Quick answer
There's no universal winner: it depends on your volume and the density of your area. Platforms cost a lot in commission (28-35% plus tax) but carry no fixed cost: you only pay when you sell. In-house delivery has zero commission but high fixed costs (riders, vehicles, software). Below a certain order threshold the platform wins; above it, your own delivery does. The break-even point is calculated, not guessed.
The two cost logics side by side
The key difference between the two models isn't the percentage: it's the cost structure.
- Platform = variable cost. You pay a percentage on every order. Zero orders, zero cost. The risk is low, but the cost per order stays high forever, even as you grow.
- In-house delivery = fixed cost. You pay wages, vehicles and software regardless of orders. The cost per order is sky-high if you sell little, but collapses as volume climbs.
It's the same logic as renting a car for a day (variable) versus buying one (fixed): below a certain mileage renting wins, above it buying does. The whole own-delivery-vs-platforms debate revolves around where that break-even point falls.
What the platform channel costs
On platforms you pay a percentage commission on the gross order value, tax included, and on that commission you pay tax again. Typical market bands:
| Platform model | Typical commission | Fixed cost | |---|---|---| | Marketplace only (your riders) | 12-20% | None | | Delivery included | 28-35% | None | | With sponsored placement | +5-10% extra | Variable |
On a £30 order with delivery included at 30%, the real cost climbs past £10 once you add tax on the commission. To avoid getting the numbers wrong on each setup, use the delivery commission calculator: it shows what you actually keep after commission and tax.
What in-house delivery costs
Here the costs are fixed and must be estimated honestly. The main lines for a full-time rider:
| Line item | Indicative monthly cost | |---|---| | Gross wage + payroll taxes | £1,400-1,800 | | Vehicle (lease or depreciation) | £130-260 | | Fuel / energy | £70-130 | | Insurance and maintenance | £45-90 | | Order-management software | £25-70 | | Monthly total | ~£1,670-2,350 |
Take an average of £2,000 a month. If that rider delivers 15 orders a day over 26 days, that's 390 orders a month: the cost per delivery is £5.13. If they deliver 25 a day (650 a month), it drops to £3.08. Volume is everything.
The break-even point: a worked example
Let's compare both models on the same average £30 order, with a 30% platform commission plus tax.
Platform channel cost per order:
Commission 30% on £30 = £9.00
Tax 20% on the commission = £1.80
Platform cost per order = £10.80
In-house delivery cost per order, by volume (rider at £2,000/month):
| Orders/day per rider | Orders/month | Cost per delivery | |---|---|---| | 10 | 260 | £7.69 | | 15 | 390 | £5.13 | | 20 | 520 | £3.85 | | 25 | 650 | £3.08 |
In-house delivery beats the platform (£10.80 per order) even at 10 orders a day in this example. But beware: this comparison only holds if those orders exist. The platform brings demand; in-house delivery requires you to generate orders yourself.
The real hidden cost of in-house delivery: demand
The calculation above has a huge limit: it assumes the rider is always busy. In reality, dropping platforms means losing their shop window. Most delivery orders happen because the customer opens the app and finds you there.
To fill your own delivery you have to build demand with your own tools: an ordering website, an optimised Google profile, WhatsApp Business, loyalty, local flyers. That's real work and it has a cost. This is why the model that almost always works is hybrid:
- You stay on platforms for visibility and distant areas.
- On every package you add a note: "Order direct, same price, faster delivery."
- You gradually migrate loyal customers to the direct channel, where commission is zero.
That way your own rider mostly delivers high-margin direct orders, and platforms stay an acquisition channel, not your only source.
Quick decision table
| Situation | Recommended model | |---|---| | Few orders, wide area | Platforms only | | High volume, tight area | Own delivery (+ platforms for acquisition) | | Little-known brand | Platforms to get found | | Loyal, repeat customers | Push the direct channel | | High average order value | Own delivery pays off sooner | | Low average order, many orders | Percentage commission hurts a lot: consider in-house |
Common mistakes
- Comparing only the percentage. A 30% commission isn't compared to "zero" for in-house: in-house has fixed costs that must be divided across real orders.
- Underestimating the rider's cost. Many count only the wage and forget vehicle, fuel, insurance and idle time between deliveries.
- Assuming dropping platforms won't dent volume. The platform's shop window brings demand: without a direct acquisition plan, your own delivery stays empty.
- Ignoring price-parity clauses. Some contracts forbid showing lower prices on your channel: read before promising discounts on your site.
- Not splitting by area. In-house works in the short, dense radius; leave distant zones to the platform.
- Forgetting idle time. A rider doesn't deliver 25 orders back to back without breaks: use realistic, not theoretical, volumes in your cost-per-delivery maths.
Related resources
- Delivery commission calculator — what you actually keep on each platform order, tax included, to compare against the cost of your own delivery