A takeaway counter is a different problem from a dining room. There are no tables, the orders arrive from three places at once, and the thing that goes wrong is not a split bill but a collection time nobody agreed on. A till sold for restaurants will take the money; whether it makes Friday at eight easier depends on a short list of things that have nothing to do with the drawer.
Orders arrive from three places and have to become one queue
Walk-ins at the counter, calls, and orders from your own site all have to end up in one list the kitchen works down, in the order they are actually due out. The failure that costs you is not a lost order, it is three separate lists with three separate clocks, which produces a bag sitting on the shelf going cold while a later order is being cooked. Ask to see a collection order from a website land in the same queue as a walk-in.
The collection time is the product
Everything a takeaway till does well or badly comes back to a promised time. It has to be set from what the kitchen can actually do at that moment rather than a fixed twenty minutes, it has to reach the customer, and it has to be visible on the ticket. A system that quotes the same wait at seven and at nine will be lying to your customers twice a night, and they will remember it as your fault.
What takeaway software stops doing
The category is sold as though it covers everything from the order to the driver. It usually does not: dispatch, route planning and driver tracking are a separate product, and where they are bundled they are often thin. If you deliver with your own drivers, ask the delivery questions separately and expect to hear a smaller answer than the marketing suggested. If you deliver only through a marketplace, none of it matters and you should not pay for it.
A collection promise, worked from the kitchen’s minutes
Eight o'clock on a Friday: nine orders in the queue, the pass plates one every three minutes, so the next order is thirty minutes away and the confirmation should say thirty, not the twenty the setting defaults to. Ten past: two orders collected, one large order landed, still thirty. Half past: the queue clears to four, the promise falls to fifteen. A till that works the promise from the queue tells the truth twice a night without a person guessing; one that quotes a fixed wait tells a lie that the customer remembers. The setting is the queue length the kitchen can hold and the minutes per order, and both are numbers the kitchen already knows.
The phone order, typed once
Half of a takeaway counter's evening still arrives by phone, and the phone order is the one that gets typed twice: once on the pad by the phone and once into the till when somebody has a hand free. Take it into the till while the customer is on the line, with the collection time read from the queue and repeated back, and the phone order joins the same list as the walk-in and the website order with the same clock. That is the whole of what a takeaway till has to do differently from a restaurant one, and it is the reason the phone should sit next to the screen rather than next to the fryer.
Questions people ask about takeaway epos
Is a takeaway EPOS different from a restaurant one?
The record is: collection slots instead of tables, promised times instead of courses. Plenty of products do both, and the ones that do it well let you turn the table side off entirely rather than leaving it on the screen forever.
Do I still need a till if I only take orders through marketplaces?
You need something that holds the record, and marketplace dashboards do not, at least not in a form you can compare across two of them. Rooms that sell only through apps still end up wanting one place where the day's orders are.
Can one system handle collection and delivery?
For collection, comfortably. For delivery with your own drivers, ask specifically about assigning and tracking, because that is the part most often described in a brochure and thin in the product.