How we compare UK POS costs
Every cost figure on this site comes from one model, applied the same way to every provider including our own. This page shows the assumptions, the arithmetic and the sources, so you can check the numbers or rerun them on your own trade.
Prices checked July 2026. All figures exclude VAT.
The venue we model
Comparisons only mean something against a fixed profile, because the answer changes with card volume. We use one independent full-service venue throughout:
1,600 transactions is simply £40,000 divided by £25. It matters because most card pricing has a fixed pence component per transaction, so a venue with a low average order pays more in fees on the same turnover.
What we count, and why
Software subscription
The published monthly price of the plan that a venue of this profile would actually need, not the cheapest tier on the page. Where a provider prices per till or per device, we count one.
Feature parity add-ons
If QR ordering, kitchen display or inventory are not in the base plan, we add what it costs to reach the same feature set, using the provider's own add-ons where they exist and third-party apps where they do not. This is the line most comparisons leave out, and it is usually the one that decides the answer.
Hardware, spread over twelve months
Terminals and printers are a one-off purchase, so quoting them as a monthly figure would flatter a subscription and quoting them as a lump sum would flatter hardware-free systems. We spread the purchase across twelve months. A venue that keeps hardware for three years pays less per month than shown; one that replaces a dropped terminal pays more.
Card processing fees
Published rate applied to £40,000, plus the fixed pence charge applied to 1,600 transactions. At this volume card fees are the largest line for every provider, which is why a comparison that stops at the software price tells you very little.
What we do not count
Staff training time, internet connectivity, receipt paper, and anything a venue would pay regardless of which system it chose. We also exclude promotional first-year discounts, because they expire and the comparison is meant to describe the ongoing bill.
The four models, worked
Same venue, same features, same volume. Every figure below is derived from the constants above rather than typed in, and a CI check fails the build if the published totals stop matching the arithmetic.
Reading it honestly. Servio comes out lowest here because QR ordering, kitchen display, inventory and analytics are in the £199 plan and there is no hardware to buy. Square's software is cheaper and its card rate is the lowest of the four; it lands higher because reaching the same feature set costs £99 a month and the hardware stack is real money. Zettle and SumUp start free or near free, then pay for it twice: £178 a month of third-party apps, and a higher effective card rate once orders route through them.
On the founding-venue pilot at £99/month, the same venue's all-in figure is ~£1,019. That is a promotional rate for the first six paid months, so it is shown separately rather than used as the headline comparison.
Sources
Every price is taken from the provider's own published pricing page, read in July 2026. Where a provider does not publish a price we say so rather than estimate: Toast, Lightspeed, Epos Now and Revel are quote-based in the UK, which is why they appear in our system comparison but not in this table.
Two lines in the table are estimates rather than published figures, and we would rather flag them than bury them. The hardware stack values are typical configurations rather than a single published bundle, and the 1.9% + 25p rate for app-routed orders is the rate those apps commonly quote rather than a headline reader rate. Both are marked as estimates wherever the figures appear.
What would change the answer
A lower average order value. The fixed pence per transaction hurts more. A dessert lounge at £12 average pays fees on more than three thousand transactions for the same turnover.
Much lower card volume. Below roughly £10,000 a month the subscription dominates and free tiers look genuinely better, which is why we say plainly that a simple counter may never need more than a free app.
Hardware you already own. If a venue already has a Square Register, the hardware line falls away and the gap narrows considerably.
Not needing the features. The model assumes a venue wants QR ordering and a kitchen display. Without them, the parity add-ons come out and the comparison changes shape entirely.
Run it on your own numbers with the calculator on the pricing page, or see the full line-by-line walkthrough in the UK POS cost guide.
Our bias, stated
Servio is our product and it is in the table. We publish the model instead of only the conclusion so the working can be checked rather than taken on trust, and the arithmetic is enforced in CI so the numbers on the page cannot quietly stop matching the model. Where a competitor wins a line, the table shows it: Square's card rate is lower than ours and its software is cheaper.
If you find an error, tell us at hello@servio.uk and we will correct it and note the change.
Common questions
Why £40,000 a month?
It is a realistic figure for a busy independent full-service venue, and it is high enough that card fees dominate the bill, which is the point the comparison exists to make. At much lower volumes the answer genuinely changes, and we say so.
Why spread hardware over twelve months rather than three years?
Twelve months is the conservative choice against our own interest, because a longer amortisation would make hardware-based systems look cheaper per month and Servio look relatively better. We would rather understate our advantage than overstate it.
How often is this updated?
Prices are re-checked quarterly and whenever a provider announces a change. The date at the top of this page is when the figures were last verified, and it is the same date used across every comparison on the site.
Can I reuse these figures?
Yes. Cite the page and the date the prices were checked, and please link back so a reader can see the assumptions rather than the number alone. A figure without its model is not much use to anyone.