POS system cost: the line that isn't the software
Ask what a POS system costs and you get a monthly figure. That figure is real, and it is usually the smallest line on the invoice.
The cost of a POS system has four components, and they scale in completely different ways. One is fixed, one grows with your headcount, one grows with your store count, and one grows with your turnover. Most people compare vendors on the fixed one — which is the equivalent of choosing a car by the price of the floor mats.
This page is how to work out your actual number. We do not quote vendor prices, because they change and because we do not rank POS vendors. What follows is the structure, so you can put your own figures in and get an answer that is true for your shop.
The four components of POS system cost
1. Software subscription — fixed, and the one everyone compares
A monthly or annual fee, sometimes per register, sometimes per location. This is the number in the pricing table and the number in every comparison article.
It is genuinely worth comparing. It is also, for most retailers doing meaningful card volume, not where the money goes.
2. Payment processing — a percentage of everything you sell
Here is the part that changes the arithmetic entirely.
Card processing is charged as a percentage of turnover plus a small fixed amount per transaction. Because it is a percentage, it does not care about your software choice, your register count or your opinion — it scales directly with success.
The consequence is arithmetic that most POS buyers never do:
Annual processing cost = annual card turnover × effective rate
Annual software cost = monthly fee × 12
Put your own turnover in. For most retailers past a modest volume, the first line is several times the second — often an order of magnitude. A difference of a few tenths of a percentage point in your processing rate can outweigh the entire software subscription, in either direction.
This is why "free POS software" is usually free. Many of the systems advertised at no cost are payment processors that give you the software so that they get the transaction stream. That can be a genuinely good deal — but you are choosing a payment contract, not a software package, and it should be evaluated as one.
Two things to establish before you sign anything:
- What is the effective rate, all-in, including the fixed per-transaction component and any monthly minimum — not the headline percentage.
- Are you locked to their processor? If the software only works with one payment company, you cannot renegotiate the biggest line on this page without replacing the whole system. That single question is worth more than any feature comparison.
3. Hardware — up front, then again in four years
Terminal, card reader, cash drawer, receipt printer, barcode scanner, and a stand. Sometimes leased into the monthly fee, which makes the subscription look higher and the up-front cost look like zero.
The line people forget is replacement. Retail hardware lives in a hostile environment and gets replaced on a cycle. Price it as an amortised annual cost rather than a one-off, and check whether the hardware is locked to the vendor — proprietary hardware means the switching cost of leaving includes buying all of it again.
4. Setup, migration and the staff day
Getting your product catalogue in, mapping tax rates, configuring receipts, and training staff. Usually quoted as a one-off if it is quoted at all, and usually underestimated, because the real cost is trading hours rather than a fee.
How POS system pricing is structured
Before the arithmetic, know which pricing model you are being quoted, because the same headline number means very different things across them. There are four, and vendors rarely name which one they use.
Per register. A fee for each till. Simple and predictable, and it punishes exactly the moment you succeed — a second checkout at Christmas doubles the software line. Ask whether a register can be deactivated seasonally or whether you pay year-round for a till you use in December.
Per location. A fee per shop, unlimited registers inside it. Better for busy single sites, worse for a chain of small ones. If you run three tiny shops this can cost more than one large one, which is the opposite of how your revenue works.
Flat rate. One price, everything included. Rare, and usually a higher headline that turns out cheaper for anyone with more than one till.
Revenue share, or "free". No software fee; the vendor takes a cut of processing. Covered in detail below — the price of a POS system on this model is not zero, it is a variable you cannot see on an invoice.
Two modifiers cut across all four. Per-user charges appear when every staff member needs their own login — which they should, because shared logins destroy the audit trail that makes a POS worth having. And module pricing puts inventory, loyalty, purchasing or multi-location behind separate line items, so the entry price of a POS system is rarely the price of the system you end up needing.
When you ask for the price of a POS system, ask for it as: base, per register, per user, per module, per location, plus processing. A quote that resists that breakdown is telling you where the money is.
Work out your three-year number
Compare on three years, not one. One-year comparisons flatter whichever option front-loads its cost.
3-year total = (monthly software × 36)
+ (annual card turnover × effective rate × 3)
+ hardware, including one replacement cycle
+ setup and migration
+ the cost of switching away, if you ever have to
Run it twice — once at today's turnover, once at your three-year target. The second run is the one that matters, because the processing line grows with the business and the software line usually does not. Systems that look identical at today's volume can differ substantially at twice the volume.
A worked example — put your own numbers in
The shape matters more than any figure, so here is the arithmetic with placeholders. Fill in three things you already know: your annual card turnover, the effective processing rate you have been quoted, and the monthly software fee.
Shop A — £300,000 annual card turnover
processing at 1.6% = £4,800 / year
software at £70/mo = £840 / year
→ processing is 5.7× the software
Shop B — £60,000 annual card turnover
processing at 1.6% = £960 / year
software at £70/mo = £840 / year
→ roughly equal
Same software, same rate, opposite conclusions. Shop B should optimise for software features and barely think about the rate. Shop A should treat the processing rate as the primary negotiation and the software as a rounding error — and should be extremely wary of anything that locks the two together.
Find your own crossover: divide your annual software cost by your effective rate. Above that turnover, processing dominates and you are shopping for a payments deal. Below it, you are shopping for software.
What you should actually be getting for it
Price comparison is only meaningful against a fixed idea of what the system must do. We maintain a hand-counted inventory of 60 real retail POS products, and the frequency distribution is unusually clear-cut:
| Capability | Found in |
|---|---|
| Cashier sell / checkout — search, scan, quick keys, editable basket | 60 of 60 (100%) |
| Actual stock visibility — real stock rows with movement history | 58 of 60 (97%) |
| Source-backed reporting — widgets that drill to exact source rows | 51 of 60 (85%) |
| Customer context — attach a customer, purchase history, loyalty | 44 of 60 (73%) |
| Tender handling — cash and change core; split and card as branches | 36 of 60 (60%) |
| Multi-location and transfers | 34 of 60 (57%) |
| Returns, refunds and exchanges — reversing every ledger the sale wrote | 29 of 60 (48%) |
| Register / shift / cash close — float, cash in/out, counted close, variance | 18 of 60 (30%) |
| Staff permissions and approval at the point of work | 16 of 60 (27%) |
Two things stand out.
Checkout is universal — every single product has it — so it is not a differentiator and should never be a line item you are sold on. What separates systems is the bottom half of that list.
The register and cash-close lifecycle appears in under a third. Open float, cash in and out during the day, a counted close and a variance figure. It is underreported in marketing and it is the thing that tells you whether the till was short on Tuesday. If you take cash at all, ask for a demonstration of the full trading day rather than the sale.
Likewise, returns appear in fewer than half — and a refund is the hardest thing in a POS to get right, because it must reverse every ledger the original sale wrote: stock, cash, tax, loyalty. Ask to see an exchange with a price difference, not a straight refund.
The number that makes building rational
For most retailers, buying is correct. The software is cheap relative to the processing line, the products are mature, and somebody else carries the maintenance.
Building becomes rational at a specific and identifiable point: when your processing arrangement is locked to your software, and the lock is costing you more than a build would.
That is the calculation nobody puts in a comparison table. If your POS forces a processor, and an independent processor would give you a materially better effective rate, then the annual difference on your card turnover is the true price of that software — and it recurs every year, growing as you grow. Against that number, a one-off build whose only ongoing cost is maintenance can pay back quickly.
The other trigger is fit. If your counter workflow is genuinely unusual — trade counters with account customers, service work that becomes a sale, deposits against future collection, weighted goods, split ownership across trading entities — you will pay for that mismatch every day in staff time, and no subscription line captures it.
If you do build, the capability order above is the specification. Checkout first, because it is the product. Then stock that reconciles, then reporting that drills to source rows rather than displaying totals nobody can trace. The register lifecycle and the returns workbench are where self-built systems usually stop — and they are exactly what makes the difference between a demo and a till you can trust on a Saturday.
If free is where you are starting, free POS software covers what those systems actually are and what they cost you elsewhere. And if you are building, the Retail POS planner is the specification we hand our own coding agents — the cashier-first counter, durable receipts, reversible money, and the controls that hold.