A restaurant can be busy on DoorDash, Uber Eats or another delivery marketplace and still have a weak direct digital relationship with its customers.
The opposite can also happen: a restaurant builds an expensive online-ordering website, then discovers that most new customers still find it through the marketplaces it was trying to replace.
So the useful question is not “Are delivery apps bad?” or “Should every restaurant take orders on its own website?”
It is:
Which parts of discovery, ordering, payment and fulfilment should the restaurant own directly, and which parts are worth buying from a marketplace?
For many independent restaurants, the answer is a hybrid system rather than an all-or-nothing choice.
Delivery apps and your website solve different problems
A delivery marketplace can put a restaurant in front of people who have not decided where to eat yet. The customer opens the app, compares options, filters by cuisine, sees delivery estimates and places an order inside an environment they already trust.
That reach has value.
The restaurant’s own website has a different advantage. It serves people who already know the business, search for it by name, follow a link from Google or Instagram, scan a QR code, or return after a previous visit.
Those customers may want to:
- see the current menu;
- order pickup or delivery;
- reserve a table;
- ask about catering or group orders;
- check opening hours and location;
- view dietary information;
- contact the restaurant directly.
A strong website makes those branded journeys simple without pretending the marketplace discovery channel no longer matters.
Compare the economics at the order level, not by slogan
Marketplace commission is one reason restaurants explore direct ordering, but the comparison should be specific.
For example, DoorDash currently publishes US Marketplace delivery commission tiers of 15%, 25% and 30%, with pickup commission at 6%. Its own merchant pricing page also states that orders placed through DoorDash Online Ordering are commission-free but still carry payment-processing costs.
DoorDash separately describes its Online Ordering product as a way to accept orders through a restaurant’s own channel while integrating with supported POS systems.
Those figures and products are market-specific and can change. Uber Eats pricing also varies by country and product. Uber’s merchant material distinguishes its marketplace from Uber Direct, which is designed to let businesses use Uber’s delivery network for orders that originate through their own channels in supported markets.
The important comparison is therefore not simply:
marketplace fee vs zero fee
It is:
marketplace order cost vs direct order cost + operating complexity
For direct ordering, account for things such as:
- payment processing;
- ordering software or custom development;
- delivery costs if you do not have drivers;
- POS or kitchen integration;
- menu and stock synchronisation;
- refunds and chargebacks;
- support when an order fails;
- maintenance and monitoring.
A direct channel is only better if it works reliably enough to protect the customer experience and the restaurant’s margin.
Do not build custom ordering when a simpler integration solves the problem
There is a temptation to treat “owning the channel” as meaning the restaurant must build every part itself.
It does not.
Your own domain can remain the branded destination while a specialist ordering provider handles the transactional complexity. The important questions are ownership, portability, customer journey and operational fit.
A restaurant website can, for example:
- present the brand, menu, locations and trust information;
- send the customer into a well-integrated ordering experience;
- keep the transition clear and visually consistent;
- return useful confirmation and support information;
- measure where orders originated.
That may be a better system than paying to recreate mature ordering infrastructure from scratch.
Custom development becomes more defensible when the restaurant has requirements that standard products cannot handle well: unusual fulfilment rules, multi-location logic, complex catering workflows, membership, loyalty, wholesale ordering or deep integrations that materially affect operations.
The hardest part is not the Order Now button
A direct-ordering interface can look finished long before the ordering system is safe to depend on.
The real flow is closer to:
menu → item options → availability → cart → taxes/fees → payment → order creation → kitchen/POS notification → preparation → pickup/delivery → customer confirmation → cancellation/refund if needed
Every arrow is a failure point.
Before launching, test situations such as:
- payment succeeds but order creation fails;
- order is created but the kitchen never receives it;
- an item sells out after a customer adds it to the cart;
- the customer submits twice because the button appears unresponsive;
- the restaurant closes early but online ordering remains open;
- a delivery address is outside the supported area;
- the restaurant cancels an order after payment;
- a refund is required;
- a menu price changes in one system but not another.
This is where product thinking matters more than simply producing attractive screens.
If money changes hands, define the order state clearly. The website should know the difference between an attempted payment, a confirmed payment, an accepted order, a cancelled order and a refunded order rather than treating every click on “Pay” as a successful sale.
Menu sync can decide whether direct ordering succeeds
Restaurants change prices, modifiers, availability and specials frequently.
If staff must maintain one menu for the POS, another for the website and separate copies for multiple delivery apps, inconsistency becomes an operational problem.
Before choosing an ordering setup, ask:
- Which system is the source of truth for menu items and prices?
- Can availability update quickly?
- Do modifiers and add-ons stay consistent?
- What happens when an item is sold out?
- Can staff change the menu without calling the developer?
- Does the ordering tool integrate with the existing POS?
The best-looking website can still create angry customers if it confidently sells food the kitchen cannot fulfil.
Use your own website for the enquiries marketplaces handle poorly
Not every valuable restaurant conversion is a standard meal delivery.
Your website can own higher-context enquiries such as:
- catering;
- private dining;
- event bookings;
- large group reservations;
- corporate lunch orders;
- venue hire;
- partnership enquiries;
- gift cards or special packages.
These journeys often need more explanation than a marketplace listing provides.
A catering page, for example, can explain minimum order sizes, lead times, service areas, menu packages and the information the restaurant needs before quoting. That reduces back-and-forth and gives the enquiry form an actual job.
Make branded search lead somewhere useful
When someone searches the restaurant’s name, the official website should help confirm that they have found the right business and give them a clear next step.
Keep core information accurate and easy to maintain:
- menu;
- opening hours;
- address and directions;
- phone number;
- reservation route;
- direct-ordering route if available;
- delivery marketplace links where useful.
Google’s Business Profile platform supports structured business offerings such as food menus; its Business Profile API documentation describes publishing menu data for locations. The website and Business Profile should reinforce each other instead of showing conflicting hours, menus or links.
If you are deciding whether your Google presence alone is enough, my guide on Google Business Profile vs an owned website explains that broader ownership question.
When should a restaurant add direct ordering?
Direct ordering becomes more attractive when several of these are true:
- customers already search for the restaurant by name;
- repeat customers regularly order the same way;
- the restaurant has enough order volume to justify the setup;
- marketplace commission materially affects margin;
- pickup is important;
- the POS or ordering provider can integrate cleanly;
- the restaurant has staff who can own menu and order operations;
- catering, reservations or other direct journeys already bring people to the website.
It may be too early when the restaurant has little direct demand, weak operations, frequently inaccurate menus, no one responsible for failed orders, or a website that still struggles with basic mobile usability.
In that case, improving the public website and measurement may create more value before adding transactional complexity.
A practical hybrid setup
For many restaurants, I would start with this model:
1. Keep marketplaces for discovery. Let them introduce the restaurant to customers who are browsing broadly.
2. Make the restaurant’s domain the branded home. Menu, location, proof, reservations, catering and direct contact should be easy to find.
3. Add direct ordering only when operations can support it. Use a proven integration before assuming custom software is necessary.
4. Connect the menu and POS where practical. Reduce duplicate data entry and stale availability.
5. Track channel performance. Measure completed orders, average order value, repeat orders, failed orders and acquisition cost where data allows.
6. Test the full order lifecycle. Do not stop testing at payment success.
7. Keep account ownership with the restaurant. Domain, DNS, analytics, ordering-provider account, payment account and recovery access should remain under business control.
That gives the restaurant optionality without removing a channel that may still be generating valuable new demand.
Your website should become the place repeat customers know to return to
Delivery marketplaces can be useful rented distribution. Your website is the place where the restaurant can organise its own brand, information and direct customer journeys.
The goal is not to force every customer away from an app they prefer. It is to make sure a customer who actively looks for your restaurant has a reliable direct destination.
If your restaurant website cannot support direct enquiries, reservations or ordering cleanly, or you are unsure whether you need a simple integration or a more custom ordering flow, see my web design service or tell me what the current setup needs to do.
