What a Restaurant Actually Pays for Software in 2026

Ask a restaurant owner what their software costs and they will name the subscriptions, because those arrive as invoices with a date on them. Then look at a month of statements. The subscriptions are the smaller half. The larger half is charged as a percentage of every transaction, never appears as a bill, and grows on exactly the nights the restaurant does well.

Most industries have one software cost: a set of subscriptions. Restaurants have two, and they behave completely differently.

The subscription bill is the familiar kind. Email marketing, reservations, scheduling, review tools, a website. Fixed monthly amounts, visible on a card statement, easy to audit and easy to cancel.

The percentage bill is card processing and marketplace commission. It never arrives as an invoice, because it is deducted before the money reaches you. It is not a line you pay, it is a slice that never turns up, which is why it is so much easier to stop noticing.

For a delivery-heavy restaurant, marketplace commission is usually the largest technology cost in the business by a wide margin, and it is not close. Card processing sits behind it. Neither is optional in practice, and this post is not going to pretend otherwise.

What is worth doing is the arithmetic, once, because most operators have never put the two bills side by side. Take last month's marketplace commission and processing fees, add them, and compare that figure to the total of every software subscription you pay. For most restaurants with meaningful delivery volume the comparison is uncomfortable, and it reframes what is worth optimising.

The strategic consequence is straightforward. If the largest technology cost is a percentage of orders placed through somebody else's app, then the highest-value software work in the business is anything that moves a repeat guest from that app to your own channel. Which is what the rest of this is about.

The marketing and admin layer is the part that behaves like ordinary software, and it is the part we can price honestly. Across the 68 go-to-market tools we track, the median per-seat entry price is $19 per seat per month and the median flat entry price is $97 a month.

A single-site restaurant running email marketing, a booking tool, a review request tool and social scheduling is typically in the low hundreds a month at entry tiers. Point-of-sale, reservations platforms and delivery integrations are priced per terminal, per cover or per order and are not in the set we track, so anything you read that adds them into one confident monthly figure without saying which products it assumed is guessing.

The duplications are predictable, because every layer of a restaurant stack has quietly grown a marketing feature.

Guest records exist in four places. The reservations platform knows who booked. The POS knows who paid. The delivery marketplace knows who ordered, and largely will not tell you. The email tool knows who subscribed. None of them agree, and no single one can answer the only question that matters: who used to come in and has stopped.

Review requests exist in three. The POS can text one, the reservations tool can email one, and the dedicated review tool exists to do exactly this. Restaurants routinely pay for all three and fire two at the same guest.

Email exists in two. The reservations platform sends marketing, and so does the email tool. Whichever list is smaller is the one somebody forgot about, and it is usually the one still being paid for.

In most businesses, consolidating tools is about efficiency. In a restaurant it is about the percentage bill, which makes it a different argument with a much larger number attached.

A guest who orders through a marketplace costs commission every single time, forever. The same guest ordering directly costs processing only. The difference on one order is small; the difference across a year of a regular's orders is not, and it compounds with every regular you convert.

Doing anything about that requires knowing who your regulars are, which requires a guest record that survives across channels. That is the actual argument for consolidation in a restaurant: not to save a subscription, but to make the expensive channel optional for the people who already like you.

First, the duplicate review tool. Pick whichever layer already asks and turn the others off. This usually removes a subscription outright and stops double-messaging guests, which is a small kindness as well as a saving.

Second, the second email list. Consolidate onto one, export the other, cancel it. Two half-lists are worth less than one whole one, and you are paying for the privilege.

Third, fix the guest record before touching anything operational. Do not start by ripping out the POS. Start by making sure that a person who books, eats and orders is one person in your records rather than three. Everything else you might want to do depends on that.

Last, look at the percentage bill. Not because it is least important, but because it is the one you change by building a direct channel rather than by cancelling something, and building takes longer than cancelling.

The general sequence and the migration plan are in the GTM stack consolidation playbook. What is particular to restaurants is that the biggest technology cost is not a subscription at all, so the usual advice to audit your software bill only ever finds the smaller half. For how the rest of the workflow fits together, the restaurant overview covers the wider picture.