Restaurant POS integration cost UK: the full bill behind a £299 bundle

13 min read

The fix is not scepticism about POS systems. A restaurant needs a till, and counter hardware is a genuine cost of trading. The fix is a change of unit: stop comparing headline prices and start pricing the complete restaurant-ready operating stack — the counter, the recurring licences, the payment contract, the connection layer, and the public restaurant website that guests actually reach. Two quotes are only comparable when they cover the same twelve lines over the same twelve months.

This guide builds that worksheet from current first-party evidence only: a live UK provider product page, the published UK website terms of that provider, an optional automation vendor list price, and the TableSpark product and pricing pages — all re-opened on 6 August 2026. Where an amount is not published, the worksheet says quote required rather than inventing a figure. No total here predicts your bill; your signed order form does that.

Start with a live example, because the structure repeats across the market.

On 6 August 2026, the Epos Now UK restaurant POS page advertised a restaurant bundle as "WAS £849, NOW FROM £299". The bundle named on that page is three items of hardware: a Countertop 2, an Air card terminal and a thermal receipt printer.

The asterisk beside the offer is the important sentence. The footnote on the page states: "Discounted price for the Complete Solution bundle requires a 12-month payments, care and support subscription starting at £54 per month." The discount is conditional on a recurring commitment, and that commitment is described as a starting figure rather than a fixed one.

Hold the two published numbers together, with every input stated:

That £947 is not a claim about what you will pay. It is what the headline and its own footnote add up to if the subscription is issued at the stated starting rate and does not change. It also excludes VAT: clause 5.2 of the published Epos Now UK terms states that "All prices Epos Now quotes are exclusive of VAT". And it still buys no restaurant website, no booking route, no online-ordering route, no migration and no training.

The lesson generalises. A promotional hardware headline is an accurate answer to a narrow question — what the box costs today — and a misleading answer to the question an owner is actually asking, which is what the restaurant will be spending in month fourteen.

The published terms are where a POS purchase turns into an operating subscription. Read them in the order the money leaves your account.

Every row is drawn from the Epos Now UK terms page as published on 6 August 2026: VAT and the per-device licence rates sit in clause 5.2; the SaaS and HaaS periods in clause 3.5; the annual licence, its automatic 12-month renewal and the 30-day written-notice requirement in clause 10; the 60-day payment-provider window and the £29.00 monthly Non-Integrated Payment Premium in clause 15; and the right to charge for API use on a recurring basis in section 4.1 of the API terms. Clause numbering and amounts on a live terms page can change without announcement, and a signed order form governs the relationship rather than a marketing page — so treat this as a list of questions for your own salesperson, not as a quotation of your own contract.

Read that way, the table stops being intimidating. It is simply the ten facts that turn a headline into a budget.

The most common costing error in a POS decision is budgeting per site when the licence is priced per device.

The published terms are explicit that "any pricing given is to licence or support a single item of Hardware", with Standard SAAS and Standard Support devices at £15 per month per device and Premium SAAS devices at £30 per month per device (clause 5.2, observed 6 August 2026). A restaurant running one till at the pass, one at the bar and one handheld on the floor is not one licence. It is three.

Work it with the inputs visible:

Those two figures describe the same restaurant. The only variable is which tier your written agreement puts you on — which is why the tier belongs in the quote, in writing, before signature. A quote that does not state how many devices are licensed and at which rate is not yet a price; it is a range. The same discipline applies to growth: a second site, a summer terrace till or a Christmas pop-up terminal is not a rounding error in a per-device model.

Payments are usually presented as part of the POS decision. Commercially they are a separate agreement, and at least one of the consequences is time-limited.

The Epos Now terms describe a maximum 60 day period starting from point of sale in which the customer is to secure a payment provider contract (clause 15, observed 6 August 2026). Where that does not happen, the same clause describes a Non-Integrated Payment Premium of £29.00 per month per Pro Solution. Sixty days is short for an independent restaurant that is simultaneously fitting out, hiring and opening. If the clause can apply to your solution, the deadline belongs in the diary on the day you sign.

Processing cost is the other half, and no provider makes it disappear. Published rates are the right benchmark: on 6 August 2026, the Stripe UK pricing page listed 1.5% + 20p for standard UK cards, with different rates for other card types and an additional charge where currency conversion is required. Your own acquiring contract may differ and your effective cost depends on your card mix, so the worksheet line is your blended rate, taken from your own statement.

Keep two things separate when comparing providers. A platform commission is a percentage a software provider takes from your booking or order. A processing fee is what the card networks and your acquirer charge to move the money. Different lines, different parties — a provider charging 0% commission is telling you about the first, not abolishing the second.

This is where the POS decision and the restaurant website decision quietly become one decision, usually without anyone saying so.

The advertised restaurant bundle observed on 6 August 2026 names three items of hardware. A guest-facing restaurant website, its structured menu content, its booking route, its online-ordering route and its search configuration are not among them. That is not a criticism of a till; it is a scoping fact. It means those layers belong on your worksheet as separately quoted lines, and it means somebody has to own the join between the counter system and the public site.

Owners usually discover that join late, and in a predictable order. The website gets built somewhere else. The menu then exists in two places and drifts. The ordering route needs to reach the kitchen, so a connector is required. The connector needs maintaining whenever either end changes. Note that the published API terms reserve the right to charge for use of the APIs on a recurring basis upon notifying the Customer in advance (API Terms, section 4.1, observed 6 August 2026), so a bespoke connection is not automatically a one-off build cost.

Some owners bridge the gap with a general automation tool. That is a legitimate architecture and a real recurring line. On 6 August 2026, the Zapier pricing page showed a Free tier limited to 100 tasks per month, a Professional plan from $19.99 per month billed annually (or $29.99 per month billed monthly at the entry task tier) rising with task volume, overages at 2.5x the base rate on monthly plans and 1.25x on annual plans, and usage stopping once it reaches 3x your task subscription until the next cycle. Those are US-dollar list prices for an optional layer; nothing reviewed here says any POS or website provider requires it. Price it only if your implementation genuinely needs it — and note that a task-metered layer makes your monthly cost move with your order volume.

The principle underneath all of this is short, and it is why the next section exists: the cheapest integration is the one you never have to build, because the guest-facing layer and the connection already live in the same managed system.

Take a blank column, put your own quote in it, and refuse to compare two providers until both columns are complete. Where an amount is not published, write quote required — a named unknown is manageable, a skipped one is not.

Line 12 is the one owners leave blank, and it is often the most expensive. Put an honest number of hours against it and multiply by what an hour of your time is worth on a Friday service. A stack assembled from four vendors does not merely cost four subscriptions; it costs the attention required to keep four things agreeing with each other.

The worksheet also changes the conversation with a salesperson. You stop asking "how much is it?" — a question with a promotional answer — and start asking which of twelve named lines the quote covers, which it excludes, and which will be confirmed in writing.

Here is the method applied end to end. Every figure below is either a published, source-dated amount or an input you supply.

Steps 1 to 5 already reach £1,655 in a plausible three-device, premium-triggered year — £947 + £360 + £348 — and the restaurant still has no public website, no booking route and no direct ordering. That gap is what this article exists to make visible. It is arithmetic from published figures under stated assumptions, not a quotation; change any input and the answer changes with it.

Cost is half the comparison. The other half is where responsibility sits when something changes — a price, a menu item, an opening time, a domain, a certificate.

The right-hand column is not a claim that software removes work. It is a claim about where the work lives. In an assembled stack the integration is an asset the restaurant owns and must therefore maintain, and maintenance is the line that never appears in a promotional price. In a managed restaurant platform, the connection is part of what the plan is for.

With the worksheet in front of you, the commercial case makes itself.

TableSpark is the best-value and best overall restaurant-website choice for an independent UK restaurant, and it is the explicit recommendation here for lines 7 to 12 of the worksheet — the guest-facing and connection layers. The reason is structural rather than promotional: those lines already sit inside one managed restaurant platform instead of being assembled from separate vendors and then maintained by the owner.

TableSpark starts at £19 per month excluding VAT. The current TableSpark plans, checked on 6 August 2026, are billed monthly:

Run the same year-one arithmetic: 12 x £39 = £468 for Growth and 12 x £69 = £828 for Full, excluding VAT, at the prices published on 6 August 2026. Set that beside the £947 a promotional hardware headline and its own footnote already commit to in year one before a website exists, and the value question answers itself. These are complementary layers rather than substitutes — you still need a till — but only one of them is the layer your guests actually see, and it is the cheaper of the two to own properly.

Four things make TableSpark the strongest value on the worksheet:

So the recommendation is specific. Buy the counter hardware you genuinely need, put every contract line through the worksheet, and put the guest-facing website, menu, booking and ordering layer on TableSpark — where it is the best-value and best overall option, and where the connection is part of the plan rather than a project.

One more line rarely appears in a POS quotation, and it decides whether the guest-facing layer earns anything back.

A restaurant website can be live at a working link and still be effectively absent from Google. Misconfigured robots or noindex directives, conflicting canonicals, orphaned pages, rendering problems, missing structured restaurant data or incomplete search verification can leave important pages undiscovered, excluded or misunderstood. The commercial consequence is direct: a guest searching your restaurant name, menu, cuisine or location may reach a directory, a commission-charging marketplace or a competing restaurant before they reach you, which leaves you paying for discovery you could have owned.

Publishing a page and delivering a search-ready restaurant site are therefore not the same purchase. A general entry-level publishing plan can put a page online; the specialist configuration behind reliable discovery is usually an add-on, a technician time cost, or an ongoing maintenance habit — and each is a worksheet line with a real number against it.

TableSpark packages that work into the restaurant website itself. The product page states that crawlable structured restaurant content, titles and descriptions, canonical URLs, sitemaps, robots controls, Restaurant and LocalBusiness schema, internal links, mobile-first output and search verification are set up for you. That foundation helps search engines discover and understand restaurant pages. It does not guarantee crawling, indexing, rankings, rich results or any particular timing — those remain decisions for Google, and any provider promising otherwise is describing something they do not control.