Taking payment in a Singapore supermarket, minimart or convenience store

Why a small basket is a different pricing problem

A fee with a fixed per-transaction component and a percentage behaves very differently at different ticket sizes. At $180 the fixed part is a rounding error. At $2.30 it can be most of what you pay, and no amount of negotiating the percentage changes that.

This is the trade where how fees are built stops being background and becomes the decision. Two providers quoting numbers that look similar can be very different for you specifically, purely because of where the fixed component sits.

The instruction is the same either way: compare effective rates. Total fees divided by total gross takings for one month, with every non-percentage line item folded in. A headline number is not comparable across two providers unless you know which shape each is quoting — and at this basket size, the shape is the whole thing.

Uniweb Pay quotes per merchant and does not publish rates. For this trade that means arriving with your real average basket and your real monthly volume, because those two numbers are what the quote should be built against.

Seconds per customer, across every lane

A minimart at 7pm and a supermarket on a Sunday are throughput problems. The cost of a slow method is not the fee, it is the queue — and unlike a restaurant, there is no seating to absorb it and no reason for the customer to stay.

So the terminal question is the same one F&B asks, for the same reason: every method the customer might reach for should be on the one device, because switching boxes mid-transaction costs seconds that multiply by every customer behind them. Contactless is the fastest path and it is what most local customers reach for first.

The multi-lane version adds a second requirement: the devices have to behave identically, because staff move between lanes and a lane that works differently is a lane that generates mistakes. Uniweb Pay supplies a Sunmi P3, which takes contactless, chip and insert, magnetic stripe and QR wallets on one unit, prints receipts, and runs on 4G or Wi-Fi so a flaky shop router does not stop the tills.

The rails that are switched off

The most common unnecessary cost in this trade is a rail that is enabled on paper and not in practice — nobody at the counter offers it, so customers use a more expensive one by default, and the merchant pays for a capability they are not using.

The check takes an afternoon: look at one month split by method, and compare it against what you believe is enabled. Merchants who do this usually find one of two things — a method they are paying to have available that nobody has ever offered a customer, or a rail switched off that would have been cheaper than the one being used in its place.

Dynamic PayNow is worth naming specifically. A static PayNow sticker at a till has no amount in it, so the customer types the figure and your cashier verifies by looking at the customer’s phone — during a queue, on a screen someone else is holding. Dynamic PayNow generates the code per sale with the exact amount and notifies your own system. The fraud guide explains why that distinction matters more than it sounds.

Where the shop is — near a tourist district, near a dormitory, in the heartlands — decides whether the visitor wallets are earning their place. That is a per-site question, and it is answered by the method split for that site rather than by an assumption.

Several lanes, several shops

Once there is more than one till, the useful view is per device and per outlet rather than a blended total. Which lane is slow, which site is actually taking visitor volume, and where refunds are concentrating are all questions that a monthly total cannot answer.

With Uniweb Pay the merchant portal shows live sales and a GMV trend, the split by payment method, and every outlet in one view, with transactions filterable by outlet or by method and exportable to Excel. Cards, wallets and Dynamic PayNow settle into one payout in Singapore dollars in T+1 batches. Several outlets covers the operational side.

What is not published: rates, any approval-time guarantee, chargeback handling specifics, and which third-party POS systems acceptance can sit behind — the last one being a fair first question for any shop already running a till system.