What a Six-Person Café Pays for Software Every Year — Line by Line

Nobody decides to spend two thousand dollars a year on café software. They decide five separate times that fourteen dollars a month is fine.

This is that bill, itemized. Every figure is quoted from the vendor's own pricing page with the date it was read, and at the end there is an honest account of which lines are worth paying and which are not.

The café: six staff, one site, table service, takes cards.

The bill

(Each row's verbatim quote, source link and read date is held in the Price Observatory.)

And note the Deputy line. Six staff on Lite is exactly $30 — the minimum and the arithmetic happen to meet. At five staff you would still pay $30, because the floor does not care.

Where the money actually is, and it is not on that table

Card processing dwarfs every line above. A café turning over $300,000 a year on cards, at a processing rate 0.3 percentage points higher than it needs to be, is paying $900 a year — more than half the entire software bill, invisibly.

Before optimizing a single subscription, find your effective rate: total processing fees divided by total card turnover, from your statement. Not from a rate card — those exclude interchange and scheme fees.

Line by line: keep, cut, or own

Till — keep paying, but check the tier

Square has a $0 plan that takes cards, runs a till and does basic inventory. The $49 buys advanced inventory, staff permissions and deeper reporting.

But the subscription is not only buying features — it is buying a processing rate, and that changes the arithmetic. From Square's own fee page (read 1 August 2026; the same rates were on the page on 2 September 2026 — a 5 September re-read returned the plan headings without the rate table, so these four figures keep their earlier read date rather than borrow a newer one): in-person is "2.6% + 15¢" on the free plan and "2.5% + 15¢" on Plus. Online is "3.3% + 30¢" free versus "2.9% + 30¢" on Plus.

So downgrading does not save you the whole $49. At $300,000 a year on cards — $25,000 a month — that extra 0.1 percentage point in person costs about $25 a month. The honest net saving is therefore around $24 a month, or roughly $288 a year, not $588.

And the direction flips with volume. Above roughly $49,000 a month of in-person card turnover, the 0.1-point discount alone is worth more than the $49 subscription — at which point Plus pays for itself and the features are free. If a meaningful share of your takings is online, the 0.4-point gap moves that break-even down to about $12,250 a month.

And a till is the one thing in a café that must never be somebody's side project — self-hosting your point of sale is a bad idea and we will not recommend it.

Verdict: stay on Square, and work out your own break-even before changing tier. Low card volume: downgrade, and expect roughly $288 a year rather than $588. High card volume: stay on Plus — the rate discount is doing more work than the subscription costs.

Rota — own it

Deputy at its minimum is $360 a year to record who is working. A rota has no payment surface, no compliance exposure, and no consequence worse than inconvenience if the server hiccups for an hour.

Verdict: the clearest own-it line on the table. Rostera is $69 once. Saving after year one: ~$360/year.

Try the Rostera demo ↗Live, on sample data, no sign-up.

Food safety — it depends on your inspectors, and the unit is the trap

FoodDocs is $79 per site on annual billing ($99 billed monthly, and the page states prices are exclusive of VAT) (read 5 July 2026, re-read 5 September 2026 — unchanged). SafetyCulture is $24 per seat on annual billing (free plan available; read 5 July 2026, re-read 3 September 2026 — unchanged, though the pricing address now redirects to mitti.com, where the product is called Mitti (by SafetyCulture), and the free plan is stated as being for teams of up to ten people).

For a one-site café with two people doing checks, SafetyCulture is $48 — cheaper than FoodDocs, and there is a free tier to test. For a one-site café with six people all logging fridge temperatures, FoodDocs' per-site model wins outright.

This is the single most miscompared line in café software, because the two headline numbers are not the same kind of number — the crossover, worked out.

Verdict: work out your inspector count first. A self-hosted logbook — Safora — is $79 once — but only if your readings are taken by hand. If you run wireless probes, buy a sensor platform and none of this applies.

Leave — own it

$9/month for six people is $108 a year to record holidays. And it grows: at fifteen staff it is $270.

Verdict: own it. Leavora is $99 once for unlimited staff.

Expenses — delete this line

A six-person café does not need expense-claim software. If your process is the owner keeping receipts in an envelope and reconciling monthly, adding $96 a year of software does not improve it.

Verdict: cancel. Saving: $96/year, immediately, for nothing.

The rebuilt bill

Expenses need not be deleted to reach zero. Zoho Expense publishes a free plan for up to three users (read 5 September 2026) — receipt capture, expense reports, mileage and multi-currency, without the corporate-card feeds, multi-level approval and policy engine the $4 Standard tier adds. A six-person cafe with two people filing expenses fits inside it. Deleting the line is the honest answer when nobody files expenses at all; the free plan is the honest answer when somebody does.

That is the maximal version and it will not be right for everyone. A more realistic outcome — keep Square Plus because you genuinely use the reporting and the lower processing rate, keep SafetyCulture's paid tier because you want the audit trail hosted — still lands around $73/month instead of $175. $1,224 a year saved, most of it by canceling one line and owning three others.

What this exercise actually shows

Only one of the five subscriptions was clearly worth its price — the till, and even then possibly at $0.

The other four had the same shape: a small monthly number, attached to a job that does not change, growing with headcount, and never reviewed since the day it was set up. That is not a story about bad vendors. It is a story about a decision nobody revisits.

The exercise transfers. Take your own five lines, write the monthly figure next to each, and ask three questions: Do we use the paid features? Does the price grow when we hire? What happens if this is down for a day? A "no, yes, nothing" answer is an own-it line.

The honest counterweight

Owning three tools means running three tools. Backups, PHP updates, and a server that is your problem. For a rota, a leave register and a hand-written temperature logbook, that responsibility is genuinely small — and it is small precisely because those tools have no payments, no public logins and no compliance surface.

If nobody in the business will do a monthly backup check, do not own anything. Downgrade the tiers, cancel the expense line, and take the $1,224 anyway. Most of the saving on this page does not require buying anything from us.

Next steps

• The SaaS Price Observatory — every vendor above, quoted from its own page with a verification date
• Self-hosted software for restaurants and cafés
• The Restaurant Ops Suite — the food-safety diary, the menu into the till and the review ask, bought once; the stack explained
• The Rostera demo and the Leavora demo — open without a sign-up
• Your backup is not a backup until you have restored it — read before owning anything

Prices are quoted from each vendor's own pricing page with the date read. Card processing rates are negotiated per merchant and are not quoted here — yours is on your statement.