Franchise Delivery Profitability: What the Channel Actually Costs

Franchise delivery profitability is decided one order at a time, not by a verdict on the channel. Four things set it: the commission your plan charges, packaging, the refunds and adjustments taken off your payout, and labour that does not fall when transactions move off the counter.

The argument about delivery is over in most networks. Operators are on the marketplaces, brands have stopped pretending otherwise, and the question has moved from whether to be there to what it is worth on a Tuesday.

That is harder than it sounds, because the channel arrives pre-netted. A payout lands, already reduced by things you did not itemise, and the sales line in your P&L stops meaning what it means everywhere else.

The four costs of a delivery order, and the two nobody counts

• Commission. The visible one, and not a single rate: the marketplaces publish tiered plans, higher tiers buy more exposure, and pickup is priced separately from delivery. Read your own agreement rather than the number another operator quoted at a meeting. You may not be on the same plan.
• Packaging. Containers, a bag, a seal, sometimes an insulated liner. Small per order, entirely real, and at most locations it sits in the same account as counter packaging, where nobody can separate it.
• Refunds and adjustments. Deducted from the payout after the fact, often weeks later, and reported in a portal rather than on the statement. This is the line that surprises people.
• Labour that does not scale down. The order still gets made, bagged and handed over. Quick-service labour averaged 26.69% of sales in 2024, down from 28.35% the year before, on PAR's QSR operational index — and none of that structure gets cheaper because the customer is at home.

The last two are missing from most operators' mental arithmetic, and they run in opposite directions: labour flatters the channel on a slow shift, refunds punish it over a quarter.

Franchise delivery profitability, worked on one order

Here is the calculation, on illustrative numbers rather than a real location's. Take a $30 menu subtotal with product cost at 30% of menu price: $9 of food, leaving $21 towards labour, occupancy and profit when that order is sold at your counter.

Now put it through a marketplace. Assume a commission of 20% — a placeholder for the arithmetic, not a published rate; your own plan sets yours. That is $6, plus $1.10 of packaging, plus the same $9 of product. The order contributes $13.90 instead of $21.

That gap is why marketplace menu prices are what they are. Lift the delivery menu 15% and the subtotal becomes $34.50. Product does not change — same sandwich — so the commission becomes $6.90 and the order contributes $17.50. Still below the counter, and much better than pretending the channel is free.

Two honest adjustments. Royalty and marketing fees are charged on gross sales, and the uplift raises the gross, so you keep less of the increase than the menu suggests. And a refunded order still cost you $9 of product and $1.10 of packaging — one lost dispute wipes out well over half of what a good order contributed.

Run those lines with your own contract in front of you. It is the only version of this arithmetic that means anything: not one of the four inputs is standard across a network.

Menu pricing on the marketplace, and who gets to decide it

A price uplift on delivery menus is ordinary practice now rather than a clever trick, and customers largely understand that a delivered item costs more than a collected one. What is not settled is who decides. Three positions exist, and each is defensible:

1. The brand sets it, as a network-wide percentage, usually to protect price perception across a metro where locations share a delivery radius.
2. The operator sets it, within a ceiling, on the argument that trade areas differ and the operator carries the P&L.
3. Nobody decides, and the result is three neighbouring locations at three different delivery prices, visible to the same customer in the same app.

The third is the common one, and the one that generates complaints. Whether a brand can mandate a delivery price depends on the agreement, the jurisdiction and advice your counsel gives — but it is a franchisee franchisor relationship conversation before it is a pricing one, and it goes better when the brand brings the arithmetic rather than the instruction.

None of it shows up in the industry's headline numbers, which count establishments and output — the picture in franchise industry statistics. Channel mix is in none of them, so nobody outside your own reporting knows what share of a network's sales now arrives through a marketplace.

The refund line, and the only way to fight it

Refunds are where operators lose money quietly, because the loss arrives detached from the order that caused it.

A customer reports a missing item, a cold meal, an order that never came. Some are true, some are not, and the marketplace decides. The operator's position is weak in one specific way: no evidence, because nobody photographs a sealed bag on a Friday night.

The routine that works is unglamorous and takes a fortnight to establish:

• Seal every bag and photograph it with the order number visible. It changes dispute outcomes more than any other single habit.
• Appeal inside the window. Every platform has one, it is short, and an unappealed adjustment is settled.
• Track appeal recovery as a line you own — how many contested, how many came back. If nobody owns that number, nobody is contesting.
• Reconcile the payout to the register weekly, not monthly. The deposit will not match the sales, and the difference is what you are managing.

That reconciliation is the most common place I see AI earning its keep at unit level — matching payouts to tickets and flagging what does not tie, rather than answering questions in a chat window. Dull uses are most of what landed this year, the pattern in state of AI franchising 2026.

What networks are standardising

Brands that used to leave this to operators are writing down four things, and writing them down does more work than whichever answer they land on:

Which channels the brand participates in. How delivery menu pricing works and who approves an exception. Who owns the storefront account — which decides who keeps the ratings and the customer history when a location changes hands. And whether marketplace data flows back into brand reporting at all, because a network that cannot see the channel cannot benchmark it.

The last one is where most brands still are: a channel that behaves like a separate business, priced differently, refunded differently, staffed the same — and no view of it beyond a franchisee's word that it is fine.

That is the real cost of the channel: not the commission, a known deduction, but the years of a location's economics nobody wrote down while everyone argued about the percentage.

Look for the gap first in what franchising does not publish about itself.