, and the one drifting from the others.
An alert that rings ten times a day is an alert nobody opens. We tune three to five, to your business, not thirty.
The evening routine, in five steps
The system does not replace closing: it makes it short and verifiable. Ten minutes at the business, two minutes in France.
Closing takes one tap, and the day's total is calculated, not copied. The cash count is entered by the person on duty, and the gap between what is in the drawer and what the till expects shows on its own. The day's expenses, paid in cash from the till, each have their amount and a photo of their receipt. Sending is automatic: nothing to photograph or send on WhatsApp. Your evening screen reaches your phone; in winter Algeria and France share the same time, in summer Paris is one hour ahead, and a 9 pm closing in Oran shows at 10 pm in Lyon.
The cash gap is calculated by the system, not announced by someone. That protects the person closing as much as the person watching, and it is the argument that gets the tool accepted: the numbers speak instead of the person. Once a week, a rolling count of ten items is enough; the full stock count stays yearly.
«The cash gap is calculated by the system, not announced by someone.»
Eight weeks to set it up, without closing
On a business that is already trading, the honest calendar looks like this.
Week 1, scoping and the choice of hardware: tablet, receipt printer, a UPS if outages are frequent, and the card terminal you already have. Weeks 1 to 2, the opening stock count: one full count, once, without which no stock gap means anything. Weeks 2 to 4, the till runs in parallel with the notebook: the team compares both every evening, sees that they match, then trusts it; this is the step that gets the tool accepted, and the one you do not skip. Weeks 3 to 5, stock deliveries and expenses. Weeks 4 to 6, the evening screen on your phone. Weeks 6 to 8, alerts are tuned and the notebook is stopped.
This is the low end of the six-to-ten-week range we give for a web app in how long it takes to build an app in 2027; each extra outlet then takes one to two weeks. And if an off-the-shelf till covers your business, we tell you so: custom work is justified when the trade departs from it, several outlets to bring together, rental, customer credit, deliveries, or one screen for several businesses.
What makes this kind of project fail
- Presenting it as surveillance. The person running the business pushes back, and rightly so. Present it as what protects them from suspicion: the numbers speak for them.
- No opening stock count. Without stock counted on day one, every gap becomes an argument. One full count, once, at the start.
- A till that stops when the network drops. The team goes back to the notebook at the first outage and never leaves it. Offline first, sync after.
- Editable sales. If a paid ticket can be corrected without a trace, the numbers are worthless. You cancel, you never edit.
- Thirty alerts. Nobody reads them after a week. Three to five, tuned to your business.
- Accounts in the manager's or the contractor's name. The day either leaves, the history leaves with them. Server, domain and data in your name, from week 1.
- Doing everything at once. Till, stock, expenses, deliveries and accounting in the same month: the team gives up. One layer after another, in calendar order.
Frequently asked questions
- Can I follow the till of a business in Algeria from France in real time?
- Yes. Every sale recorded at the till reaches the server and your phone within seconds when the network is up, and as soon as it returns after an outage. In practice, most owners look at a closing screen every evening rather than in real time.
- Aren't cameras enough?
- They show who was there, not what was sold or what is missing. A camera helps understand a gap the numbers have flagged; it does not flag it.
- What happens during a power or internet outage?
- The till keeps working on the tablet, with a UPS if outages are frequent. Sales are kept locally and sent when the network returns, in the order and at the time they happened.
- Will my manager accept it?
- If they gain from it, yes: a ten-minute closing instead of a notebook, fewer calls, and numbers that protect them from suspicion. Presenting it as a control tool is the surest way to see it worked around.
- How long does it take to set up?
- Six to ten weeks on a business already trading, without closing: opening count, till in parallel with the notebook, deliveries and expenses, evening screen, then alerts. Each extra outlet then takes one to two weeks.
- Do I need custom software?
- Not always. If an off-the-shelf till covers your business, we tell you so. Custom work is justified when the trade departs from it: several outlets to bring together, rental, customer credit, deliveries, or one screen for several businesses.