A family in Arcadia opens an app on a Tuesday night and orders $52 of enchiladas and a side of your smoked-brisket tacos from your Phoenix kitchen. Your line cooks fire it, your driver-handoff runs clean, the guest is happy. Then the marketplace takes its cut — up to 30%, or roughly $15.60 of that ticket — and hands you back a customer you’re not even allowed to keep. You never learn their name, their phone number, or that they love the brisket. Next Tuesday, the app shows them your competitor’s promoted listing, and you pay again to win back a guest you already fed.
That’s the quiet math crushing independent restaurants across the Valley — from Roosevelt Row to Old Town Scottsdale to Tempe. This is a practical, numbers-first guide to the fix: custom online-ordering software you actually own, so the margin and the guest data both stay in your restaurant instead of on someone else’s balance sheet.
Table of Contents
- The 30-second answer
- The margin math: why 30% commissions erase your profit
- Off-premise ordering is your main channel now
- The bigger loss: you never own the guest
- Your guests already want to order direct
- What “own your ordering” custom software actually looks like
- Marketplace app vs. a portal you own
- How we build it (and what it costs)
- A practical first-party rollout for a Phoenix restaurant
- Frequently asked questions
The 30-Second Answer
Third-party delivery apps are a rented channel that taxes every order and keeps your customer; a custom online-ordering portal is an owned channel that keeps both the money and the data. For a Phoenix restaurant running on a 3–5% margin, handing 20–30% of a ticket to a marketplace isn’t a marketing cost — it can be the whole profit and then some. The apps are worth keeping as a discovery tool for guests who’d never find you otherwise. But your regulars, your repeat orders, and anyone who came in once and loved the food should be ordering through your system, on your domain, into your database.
Custom software makes that switch practical: a branded ordering page that takes payment directly, captures name, phone, and order history on every checkout, and pipes it into your CRM so your loyalty club, birthday offers, and win-back texts run themselves. You pay card processing (~3%) instead of a 30% marketplace tax — and you finally own the guest relationship the apps hold hostage. Everything below is the evidence and the build path.
The Margin Math: Why 30% Commissions Erase Your Profit
Start with what a restaurant actually keeps. The National Restaurant Association pegs typical pre-tax profit margins at about 3–5%, and in its 2025 read on the industry, 42% of operators said they were not profitable as food, labor, and occupancy costs stayed elevated (NRA, 2025). This is a pennies-on-the-dollar business even when it’s healthy.
Now layer on the marketplace take rate. DoorDash’s own published pricing offers restaurants three Marketplace tiers — 15% (Basic), 25% (Plus), and 30% (Premier) commission on delivery orders, plus a 6% rate on pickup (DoorDash Merchant Pricing, 2025). Uber Eats mirrors it almost exactly, with 15% / 25% / 30% delivery tiers (Restaurant Dive, 2024). The higher tiers buy you more visibility in the app — which means the more you want the app to send you orders, the more of each order it takes.
Put the two numbers next to each other and the problem is obvious: the commission is several times the size of the margin.
Make it concrete. On that $52 Arcadia order, a 30% Premier commission is $15.60. Sell the identical order through your own site and you pay roughly 3% card processing — about $1.56. The difference, ~$14, is real money you either keep as profit or reinvest in better ingredients and staff pay. Do 25 direct orders a night at a similar delta and you’ve protected roughly $10,000 a month that was walking out the door in commissions. That’s not a rounding error on a 3–5% margin — it’s the difference between the 42% who aren’t profitable and the ones who are.
Cities noticed how punitive the take rate is. New York City made a permanent 15% cap on delivery commissions (plus a 5% cap on other fees) (Restaurant Dive, 2021), and Portland locked in permanent 15% delivery / 4% takeout caps (City of Portland, 2023). Arizona hasn’t legislated a cap — so in Phoenix, protecting your margin is on you.
Off-Premise Ordering Is Your Main Channel Now
It would be easy to shrug this off as a small slice of the business. It isn’t. The NRA reports that off-premises orders — takeout, drive-thru, and delivery — now make up roughly 75% of restaurant traffic, and a majority of operators say off-premise is a larger share of sales than it was in 2019 (NRA, 2025). Three of every four orders never sit down at a table.
The market backs the behavior: the U.S. meal-delivery segment is projected to top $100 billion in 2025 (Statista, 2025). Whatever channel most of your orders flow through is your restaurant’s storefront. If that storefront is a third-party app, then a marketplace — not you — owns your busiest register, sets the fees on it, and decides which competitor to show your customer next.
For a Valley operator, that’s the whole argument for building an owned channel. Phoenix is a serious restaurant town — the Arizona Restaurant Association counts roughly 12,000 restaurants, about 300,000 employees, and around $23 billion in annual sales statewide (Arizona Restaurant Association, 2024). In a market that dense and that competitive, the operators who own their ordering and their guest list will out-market the ones renting both.
The Bigger Loss: You Never Own the Guest
The commission is the cost you can see. The data you never receive is the one that quietly caps your growth. When an order comes through a marketplace, the customer belongs to the app. You typically don’t get a usable phone number or email, you can’t text them a Tuesday-night special, you can’t drop them into a birthday club, and you can’t win them back when they lapse. You rented a transaction; you didn’t earn a relationship.
That matters because repeat guests are where restaurant profit actually lives. A first-party ordering system flips the model — every checkout captures the name, phone, order history, and preferences, and lands them in your CRM. That’s the raw material for everything that compounds: loyalty, win-backs, and lifetime value. We’ve written the playbooks for exactly this — how to win back lapsed diners with a timed sequence, and why customer lifetime value beats one-off acquisition every time. None of it works if a marketplace is holding your guest list.
Your Guests Already Want to Order Direct
Here’s the part operators underestimate: switching guests to your own channel isn’t a fight against consumer habit — it’s aligned with it. In NCR Voyix’s 2025 customer experience research, 58% of customers said they prefer to order delivery through the restaurant’s own app or website (split evenly between app and site), citing convenience, customization, and loyalty rewards as the reasons (NCR Voyix, 2025). And in a separate HungerRush consumer survey, 67% of diners said third-party fees at least sometimes deter them from ordering (HungerRush, 2024).
The takeaway is that the barrier was never demand — it was friction. Guests didn’t order direct because your site was slow, your ordering link was buried, or checkout was clunky, so the frictionless app won by default. Remove the friction with a fast, branded, one-tap ordering flow and a majority of your guests will happily give you the order and their data. That starts with a fast site — the same reason a prebuilt restaurant website built for speed matters — and finishes with an ordering portal wired to your CRM.
What “Own Your Ordering” Custom Software Actually Looks Like
To be clear about what we mean — and what we don’t. We’re not a POS, a reservation platform, or a delivery network, and custom software doesn’t rip out the tools you already run. It’s a branded layer you own that sits in front of your existing stack and captures the order and the guest. A typical first-party build for a Phoenix restaurant includes:
- A branded online-ordering portal on your own domain — menu, modifiers, upsells, scheduled pickup and delivery, and direct card payment, so the full ticket is yours minus processing.
- First-party data capture on every checkout — name, phone, email, and order history flowing straight into your CRM, tagged and segmented automatically.
- A guest CRM and loyalty ledger built around how restaurants actually run — covers, order frequency, favorite items, VIP tiers, and points that match your concept.
- Automated marketing hooks — birthday offers, lapsed-diner win-backs, and slow-night specials that fire off the order data without anyone lifting a finger.
- A POS-to-CRM sync so orders and check-ins from Toast, Square, Clover, or your reservation tool reconcile into one guest record instead of five disconnected silos.
If the idea of unifying scattered data appeals, that’s the core of what our custom software and GoHighLevel custom development work does — build the plumbing that makes your restaurant run on one source of truth instead of a stack of apps that don’t talk.
Marketplace App vs. a Portal You Own
Here’s the honest scorecard. Marketplaces genuinely earn their keep as a discovery engine — leave them on for reach. The point isn’t to quit them; it’s to stop paying a 30% tax on the guests you already have.
Third-party marketplace vs. a custom ordering portal you own
| Plan | Your own ordering portal recommended | Third-party marketplace |
|---|---|---|
| Price | You keep the guest | 15–30% per order |
| Feature 1 | You keep the full ticket minus ~3% card processing | 15%, 25% or 30% commission on every delivery order |
| Feature 2 | Name, phone, email & order history captured on every order | You rarely get usable guest contact data |
| Feature 3 | Feeds your loyalty club, birthday & win-back automations | Can't remarket — the app owns the relationship |
| Feature 4 | Branded on your domain — your look, your voice | Your listing sits beside promoted competitors |
| Feature 5 | First-party data you can market to forever | Great for discovery of brand-new customers |
| Feature 6 | One-time build; no per-order tax as volume grows | Fees scale with every order, forever |
| Feature 7 | Best for regulars, repeat orders & anyone who found you once | Keep it on — just stop routing your regulars through it |
| Build your portal → | Fine for discovery |
The winning play isn’t either/or. Keep the apps for the first-time diner who’d never have found you, and route everyone else — the regulars, the repeat orders, the guest who came in once and loved the brisket — into the channel you own.
How We Build It (and What It Costs)
You don’t need an in-house engineering team to own your ordering. Our custom software practice builds exactly this kind of first-party system for restaurants, using Claude Code (Anthropic’s AI-assisted development tooling) to ship the same quality a traditional dev shop would — typically in about half the time. Fixed-price projects generally run $3K–$50K+ depending on scope, or $75/hour on an hourly retainer, and every fixed-price build ships with a 30-day bug-fix warranty. A branded ordering portal wired into a guest CRM is squarely in that range, and the whole thing is yours — fully owned and deployable on your own infrastructure.
The economics are simple: if a mid-sized Phoenix restaurant is paying five figures a month in marketplace commissions, an owned ordering portal that shifts even a third of that volume direct pays for itself in a single-digit number of months — and then keeps paying, because you’re trading a permanent per-order tax for a one-time build.
A Practical First-Party Rollout for a Phoenix Restaurant
Owning your ordering is a phased move, not a rip-and-replace. Here’s the sequence we use so you never lose an order in the transition.
- Measure the tax. Pull three months of DoorDash and Uber Eats statements and total the commissions. That real dollar figure — usually far bigger than operators guess — is your budget and your motivation.
- Stand up the owned portal. Launch a fast, branded ordering page on your domain with direct payment and full menu, modifiers, and scheduled pickup/delivery. Speed matters — a slow page is why guests defaulted to the app in the first place.
- Wire it to your CRM. Every checkout writes name, phone, email, and order history into your guest database, tagged and segmented, ready for marketing.
- Turn on the automations. Birthday offers, slow-night specials, and win-back sequences fire off the order data automatically — the payoff for owning the guest, and the reason your email marketing finally has a list to talk to.
- Redirect your own traffic direct. Put your ordering link everywhere you control — Google Business Profile, Instagram bio, receipts, table tents, the map pack listing — and offer a small direct-order perk. Keep the apps on for pure discovery; send everyone else home to your portal.
Run that, and the next time the Arcadia family orders $52 of enchiladas, the ticket, the margin, and the relationship all stay in your restaurant.
Frequently Asked Questions
Custom online ordering for Phoenix restaurants — FAQs
How much do third-party delivery apps actually charge Phoenix restaurants?
DoorDash and Uber Eats both use a three-tier Marketplace model: roughly 15% commission on the basic tier, 25% on the mid tier, and 30% on the top tier for delivery orders, with a lower rate (about 6%) on pickup. Higher tiers buy more visibility in the app, so the more orders you want the marketplace to send you, the larger the cut it takes. Against a typical restaurant profit margin of 3–5%, a 25–30% commission can exceed the entire profit on the order.
Isn't card processing on my own site also a fee?
Yes, but it's an order of magnitude smaller. Direct card processing runs roughly 3% of the ticket, versus 15–30% for a third-party marketplace. On a $52 order that's about $1.56 in processing instead of up to $15.60 in commission — and unlike the marketplace, your own portal also hands you the guest's contact info and order history to market to later.
Should I stop using DoorDash and Uber Eats entirely?
No. Marketplaces are a genuinely useful discovery channel for reaching brand-new customers who'd never otherwise find you, and the NRA reports that off-premise is now about 75% of restaurant traffic. The smart play is to keep the apps on for discovery but route your regulars, repeat orders, and anyone who's already visited into an ordering portal you own — so you stop paying a 30% tax on guests you already earned.
What guest data do I actually get with a custom ordering portal?
Everything the marketplace hides from you: the guest's name, phone number, email, full order history, and item preferences, captured on every checkout and written straight into your CRM. That first-party data is what powers loyalty points, birthday offers, slow-night specials, and win-back campaigns — none of which are possible when a third-party app owns the customer relationship.
How much does custom online-ordering software cost to build?
Our fixed-price custom software projects generally run $3K–$50K+ depending on scope, or $75/hour on an hourly retainer, with a 30-day bug-fix warranty on every fixed-price build. A branded ordering portal wired into a guest CRM sits in that range. Because we build with Claude Code, timelines are typically about half what a traditional dev shop quotes — and the software is yours, fully owned and deployable on your own infrastructure.
Will a custom portal replace my POS or reservation system?
No — it sits in front of the tools you already run. We're not a POS, a reservation platform, or a delivery network. A custom ordering portal is an owned layer that captures the order and the guest, then syncs into your existing stack (Toast, Square, Clover, or your reservation tool) so everything reconciles into one guest record instead of five disconnected silos.
Devon runs a small agency that builds and resells GoHighLevel systems, custom software, and blazing-fast websites to hospitality clients, from food trucks to fine dining. A former line cook turned funnel nerd, he’s obsessed with the unsexy plumbing of restaurant marketing: first-party data, POS-to-CRM pipelines, two-way SMS, and ordering flows that keep the margin where it belongs. His posts lean technical without losing the operator who has to live with the system.
Related reading
- Custom restaurant software — CRM, loyalty & ordering portals you own
- GoHighLevel custom development for restaurants
- Restaurant customer lifetime value: why repeat guests beat new ones
- Win back lapsed diners with an automated sequence
- Build a restaurant online reservation system that fills tables
Sources
- DoorDash — Merchant Pricing (Marketplace commission tiers)
- Restaurant Dive — Uber Eats increases marketplace fees
- National Restaurant Association — Elevated costs continue to pressure restaurant profitability (2025)
- National Restaurant Association — Off-premises dining now essential (2025)
- Statista — Online Food Delivery: Meal Delivery, United States (2025)
- NCR Voyix (via Restaurant Dive) — Majority of customers prefer ordering delivery directly
- HungerRush (via Restaurant Business) — How restaurants can get more customers to order direct
- Restaurant Dive — NYC passes permanent 15% delivery-fee cap
- City of Portland — Permanent commission caps on third-party delivery
- Arizona Restaurant Association — Arizona Restaurant Facts