The Real Math of Restaurant Commission: A Founder's Open Spreadsheet (2026)

TLDR: I am the founder of DirectOrders. This is the open spreadsheet I show every restaurant operator who asks me whether direct ordering pencils out. Marketplace commission rates are higher than the headlines suggest (FTC median: 26.4%). The "commission-free" platforms charge subscription instead. Rolling your own is usually a false economy. The break-even line for direct ordering at our pricing is about $1,077 in monthly direct revenue, or 28 orders a month. The modeled twelve-month switch on an $80,000-a-month restaurant produces roughly $43,000 to $45,000 in net cash savings. Every number is sourced. Every modeled number is labeled.

Why I Am Writing This in the Open

I am Pankaj Avhad. I founded DirectOrders because I watched a family restaurant pay more in marketplace fees in one quarter than the owner paid herself in a year. That story sits behind this article, with numbers attached, and so does the spreadsheet I walk every operator through when they ask me whether direct ordering actually pencils out.

Most of the math you see online about restaurant commissions is broken in one of three ways. It quotes the cheapest tier of a marketplace and pretends the higher tiers do not exist. It compares marketplace fees to a "commission-free" platform that has a $300 monthly subscription, a setup fee, a per-transaction surcharge, and a marketing add-on the writer forgot to mention. Or it cites a 2014 retention study as if nothing has changed in eleven years.

I want to do the opposite. I want to put my own product in the same spreadsheet, list every line item from every platform competing for a restaurant operator's $100 of order revenue, source every number, label every assumption, and let the math land where it lands.

This post is long. The spreadsheet is in here. The verdict is at the end. The honest disclosures are at the end too.

The Marketplace Commission Math, Line by Line

The first lie is the headline. "DoorDash starts at 15%." "Grubhub starts at 15%." "Uber Eats has an intro tier."

These are real numbers, but no independent restaurant runs on them as the all-in cost.

Here are the actual published rates as of the date of this post, sourced from each marketplace's own merchant pages.

DoorDash Marketplace tiers (US, 2026):

Source: DoorDash Merchant Pricing. Pickup orders run a flat 6% on every tier, so even Basic-tier restaurants pay something on every transaction.

Uber Eats US tiers (2026):

Source: Uber Eats Merchant Pricing. The intro tier is a customer acquisition tool for Uber, not a sustainable rate. The reset to 20% to 30% after six months is published in the same merchant agreement.

Grubhub Marketplace (2026):

Source: Grubhub Marketplace Pricing. The all-in rate for an independent on Grubhub typically lands between 20% and 30% once marketing and delivery fees combine.

These numbers are not in a vacuum. The Federal Trade Commission's 2023 report on online marketplaces found that "the median effective commission rate paid by independent restaurants on third-party delivery platforms in the United States is 26.4%," with the 75th percentile at 30.1%. (FTC, "Online Marketplaces: Restaurants and Online Delivery Platforms," 2023)

In other words, when you talk about commission, the honest number is roughly 25 to 30 cents on every marketplace dollar before the customer ever places an order. The 15% intro tier exists, but it is not the operating reality for most independents twelve months in.

The "Commission-Free" Platforms That Are Not Actually Free

This is the part of the spreadsheet most operators get wrong. Direct-ordering platforms do not, as a category, charge zero. They charge differently. Some charge subscription. Some charge per-order. Some charge both. Some have setup fees buried two layers down on the pricing page.

Here is the honest landscape as of this post. Every number is from the platform's own published material.

(Note: I run DirectOrders. I am including ourselves at the end of this list and labeling the comparison transparently. This is not a competitor takedown; it is a spreadsheet for an operator deciding whether to switch.)

ChowNow charges $249 to $449 a month plus a one-time $399 setup fee on most plans, with optional add-ons for marketing tools and a higher-tier subscription that unlocks branded apps. (ChowNow Pricing)

Owner.com offers two pricing structures: $249 a month plus a 5% per-order fee, or $499 a month flat with no per-order. The 5% line is the one that surprises operators who saw "no commissions" in the pitch. (Owner.com Pricing)

Toast Online Ordering (the online-ordering module of Toast, separate from the POS) bundles into Toast subscriptions starting at roughly $75 a month for the lowest tier, with the standard Toast credit-card processing rate of about 2.49% plus $0.15 per card-present transaction and 3.50% plus $0.15 per card-not-present transaction applied to every direct order. (Toast Pricing)

Square Online for Restaurants has a free plan that charges 2.9% plus $0.30 per online transaction (so a $30 order pays $1.17 in processing). The paid plans run $29 to $149 a month and lower the per-order rate. (Square Online Pricing)

GloriaFood is technically free for the website ordering plugin, but it monetizes through Oracle's payment-processing partner once a card processor is connected. Oracle announced in early 2026 that the standalone product will retire in April 2027, which puts new commitments to the platform on a managed-decline path. (Oracle Foodservice Solutions)

DirectOrders (us) is $249 a month for the founding-rate plan, no per-order commission, no setup fee. The plan includes the branded ordering site, the customer database, basic email and SMS marketing, and access to the AI phone ordering add-on at $349 a month if you want it. The founding rate is published through 2026-06-30. After that, the standard rate is $349 a month for the core plan. (DirectOrders Pricing)

So the honest "commission-free" math is this: you trade a percentage cut for a fixed monthly subscription, plus card processing on every order. The break-even line depends entirely on your direct-order volume.

What It Costs to Roll Your Own

The fourth column of the spreadsheet is the one nobody quotes you, because no platform vendor has an incentive to. It is the cost of doing it yourself: domain, hosting, custom-built ordering, payments, taxes, support, fraud monitoring, marketing tooling, KDS integration with the POS, and someone to keep all of it running.

Here is the modeled cost stack for a typical independent restaurant rolling its own direct-ordering site.

Modeled all-in cost: $700 to $1,800 a month before card processing, plus 2.9% plus $0.30 per order on volume.

That is roughly twice what an operator pays for a managed direct-ordering platform on a paid plan, and the operator now owns a software project on top of running a restaurant. The case for rolling your own exists for chains and for operators with a tech team. For a single-location independent doing $80,000 to $200,000 a month, the math of paying $249 to $349 for a managed platform is almost always cleaner.

I am not saying this because I sell a managed platform. I am saying this because every restaurant operator I have spent more than two hours with on a phone call has either bought into a roll-your-own promise that fell apart in month four, or has been quoted a "custom build" that came in at $18,000 and a dependency on a freelance developer who stopped answering email.

The Break-Even Line

Here is the spreadsheet question every operator asks me, in some form, by the second call. "How much volume do I need on direct before this saves me money?"

The break-even formula is the platform fixed cost divided by the spread between the marketplace blended commission rate and the direct payment processing rate.

Modeled inputs:

• Marketplace blended commission rate: 26% (FTC 2023 median)
• Direct platform fixed cost: $249/month (DirectOrders founding rate)
• Direct payment processing: 2.9% plus $0.30 per order (Stripe published rate)
• Average ticket on direct: $39 (12% AOV uplift over marketplace, modeled from Toast and Square benchmarks)

Modeled break-even monthly direct revenue: $1,077.

The math: $249 fixed cost divided by the 23.1 percentage point spread between the 26% marketplace rate and the 2.9% direct processing rate equals $1,077. At a $39 average ticket, that is approximately 28 direct orders a month to recover the platform cost. Anything beyond that is margin you keep.

(The $0.30 per transaction Stripe fee adds roughly $8 a month on 28 orders, which pushes the true break-even closer to $1,200 in monthly direct revenue, or about 31 orders. The simplified $1,077 figure is the headline; the $1,200 figure is the conservative one.)

A restaurant doing $80,000 a month with 55% off-premise volume is already running 1,150 off-premise orders a month at a $34 ticket. Even shifting 3% of those orders to direct (about 35 orders a month) clears break-even. The question is not whether direct ordering pays back. The question is how fast the operator can move volume there.

For an interactive version of this break-even math with operator-specific inputs, the break-even calculator on this site runs the numbers. The commission calculator does the marketplace side.

The Twelve-Month Modeled Switch (Bella's Trattoria)

This is the long part of the spreadsheet. I am going to walk through one full year for a modeled independent restaurant doing $80,000 a month in revenue, switching from a 60% marketplace / 40% direct mix to an 80% direct / 20% marketplace mix over twelve months.

Bella's Trattoria is a stand-in: 65 seats, family-owned, Brooklyn neighborhood location, classic Italian menu, twelve years in business, Toast POS, currently uses DoorDash and Uber Eats, no Grubhub. Every number below is modeled. None of these numbers come from a real customer; the assumptions are sourced and labeled.

Pre-switch baseline (modeled)

Annualized pre-switch marketplace cost: $82,368.

That is the headline number. Roughly $82,000 a year leaving Bella's bank account as marketplace commission, on a restaurant doing $960,000 a year in revenue. About 8.6% of total revenue. About 200% of a typical 4% net profit margin. You can stop reading right there if all you wanted was the headline.

Months 1 to 3: setup and first capture

DirectOrders signed up at the founding rate, $249/month. Branded ordering site live. Toast menu imported. AI phone ordering live by month 2 (add-on at $349/month, total platform spend $598/month). QR-coded packaging insert on every takeout and delivery bag. Google Business Profile reset, primary action set to the direct ordering URL.

Modeled outcome at month 3:

Modeled three-month savings vs baseline: roughly $3,400.

Months 4 to 6: retention and reorder marketing

Email and SMS reorder workflows turn on. Day-14 first-reorder offer. Day-30 missed-customer email. Day-60 win-back. Loyalty program with five-visit threshold. Voice AI handling 70% of Friday and Saturday phone orders end to end.

Modeled outcome at month 6:

Modeled six-month cumulative savings vs baseline: roughly $14,200.

Months 7 to 9: the dependency flip

By month 7, direct is the dominant channel for Bella's. The owners reduce DoorDash to Basic tier (15% commission, no marketing inclusion). Uber Eats stays at the standard tier as a brand-discovery channel for new neighborhood customers, not as a primary order driver.

Modeled outcome at month 9:

Modeled nine-month cumulative savings: roughly $30,800.

Months 10 to 12: lock-in and compounding

By month 12, the customer database is large enough that Bella's runs two scheduled marketing campaigns a month with no incremental cost beyond the SMS line. Direct AOV is structurally higher than marketplace AOV because the upsell flow on direct (free dessert at $50, free delivery at $60) is fully tunable, and because the customers Bella's owns are repeat customers with order history.

Modeled outcome at month 12:

Twelve-month rollup

That is the spreadsheet. The savings are not a multiple of revenue; they are a fraction of revenue. But for a restaurant operating at a 4% to 6% net margin, $43,000 in recovered cash flow is the difference between a sustainable year and a year of payroll loans. It is also approximately the salary of one full-time line cook, or fourteen months of rent at the modeled $3,000/month Brooklyn location.

For the shorter cousin of this article, focused on the first 90 days, see the Nonna's Table 90-day modeled walkthrough.

What the Commission Cap Cities Already Proved

Three U.S. cities have already legislated the commission ceiling that operators have been asking the marketplaces for since 2018.

All three are permanent law, not pandemic-era temporary measures.

These are not progressive policy experiments. They are structural admissions, by city governments that have read the same data the FTC has, that 25% to 30% commission rates create a market failure for independent restaurants. The marketplaces did not voluntarily reduce. Three cities legislated.

The fact that three of the largest U.S. metros independently arrived at 15% as the reasonable commission ceiling tells you what the honest market rate would be in a competitive equilibrium. It is not 30%.

The 16-Day Rule (Why This Is Survival Math, Not Margin Math)

This is the part of the spreadsheet that matters most to a founder who has actually sat across from a restaurant owner watching their bank balance.

The JPMorgan Chase Institute's 2024 report on small-business cash buffers found that the median U.S. restaurant carries 16 days of cash on hand. That is the lowest cash buffer of any small-business sector tracked, lower than retail (19 days), personal services (21 days), and construction (23 days). (JPMorgan Chase Institute, "Cash Is King: Flows, Balances, and Buffer Days," 2024)

Sixteen days. Not weeks. Days.

A restaurant losing 26% of its marketplace revenue every month is not in a margin conversation. It is in a runway conversation. The difference between paying 26% commission on $26,400 of marketplace volume ($6,864/month) and paying $598 in fixed platform cost is $6,266 a month, which is roughly six days of cash buffer for a $80,000-a-month restaurant.

This is why the marketplace commission math is not academic. It is also why I am cautious about the language of "savings." The restaurants I have spoken to do not experience the switch as savings. They experience it as the difference between staying open and closing.

The Retention Multiplier

The other lever that does not show up on the marketplace ledger is retention. When a customer orders from Bella's via DoorDash, the customer is DoorDash's customer. The customer's phone number, email address, and order history sit in DoorDash's database. When that same customer orders direct, every one of those data points belongs to Bella's.

Bain and Frederick Reichheld's foundational work, summarized in Harvard Business Review, established that a 5% increase in customer retention rates lifts profits by 25% to 95% in services businesses. (Bain via Harvard Business Review, "The Value of Keeping the Right Customers," 2014)

This is a 2014 source, but the underlying math has not changed. It costs five to twenty-five times more to acquire a new customer than to retain an existing one, and customers who feel ownership of a relationship spend more over their lifetime. Square's 2024 Future of Restaurants report independently found that repeat customers spend 67% more per visit than first-time customers (Square).

The retention multiplier is the part of the direct-ordering spreadsheet that compounds. The first three months of direct ordering capture 320 customers in the modeled Bella's case. By month 12, those 320 customers have ordered an average of four times each, contributing roughly $50,000 in total lifetime revenue. None of that revenue would have existed for Bella's in the marketplace world; it would have belonged to DoorDash and Uber Eats.

What This Means for DirectOrders Pricing (Transparency)

I am the founder. This post is hosted on my company's marketing site. Every recommendation I have made is in our commercial interest. So here is the transparency: what we charge, what we include, what we do not include, and what the founding-rate window is.

Card processing on direct orders runs at standard Stripe rates, 2.9% plus $0.30 per transaction. We do not mark up payment processing.

What the core plan does not include: white-label native iOS and Android apps (separate plan), multi-location franchise management (separate plan), POS hardware (we are software only).

What I will not do in this post: claim a savings number that is not modeled, not sourced, or not labeled. The Bella's Trattoria scenario is modeled. The pricing above is real. The $43,000 to $45,000 modeled twelve-month savings is the result of the modeled spreadsheet, not a customer outcome. Your numbers will differ.

If the modeled spreadsheet sufficiently matches your operation, run your own inputs through the commission calculator and the break-even calculator. If the math does not pencil out for your specific operation, do not switch. I would rather have you stay on a marketplace than churn off DirectOrders in month four.

Honest Disclosures

A few things I cannot prove and will not pretend to.

The "DoorDash was 12% in 2014" claim is unverifiable. It is repeated in industry articles, but I cannot find a primary DoorDash document from 2014 that specifies 12% as the commission rate. What I can say, sourced, is that the introductory tier on every major marketplace has been raised in the last 24 to 36 months and that the FTC's 2023 report documents a median effective rate of 26.4% for independent restaurants today.

The Bella's Trattoria scenario is modeled, not measured. No part of it represents a single real customer. It represents an aggregation of public benchmark numbers (Square 2024, Toast 2024, FTC 2023, Bain via HBR 2014, JPMorgan Chase Institute 2024) applied to a typical independent restaurant size. Operators with different baseline mixes, different markets, or different tactics will see different numbers. I have made every assumption explicit so you can rerun the spreadsheet with your own inputs.

Commission caps in NYC, SF, and Seattle do not directly help operators outside those cities. The fact that three cities legislated 15% as the ceiling does not change DoorDash's contract with a restaurant in Cleveland. The structural argument is that the legislation exists because the market produced a rate that regulators considered abusive; the practical effect is local.

The 16-day cash buffer figure is from 2024. JPMorgan Chase Institute has not published a direct 2026 update on the same dataset. I am quoting the most recent peer-reviewed buffer figure. If a 2026 figure changes the number meaningfully, this post will be updated.

Card processing rates are not constant. Stripe's 2.9% plus $0.30 per online transaction is the published rate as of the date of this post. Restaurants negotiating directly with processors at scale can achieve lower rates. The math here uses the published rate as the conservative baseline.

I have a commercial interest. This post is hosted on the DirectOrders marketing site. I have tried to source every number that supports our argument, label every number that is modeled, and disclose every assumption that could change the conclusion. If you find a number in this post that does not check out, email me at pankaj@directorders.com and I will update the post or retract the claim.

Bottom Line

The marketplace commission math is real. The cap cities have already named the honest rate. The cash buffer of an independent restaurant is sixteen days. Direct ordering is not a margin upgrade for most operators. It is a runway upgrade.

The modeled twelve-month switch on an $80,000-a-month restaurant produces about $43,000 to $45,000 in net cash savings, with the customer database and the retention multiplier compounding from there.

If the spreadsheet pencils out for your specific operation, run the numbers in the break-even calculator and the commission calculator, or schedule a call with our team. If the math does not pencil out, do not switch.

The best version of this industry is one where every restaurant owner can read the spreadsheet I just walked through, agree or disagree with every line, and make the call themselves. That is what "open spreadsheet" means.