Product · 6 min read
On-Demand App Business Models: Commission, Fees, Subscriptions, SaaS and Ads, With Per-Order Unit Economics
On-demand app business models compared: commission, delivery fees, subscriptions, SaaS for merchants and ads, with a worked per-order unit economics table.
Zohaib KhalidCEO & Co-founder, Innovation Insight·Published 
Short answer
On-demand apps earn money five ways: commission from merchants, delivery and service fees from customers, a monthly subscription for free delivery, software fees from merchants, and advertising. Most profitable platforms combine three or more. Whatever the mix, check the per-order contribution first: revenue per order minus provider pay, payment fees, refunds, support and acquisition. Our worked example with labelled assumptions shows how a thin positive margin turns negative.
Five ways an on-demand platform makes money
| Model | Who pays | How it works | Watch out for |
|---|---|---|---|
| Commission | Merchant or provider | A percentage of each order or job, often in tiers that include more delivery radius or marketing | Merchants raise menu prices or leave; some cities have capped delivery commissions |
| Delivery and service fees | Customer | A delivery fee by distance or demand, plus a service fee as a percentage of the basket and a small-order fee | Fee stacking at checkout is the top reason customers abandon carts |
| Subscription | Customer | A monthly or annual membership with free or reduced delivery above a minimum basket, in the style of DashPass or Uber One | Members order more but pay fewer fees; it only works if frequency rises enough |
| SaaS for merchants | Merchant | A monthly fee for online ordering on the merchant's own site, a POS integration or marketing tools, with lower or no commission | You become a software company with churn and support costs |
| Advertising | Merchants and brands | Sponsored listings in search and the home feed, promoted products, offers funded by brands | Needs traffic first; too many ads damage search trust |
At launch you will usually run commission plus customer fees, because both scale with orders and need no extra product. Subscriptions come once you have repeat customers worth keeping; advertising once there is enough traffic for merchants to compete for; SaaS when merchants ask for tools beyond your marketplace.
Worked example: unit economics per order
The table below is an illustrative food delivery order. Every number is an assumption we chose to make the arithmetic visible, not a market figure or any company's actual rate. Replace them with your own quotes and pilot data.
| Line | Assumption | Per order |
|---|---|---|
| Basket value (food) | Assumed average order | $35.00 |
| Commission from restaurant | Assumed 20% of basket | +$7.00 |
| Delivery fee from customer | Assumed flat fee | +$3.99 |
| Service fee from customer | Assumed 10% of basket | +$3.50 |
| Platform revenue | Sum of the three lines above | $14.49 |
| Courier pay | Assumed per-delivery pay for a 25-minute trip, before tips (tips pass through in full) | -$7.50 |
| Payment processing | Stripe's published US rate of 2.9% + 30 cents on the $42.49 the customer pays (checked October 2026) | -$1.53 |
| Refunds and credits | Assumed 2% of basket for wrong or missing items | -$0.70 |
| Customer support | Assumed share of support cost per order | -$0.50 |
| Maps, SMS, hosting | Assumed per-order infrastructure | -$0.20 |
| Contribution before marketing | Revenue minus the five costs above | $4.06 |
| Acquisition and promotions | Assumed $30 to win a customer who places 15 orders, plus $1.00 of promo credit per order | -$3.00 |
| Contribution per order | What is left for salaries, offices and the software | $1.06 |
Three things to take from it. First, the margin is thin: about a dollar on a $35 order under these assumptions. Second, courier pay is the largest cost, so dispatch quality (shorter trips, batching two orders, less waiting at the restaurant) has more effect on profit than any fee change. Third, a promotion of $5 off, common at launch, wipes out the contribution entirely; budget launch promotions as an investment with an end date, not as the business model.
The same order with a subscription
Assume a membership at $9.99 a month that waives the delivery fee and halves the service fee on orders over $35, and that members place six orders a month instead of four. The assumptions are ours, for illustration. Per member order, revenue falls by $3.99 and $1.75 to $8.75, and processing drops to $1.37, so contribution before marketing is about minus $1.52. The $9.99 fee spread over six orders adds about $1.67, leaving about $0.15 an order, or roughly $0.87 a month per member, against $16.24 a month from the same customer placing four orders without the plan. Under these assumptions the plan loses money. It pays only if it changes behaviour more than this: members who would otherwise have left, larger baskets, or lower courier cost because members' orders batch well. That is why real membership plans set a minimum basket and keep part of the service fee rather than waiving everything. Run your own numbers before launching one.
Own fleet vs delivery-as-a-service in the numbers
With delivery-as-a-service (DoorDash Drive, Uber Direct or a local courier API), courier pay is replaced by a per-delivery fee from the provider, and you drop the provider app, dispatch and courier support from the build. If that fee is higher than your $7.50 assumption plus the support and insurance costs of your own couriers, your contribution falls; in exchange you need no courier supply on day one. Get real quotes for your city and drop them into the table. Our on-demand app development guide explains when to switch from one to the other.
SaaS for merchants: a different business
Selling software to merchants (ordering on their own website, a kitchen display, POS sync, loyalty) gives revenue that does not depend on order volume, and it can be priced as a flat monthly fee with a lower commission. It also brings a SaaS company's problems: onboarding, churn, support and feature requests from every customer. We have built credit and subscription economies before: on Seek My Service, vendors subscribe to service areas priced by population data, buy credits through Stripe and spend daily credits to submit proposals. The opposite also works: CareFlair launched as a commission-free marketplace with no lock-in. Either way, the billing model has to be designed into the admin from the start; see our guide to SaaS development cost.
Advertising: when it is worth building
Sponsored listings are high-margin revenue because they cost almost nothing to deliver once search and ranking exist. They need three things: enough traffic that a top slot is worth paying for, a self-serve tool for merchants to set budgets and see results, and clear labelling so customers trust the results. Most platforms add them in the second or third year. Build the search and ranking so a sponsored slot can be inserted later, and leave the ad tools until merchants ask how to appear higher.
Pricing mistakes we see in plans
- Setting fees from a competitor's checkout instead of from your own cost per order. Their courier pay, density and order size are not yours.
- Leaving refunds, credits and support out of the model, then finding they take a dollar or more per order.
- Treating launch promotions as permanent. Give every promotion an end date and a target number of repeat orders.
- Hard-coding fees in the app. Put delivery, service and small-order fees in the admin by zone and time so you can test them weekly.
- Showing fees only at the last step. Display the delivery fee on the merchant page and the full total before payment, which some jurisdictions now require.
- Pricing a subscription before knowing order frequency. Run a pilot with a test group first.
What the business model changes in the build
| Model | Extra software | When to build it |
|---|---|---|
| Commission and fees | Fee rules by zone, time and basket in the admin; ledger with splits; transparent fee breakdown at checkout | MVP |
| Subscription | Membership billing, eligibility rules at checkout, cancellations and refunds, member reporting | Once repeat customers exist |
| SaaS for merchants | Merchant billing, plans, white-label ordering pages, POS integrations | When merchants ask for tools |
| Advertising | Sponsored slots in search, merchant campaign tool, reporting, budgets and billing | When traffic justifies it |
Builds for these platforms start at $45k to $82k for a food delivery MVP and $49k to $90k for grocery with a senior offshore team. See on-demand app development, food delivery app development and grocery delivery app development, or open the app development cost calculator with the food delivery preset loaded.
Sources
- Stripe pricing (2.9% + 30 cents for US domestic cards): https://stripe.com/pricing
- Stripe Connect pricing: https://stripe.com/connect/pricing
- Innovation Insight, Seek My Service and CareFlair case studies
- Innovation Insight app development cost calculator and pricing
Need a number for your project?
Send a short brief and get a written estimate.
A senior engineer replies within one business day. No sales call required.
Zohaib Khalid
CEO & Co-founder, Innovation Insight
Zohaib leads strategy, client partnerships and delivery at Innovation Insight. He has spent a decade turning founder briefs into products that ship and still reviews every proposal that leaves the company.
LinkedInRelated questions.
What is the most common on-demand app business model?
Commission from merchants plus delivery and service fees from customers. Both scale with orders and need no extra product, which is why almost every platform launches with them.
Is an on-demand app profitable per order?
Only if revenue per order covers provider pay, payment fees, refunds, support and acquisition. In our illustrative food delivery example the contribution is about a dollar on a $35 order, and a $5 launch promotion makes it negative.
Do subscriptions like DashPass make money?
Only if they change behaviour enough. In our illustrative example, a $9.99 plan that waives delivery and halves the service fee loses money even at six orders a month. Plans pay off through retention, larger baskets and a minimum order value, so model the fees you waive against the membership fee before launching one.
Should a delivery platform charge merchants a SaaS fee instead of commission?
It suits merchants with steady volume who want lower commission and their own ordering page. It turns you into a SaaS business too, with onboarding and churn to manage.
How much does it cost to build an on-demand app?
With a senior offshore team, about $45k to $82k for a food delivery MVP, $46k to $85k for ride-hailing and $49k to $90k for grocery delivery.