Nobody selling you a till will say this. Most lists of the best EPOS are written for a shop that does not look like yours. They assume three things. That you start from nothing. That you will pay a monthly fee for ever. That one product suits a corner shop and a five-site chain alike. None of that is true. There are really five kinds of independent shop owner, and the right answer for each one is different. Get the type wrong and you will overpay for years.
Connect with us about retail software · · Replies within 24 hours
The ACS Local Shop Report 2024 counts 50,387 convenience stores in the UK. Independent retailers run 71% of them, and not the large chains. This is not a corner of retail dominated by head-office IT departments. It is tens of thousands of owners making their own technology decisions, often alone, often at the end of a long shift.
And the tools in use are all over the place. The Retail Data Partnership estimates that around half of all convenience stores use no EPOS at all: a basic till, paper, or a spreadsheet. The ACS Local Shop Report 2024 found that only 17% have a self-service till. The Grocer reports that more than 16,000 UK shops trade under a symbol group name. Spar, Nisa, Premier, Costcutter, Londis, Best-one and Mace. Each one comes with its own EPOS arrangement. So when someone asks "what is the best EPOS", the answer starts with another question: which of these shops are you?
Before you compare a single product, work out which group you fall into. The pitch that suits one group is actively wrong for another.
Keep your type in mind as you read on. The comparison table below is clear about price, but price means nothing until you know which trap you are personally exposed to.
The loudest pitch in shop EPOS right now is "stop renting your till, own it instead". It is a good line, and for the cloud-subscription owner it is genuinely true. But it is sold as if it applies to everyone, and it does not.
Consider the own-it-outright owner. ICRTouch says its TouchPoint has been installed more than 150,000 times in the UK. The licence lasts for ever, with no monthly fee. One till usually costs £1,000 to £2,500, paid once. That owner already owns their system. Telling them to "stop paying monthly" is telling them to fix a bill they do not have.
Now consider the symbol-group owner. Their EPOS is free or subsidised, but it is tied to buying stock from the group. The Grocer and ShopMate report the Nisa evolution EPOS as a £3,895 saving. That is real money. It is not a gift. It is a discount in exchange for your stock loyalty. The "free" system is paid for at the wholesale invoice, every week, for as long as you stay.
Here is where the popular options actually sit. Prices are as published by each provider at the time of writing. This table does not name a winner. It shows that how you pay matters as much as the price you see.
| System | Headline price | Ongoing cost | Model and the catch |
|---|---|---|---|
| Epos Now | Advertised at £199 (reduced from £849), per Epos Now. | Support and payments subscription from £54 per month, plus card processing. | Cloud subscription on a 12 to 36 month contract. The cheap hardware is the hook; the monthly fee and the contract length are the real commitment. It holds 3.8/5 on Capterra from 726 reviews (checked July 2026), the lowest score of the tills here. |
| Square | No monthly software fee on the basic tier. | Around 1.75% on every card sale, per Square. | Pay-as-you-go. Cheap to start, but the percentage never stops. On £15,000 a month in card sales that is roughly £262.50 a month, about £3,150 a year (ESRE framework calculation from Square pricing). |
| Zettle (PayPal) | No monthly software fee on the basic tier. | Around 1.75% on every card sale, per Zettle. | Same model as Square. Illustratively about £262.50 a month, about £3,150 a year on £15,000 of card sales (ESRE framework calculation from Zettle pricing). |
| SumUp | Low entry on the 0.99% plan. | 0.99% plus £19 a month, per SumUp. | A blended model. Illustratively about £167.50 a month, about £2,010 a year on £15,000 of card sales (ESRE framework calculation from SumUp pricing). Lower rate, but a fixed monthly fee returns. |
| ICRTouch | Typically £1,000 to £2,500 per till, paid once, per ICRTouch. | None. Perpetual licence, no monthly software fee. | Own-it-outright. Over 150,000 UK installs. Higher up-front, nothing per month. Bought through dealers, so support and changes go through them. |
| Symbol group | Free or subsidised (Nisa quotes a £3,895 saving on evolution EPOS, per The Grocer and ShopMate). | Paid through tied stock buying from the group. | The "free" system. No software bill, but your wholesale spend is locked to the fascia. Leave the group and the deal goes with it. |
| ShopMate (TRDP) | Around £999, per The Retail Data Partnership. | Varies by support and services. | Convenience-focused EPOS sold by The Retail Data Partnership. A middle path between cheap cloud hardware and a full dealer till. |
Read the right-hand column twice. Square and Zettle look free until you notice the percentage runs for as long as you trade. Epos Now looks cheap until you add 12 to 36 months of fees. ICRTouch and the symbol group both charge no software fee, for opposite reasons. With ICRTouch you bought it. With the symbol group you pay at the wholesale counter instead.
Whichever till you pick, you still pay to move money. This is where a lot of the real cost hides, and where the differences are large. The figures below are an example. We modelled a shop taking about £15,000 a month in card sales, from the published prices of each provider.
| Processor | Rate | Illustrative monthly | Illustrative yearly |
|---|---|---|---|
| Square | 1.75% | About £262.50 | About £3,150 |
| Zettle | 1.75% | About £262.50 | About £3,150 |
| SumUp (0.99% plan) | 0.99% plus £19 a month | About £167.50 | About £2,010 |
| Stripe Terminal | 1.4% plus 20p | About £240 | About £2,880 |
| Revolut | 0.8% plus 2p | About £130 | About £1,560 |
The spread is the story. A move from about 1.75% to 0.8% saves that shop more than £1,500 a year. That is the card fees alone, before anything else in the shop changes. The Revolut figures include a terminal at about £169 plus VAT, paid once. They also include a Revolut Business account at £10 a month. The yearly saving covers both many times over.
Why does this matter in an EPOS article? Because most cloud tills steer you towards their own payments rail, and that rail is often the more expensive one. When the till and the card processor are bundled, you lose the freedom to put the cheaper processor behind the same screen.
It is tempting to treat the till as a minor purchase. It is not, because of what surrounds it. The ACS Crime Report 2024/25 puts retail crime at £316m in one year. There were 6.2m thefts, up from 5.6m. That is more than 600 every hour. Retailers spent more than £265m on crime prevention, and the ACS describes the net effect as a 10p "crime tax" added to every transaction. A till that follows your stock, marks a loss and keeps an audit trail helps you fight that. It is not a cost on top of it.
Labour is the other pressure. The National Living Wage rose to £12.71 an hour on 1 April 2026. That is 4.1% above £12.21. The National Minimum Wage (Amendment) Regulations 2026 set it. For context, the ACS notes it was £11.44 in April 2024, a 9.8% rise that year. The University of Stirling and the SGF counted everything an employer pays. They put the real cost at about £15.39 an hour in 2024. That figure came before the 2026 rise, so it is now higher. The same research found that 75% of convenience store owners and managers work 65 hours a week or more. The ACS found that 53% of retailers cut what they spent, and 47% took less profit, to pay the higher wages. When labour is that expensive and that stretched, a system that saves the owner an hour a day is not a luxury.
Everything above rests on one number. A system of your own for £2,000 to £3,000. Ten years ago that was impossible, and it is worth saying why it is possible now.
A till, live stock, supplier ordering, VAT by rate and an audit trail are the same in every shop. Until now every builder wrote all of it again for each customer. A build of your own therefore cost tens of thousands, and only a chain could pay.
We build each shop its own system on one shared foundation. That foundation already holds a product, a sale, a stock level, a supplier and an audit trail. Every shop we build for uses it. You pay only for what is your own, and you own the code, the data and the machine.
We explain how a whole sector shares one foundation in bespoke software for a whole sector. We set out the full cost, including your own time, in what does it cost to own and run your own systems.
Every option above is somebody else's product. You rent it, you buy a fixed version of it, or you get it free with strings. A bespoke system is the fourth model: one your shop owns outright, built around how you actually trade. This is what ESRE Media builds, and it is worth being precise about where it beats the alternatives and where it does not.
It does not win on the lowest possible up-front price for a single basic till. A £199 hardware deal or a free symbol-group box is cheaper on day one. A shop system of your own starts at about £2,000 to £3,000 for a core build, paid once. You own it outright, with no monthly software fee. More sites, a staff rota or a website in the same system cost more. So it is not the cheapest first cheque.
It wins on three things the generic platforms cannot match.
There is no subscription, no per-till licence and no contract. If the relationship with ESRE ended tomorrow, the system keeps running. Your data lives on secure UK servers, fully exportable, and you can roll the whole system back to any second through a write-ahead log. Compare that with the Square percentage, which never stops. Or the Epos Now contract of 12 to 36 months. Or a symbol group deal that ends the day you change wholesaler. Ownership is the whole point, and it is the one thing none of the rented options can give you.
Generic EPOS makes your shop fit the software. A bespoke build does the opposite. One system replaces five. The till, the stock control, the supplier ordering, the reports and the website. You get one source of truth, instead of five apps that do not talk. The working ESRE shop demonstration already does all of it. The till scans an item, or takes a typed code. It takes a card through a Revolut Terminal, or takes cash. It prints a receipt, and lowers the stock count by itself. The stock is live, with low-stock warnings and expiry dates. It proposes a purchase order from the levels you set. It gives reports most owners never see. Margin for each line, top sellers, dead stock, the busiest hours, waste and loss, and VAT by rate. You keep the screen, the USB scanner, the receipt printer and the cash drawer you have. Only the card reader changes, to a Revolut Terminal at about £169 plus VAT.
This is where the growing owner should pay attention. A system of your own can hold a staff rota, with the hours worked and the pay owed. It can show every shop on one head office screen. Most ready-made tills do neither well. The numbers turn as you grow. For a second shop, a multi-site system of your own costs about £2,000 to £3,000. A second premium dealer till costs £5,000 to £7,000. The product that looked expensive against a £199 box looks cheap against two dealer tills.
Four serious things come as standard on the first day. A full audit trail. A return to any second, through a write-ahead log. Automatic backups. Data held on UK servers. It is built in weeks rather than months, it is noticeably fast, and it is built to last and grow with the business. The catalogue loads from a CSV file. It looks up a barcode in Open Food Facts, and a telephone camera can read that barcode. Nobody types in 1,000 products. Every telephone becomes a scanner through a web app, with nothing to download from a store. End-of-day cash reconciliation is built in, and the till is offline-capable, so it keeps working if the internet drops and syncs later.
One thing changed in December 2025. AI tools became good enough for your own team to change the system in plain English. They work in safe environments, with version control and automatic backups. So the shop is not frozen on the day it launches. It changes as you change. We also give a guarantee most software companies will not. If something we built is faulty, we fix it at no cost, with no time limit. For early clients, hosting is free and permanent for a limited time. You can read the full reasoning on why bespoke, and see the kind of systems we build on the projects page.
It depends which of the five owners you are. If you are on no system, almost anything beats flying blind, and the decision is about how far you want to go. If you rent a cloud till, the case for owning is real for you. A build of your own is the strongest form of it. If you already own an ICRTouch outright, you owe nobody anything, and the case for switching is about capability, not cost. If you are tied to a symbol group, understand what the "free" till really costs at the wholesale invoice. And if you are growing, the bespoke multi-site maths is hard to argue with.
The thread running through all five is ownership. A rented till takes a share of every sale, or a fee every month. The symbol group gift takes its share at a till you cannot see. A system your shop owns outright stops the meter. It fits the way you trade, and it grows with you. That is the difference a bespoke build makes.
If you want to see what the bespoke route actually looks like, ESRE's live retail demo opens without a login. Sign in as staff and ring a sale with live stock update, or as owner and see the shrinkage report. It is a working system, not a screenshot.
Connect with us about retail software · · Replies within 24 hours