Home / Guides / Start a food delivery business

Guide · United Kingdom

How to Start a Food Delivery Business in the UK (2026)

Food delivery is one of the few businesses you can start small, test in a single postcode, and scale street by street. It is also a business where the economics are unforgiving if you get the basics wrong. This guide walks through the decisions that actually matter: which of the three viable models to run, the setup steps to check before you trade, how the unit economics work, what software you genuinely need at launch, and how the big aggregators fit into the plan.

Takeaway ordering apps →Delivery app development

In short

To start a food delivery business in the UK: pick one of three models — a local multi-restaurant platform, a single-cuisine delivery brand, or a delivery arm for an existing restaurant. Register as a food business with your local council and check current requirements on GOV.UK before you trade. Take professional advice on how you engage drivers, and talk to a broker about insurance. Keep your delivery radius tight so drivers complete more drops per hour, and launch with the smallest software stack that supports your model: an ordering channel, dispatch, and a driver view. Aggregators like Just Eat and Deliveroo are a reach channel, not a business model — most successful operators dual-run them alongside their own ordering channel.

Step 1 — pick your model

The three viable models — and their honest trade-offs

Almost every food delivery business in the UK is a version of one of these three. They look similar from the outside — food, drivers, an app — but they are very different businesses to run. Pick deliberately, because the model decides your costs, your software and your hardest problem.

Model 1

Local multi-restaurant platform

A mini-aggregator for one town or district: local restaurants list on your platform, you take the orders and run (or coordinate) the delivery.

Works when: the big aggregators under-serve your area, and you can offer restaurants a better deal than they get elsewhere.

Upside: commission on every order across many kitchens, and local network effects once it works
Hardest problem: a three-sided cold start — you need restaurants, customers and drivers at the same time

Honestly the heaviest model: most capital, most operations, most software. Start here only if you know your patch and its restaurants well.

Model 2

Single-cuisine delivery brand

One focused menu, built for delivery from day one — often from a rented or shared kitchen, with no dine-in at all.

Works when: you have a genuinely strong food concept and the discipline to keep the menu short enough to travel well.

Upside: full control of product, brand and margin; the simplest operation of the three
Hardest problem: demand is entirely yours to generate — no platform sends you customers unless you pay for the privilege

Concentrated risk: one brand, one menu. But it is the cheapest model to test, and the easiest to shut down or pivot if the numbers say no.

Model 3

Delivery arm for existing restaurants

You already run a restaurant or takeaway — or you deliver on behalf of local restaurants that do not want to run drivers themselves.

Works when: the kitchen and the demand already exist, and delivery is incremental revenue rather than the whole bet.

Upside: the lowest cold start — proven food, existing customers, a kitchen that is already paid for
Hardest problem: protecting margin, especially if you also stay on aggregators, and staffing drivers around sharp evening peaks

The pragmatic default for most founders reading this: add your own channel, measure it against the aggregators, and expand what works.

Step 2 — set up properly

Setup essentials: what to check before you trade

Not legal or regulatory advice. This section is general information to help you plan. Requirements vary by council and by nation within the UK, and they change — always check current requirements with your local council and on GOV.UK, and take professional advice for your specific circumstances.

Food business registration. Expect to register as a food business with your local council before you start trading — this is the normal first step for anyone preparing, storing or handling food commercially, whether from a home kitchen, a shared unit or a commercial premises. The process and timing are set out on GOV.UK and your council's website; check them early, not the week you launch.
Food hygiene expectations. Plan for your premises to be inspected and rated, and build hygiene into your routines from day one: temperature control, cleaning schedules, allergen information for every item on the menu, and traceability of ingredients. The current official guidance sets out the specifics — a recognised food safety course for you and anyone handling food is a sensible early investment, and packaging that keeps food safe and hot in transit is part of the job, not an afterthought.
Premises. Be upfront with your council about what you are running and from where — a home kitchen, a rented commercial kitchen and a dark kitchen unit are treated as what they are. If you rent, check that your lease or licence actually permits food preparation and delivery operations.
Drivers: employed or self-employed. How you engage drivers is one of the most contested areas in this industry, and it affects tax, rights and your obligations as an operator — take professional advice before you decide rather than copying what a competitor does. Whichever route you take, be clear about it in writing with every driver.
Insurance. Talk to a broker before your first delivery and ask what applies to your setup: vehicle cover appropriate for delivery work for anyone driving (personal policies commonly do not cover it), public liability, product liability for the food itself, and employers' liability if you take on staff. A broker will tell you what is required and what is merely wise for your circumstances.

Step 3 — learn the economics

Unit economics: the business lives or dies on drops per driver-hour

Strip away the app and the branding, and a delivery business is a simple machine: each order must earn more than it costs to make and deliver. The delivery side of that equation is governed by three linked numbers.

Delivery radius sets how long each journey takes. Order density — how many orders come from each square mile inside that radius — sets how efficiently a driver can chain drops together. Together they determine your driver cost per drop: a driver's hour costs roughly the same whether they complete many deliveries or one, so the more drops per hour, the cheaper each one becomes.

This is why the counterintuitive advice is right: shrink the map. A tight radius means short drives, more drops per hour, hotter food and better reviews. A wide radius means long drives, expensive drops and lukewarm biryani. Depth beats breadth — it is better to be the obvious choice in three postcodes than an occasional option in fifteen.

Your margin per order — what is left after ingredients, packaging, payment fees and any commission — has to clear the driver cost per drop with room to spare for marketing and everything else. Model it honestly for the quiet Tuesday lunch, not just the heaving Friday night, because the same driver-hour maths gets much worse off-peak.

The questions that decide profitability

Drops per driver-hour at peak — the single most important operational number
Driver cost per drop — a driver's hour divided by the drops they complete in it
Contribution margin per order — after ingredients, packaging, payment fees and commission
Order density — orders per square mile inside the radius, not total orders
Off-peak reality — what all four numbers look like on a wet Tuesday

Track these weekly from your first week. Widen the radius only when the core area is consistently profitable — and be as willing to shrink it again.

Step 4 — the technology

The software you actually need at launch

At launch you need three things, and only three: an ordering channel (a fast website first — an app earns its keep once repeat customers justify it), dispatch (getting the right order to the right driver, which can be a person with a screen at tiny volume but becomes software quickly), and a driver app or driver view (the run of jobs, the route, order status and proof of delivery). Everything else — loyalty, scheduling, multi-kitchen menus — can wait until the model is proven. Build the smallest tier that supports your model.

What you are launching	Typical build	Timeline	Price band (GBP, ex VAT)
Single-restaurant ordering site — your menu, checkout and order management	Custom storefront	5–8 weeks	£6,000–£12,000
Single-vendor delivery app (web) — ordering plus dispatch and delivery tracking for one brand	Focused delivery app	6–9 weeks	£8,000–£14,000
Multi-vendor delivery platform — multiple restaurants, commission handling, dispatch	Standard delivery platform	9–13 weeks	£14,000–£26,000
Delivery platform + driver & customer apps — the full aggregator-style stack	Full delivery product	10–16 weeks	£20,000–£40,000

Prices published from our Open Price Book (v1.0 · July 2026 · next review October 2026). All prices exclude VAT.

If you are running Model 2 or Model 3, start with our takeaway ordering app page — that is the storefront and single-vendor end of the table. If you are building a multi-restaurant platform (Model 1), see delivery app development for the platform and driver-app tiers. Either way you get senior engineers on UK hours and milestone billing — you pay for delivered, accepted work.

Step 5 — the aggregators

How Just Eat and Deliveroo fit into the plan

Be honest about what the big aggregators do well, because it is a lot: they put your food in front of customers you could not reach on day one, they handle payment and (on some arrangements) the delivery logistics, and their checkout carries a trust a new brand has not earned yet. For a new operator, that reach is genuinely valuable.

Now be honest about the trade-offs. They typically charge a commission on each order, which comes straight out of the margin you just modelled so carefully. The customer relationship — the name, the ordering history, the ability to bring someone back with an offer — generally stays with the platform, not with you. And your visibility depends on a ranking you do not control.

The pragmatic answer for most operators is to dual-run. Use the aggregators as a discovery channel — that is what they are best at — while building your own ordering channel for the customers who already know you. Repeat orders through your own site or app keep more of the margin and give you your own customer data to work with. Convert regulars gently: a card in the bag, a loyalty nudge, a reason to order direct next time — within whatever your agreements with the platforms allow.

Treat the split as a number to watch, like everything else: what share of orders is direct, and is it growing? A delivery business that owns none of its demand is a tenant. Dual-running is how you stop being one without giving up the reach.

Step 6 — launch

The launch checklist

Work through this in order. Each line is covered in the sections above.

Pick one model — platform, delivery brand, or delivery arm — and write down its hardest problem so you attack it first
Choose a tight launch area around your densest expected demand — postcodes, not "the city"
Register your food business with your local council and check the current requirements on GOV.UK before trading
Sort the premises and hygiene routines — temperature control, cleaning, allergen information, delivery-worthy packaging
Decide the driver model with professional advice and put the arrangement in writing with every driver
Talk to an insurance broker about vehicle cover for delivery work, public and product liability, and employers' liability if you employ staff
Stand up the launch stack — ordering channel, dispatch, driver view — at the smallest tier that supports your model
List on aggregators if reach matters more than margin at your stage, and plan the direct-channel conversion from day one
Soft-launch to the tight area, measure drops per driver-hour and margin per order weekly
Widen the radius only when the core is consistently profitable — and shrink it again if the numbers say so

Questions & answers

Starting a food delivery business — FAQ

Do I need to register a food delivery business in the UK?

You should expect to register as a food business with your local council before you start trading, and to meet the food hygiene standards that apply to your premises and menu. Requirements vary by council and change over time, so check the current guidance on GOV.UK and confirm the specifics with your local council before you launch. This guide is general information, not legal or regulatory advice.

Which food delivery model is best for a first-time founder?

It depends on what you already have. If you run a restaurant or takeaway, adding your own delivery channel is usually the lowest-risk start. If you have a strong food concept but no premises, a single-cuisine delivery brand keeps the operation simple. A local multi-restaurant platform has the biggest upside but is the hardest model, because you are recruiting restaurants, customers and drivers at the same time.

Should delivery drivers be employed or self-employed?

There is no universal answer, and driver classification has been actively contested in the UK, so take professional advice before you decide — the choice affects tax, rights and your obligations as an operator. Whichever route you take, ask an insurance broker about the cover your drivers and your business need, including vehicle cover appropriate for delivery work, and check the current guidance on GOV.UK.

How much does the software for a food delivery business cost?

From our Open Price Book (ex VAT): a single-restaurant ordering site typically runs £6,000–£12,000 as a custom storefront, or £8,000–£14,000 as a focused single-vendor delivery app with dispatch and tracking. A multi-vendor delivery platform runs £14,000–£26,000, and a full platform with driver and customer apps runs £20,000–£40,000. Most single-brand launches do not need the full tier on day one.

Should I use Just Eat and Deliveroo or build my own ordering channel?

For most operators it is not either/or. Aggregators are genuinely good at discovery and put you in front of customers you could not reach on day one, but they typically charge a commission on each order and the customer relationship stays with the platform. A common approach is to dual-run: use aggregators for reach while building your own ordering channel for repeat customers, better margins and your own customer data.

How big should my delivery radius be at launch?

Smaller than most founders want it to be. A tight radius keeps drive times short, which means more drops per driver per hour, hotter food and better reviews. Every extra minute of driving reduces how many orders a driver can complete in an hour and raises the cost of each drop. Start tight around your densest demand, measure drops per driver-hour, and only widen the radius once the core area is consistently profitable.