Dark Kitchen Business Model: 5 Fixes for Better Margins

A dark kitchen, also called a ghost kitchen or cloud kitchen, is a delivery-only kitchen with no dining room. That removes rent on a prime street unit and front-of-house wages. Yet many owners find the profit they expected never arrives. This guide is for dark kitchen owners and people planning one. By the end you will know where the dark kitchen business model leaks money, and five fixes you can start this month.
The pattern is simple. Low overhead is the advantage, delivery app commission is the leak, and the cure is control over orders, customers and data.
Where does a dark kitchen business model lose money?
Start with what apps charge. Uber Eats' US merchant pricing lists 20% (Lite), 25% (Plus) and 30% (Premium) on delivery orders, with pickup at 7% when in-store pricing is validated, as of November 2026. DoorDash's merchant pricing lists 15% (Basic), 25% (Plus) and 30% (Premier) for delivery. Rates vary by city and plan, so check your own contract. Our delivery app fees hub compares more platforms.
Here is a hypothetical dark kitchen with $20,000 of monthly sales, all through one app on a 25% plan. The cost percentages are illustrative assumptions, not industry benchmarks.
Line | % of sales | Monthly $ |
|---|---|---|
Food and packaging | 35% | $7,000 |
Kitchen labour | 25% | $5,000 |
Rent, utilities, insurance | 10% | $2,000 |
App commission | 25% | $5,000 |
Left before tax and other costs | 5% | $1,000 |
Now suppose 30% of those orders ($6,000 of sales) come through the kitchen's own ordering website instead. The app commission on that $6,000 would have been $1,500. Card processing on the direct orders costs roughly 3% to 4.5% depending on basket size ($180 to $270 at Stripe's 2.9% + 30 cents standard US rate), plus your software subscription. The profit left rises from $1,000 to roughly $2,200 or more. Same food, same kitchen, more than double the profit.
Use our commission calculator to run this with your own numbers, and read the real maths of commission vs flat fee for a longer worked example.
Fix 1: Move repeat customers to direct orders
You cannot move every order. Apps bring genuinely new customers. The aim is to move the second and third order.
Tactics that need no special software:
Put a flyer or sticker with a short link and QR code in every delivery bag.
Print the link on packaging, labels and receipts.
Add your ordering link to your Google Business Profile, Instagram bio and any social pages.
Offer a small, fixed incentive on a first direct order, and work out its cost before you start.
Check your delivery app's terms before you promote direct ordering inside its packaging or messages. Rules differ by platform and country.
Fix 2: Own the customer relationship
On a marketplace, the app holds the customer details and decides how much you see. On your own ordering website, customers create accounts with you. A branded site such as FoodFlick's keeps each customer's order history and saved addresses in your account, so you can see who your regulars are and what they order.
How you contact customers is a separate question. Marketing messages carry consent rules, such as CAN-SPAM and TCPA in the US and PECR and UK GDPR in the UK. Check what customers have agreed to before you contact them, and ask your provider which messaging tools they offer.
Fix 3: Reduce the number of screens in the kitchen
A kitchen that takes orders from two apps, a website and the phone has several order streams. Missed orders and late tickets follow.
You have two realistic options:
Limit channels. Run one app plus your own site, and drop apps that bring little volume.
Use a middleware tool that merges app orders into your system, if you need many apps. Not every ordering system offers this, so ask providers directly.
Whichever you pick, your own website orders and any walk-in or pickup-window orders should land in one place. FoodFlick handles online and walk-in orders in one system and sends them to the kitchen through kitchen workflows with thermal printing. Orders from delivery apps are managed in those apps' own tools.
Keep menus in sync too. Every price or availability change needs making in each app and on your own site, so assign one person to own that task and do it at a set time each day.
Fix 4: Use your own numbers, not instinct
Dark kitchens have an advantage here: every sale is digital. The useful questions are simple:
Which dishes sell most, and which earn the most after food cost?
Which hours and weekdays bring orders?
What is the average order value (AOV) on each channel?
How many customers order twice?
Our restaurant analytics guide covers the weekly numbers worth checking. FoodFlick's reports show sales by date range and product performance for the orders it handles. Calculate food cost percentage for each dish so that the best sellers are also the best earners.
Fix 5: Build a brand customers can find by name
Customers on an app see a list of near-identical restaurants and remember "the burger place on the app". A branded ordering website lets them search for you by name and return without an algorithm in the way.
You do not need a large budget:
Choose a short, easy-to-spell name and secure the matching web address.
Use the same logo, photos and menu wording on every channel.
Put your direct link on every printed item that leaves the kitchen.
If you run several brands from one kitchen, keep each brand's menu focused and its name distinct.
What is the best dark kitchen business model in practice?
For most operators, a blended model works best. Use apps to be found, use direct ordering to keep margin, and use your numbers to trim the menu. There is no single winner. If your kitchen sells only through apps and you have no marketing time, a delivery-only model can still earn a profit, provided you price your app menu to cover commission. When you compare, count every cost: commission, card processing fees, subscriptions, packaging and any delivery fees.
Common mistakes to avoid
Judging profit by sales. A $20,000 month on a 25% plan earns less than a $15,000 month with a third of orders direct.
Leaving app prices equal to menu prices. Many operators raise app prices to offset commission. Check the app's rules and customer reaction.
Chasing every app. More apps mean more screens and more menu updates.
Ignoring card fees. Direct orders are cheaper than commission, not free.
Opening a direct channel and not telling anyone. Customers will not find it unless you promote it.
How FoodFlick helps
FoodFlick is a commission-free online ordering system and POS, set up for you by the FoodFlick team. Dark kitchens get a branded ordering website on their own domain, delivery, pickup and dine-in ordering, delivery zones, customer accounts with order history, kitchen workflows with thermal printing, and sales and product reports. Plans start at $59 per month with 0% commission per order. Card processing fees from your payment provider still apply. Most restaurants go live within a few days of completing onboarding. See FoodFlick for cloud kitchens.
Sources
Uber Eats: Merchant pricing (US), checked November 2026
DoorDash: Merchant pricing, figures from DoorDash's published plans; the page blocks automated access, so confirm in a browser, checked November 2026
Stripe: Pricing, checked November 2026
Frequently asked questions
How do dark kitchens make money?
A dark kitchen sells delivery-only food from a kitchen with no dining room, so it saves on rent, front-of-house staff and decor. Income comes from delivery app orders and direct orders. Profit depends on food cost, labour, packaging and how much commission each channel takes.
What is the biggest risk in a dark kitchen business model?
Dependence on one or two delivery apps. They set the commission, control search ranking and own the customer relationship. If your sales come mostly from one platform, a fee change or ranking drop can cut your income fast.
Is a dark kitchen the same as a cloud kitchen or ghost kitchen?
In everyday use, yes. Dark kitchen, cloud kitchen and ghost kitchen all describe a delivery-focused kitchen without a customer-facing dining room. Some people use the terms for slightly different set-ups, such as shared hubs or virtual brands.
Should a dark kitchen leave delivery apps completely?
Not usually at the start. Apps bring new customers you could not reach alone. A common approach is to stay on one or two apps for discovery and give those customers a reason to order directly next time, such as a flyer in the bag with your ordering link.
Do I still pay fees on direct orders?
Yes. Direct orders avoid marketplace commission, but you still pay card processing fees to your payment provider and any subscription for your ordering software. Stripe's standard US rate, for example, is 2.9% plus 30 cents per online card payment.


