How to Increase Average Order Value at a Restaurant (2026)
Average order value formula, where to find it in your POS, published benchmarks, and the upselling tactics that lift the check, plus a 30-day test plan.
Updated Sep 21, 2026
+0%
Revenue Growth
TLDR
Average order value is net sales divided by order count, tracked separately for dine-in, phone, website, and delivery apps, because online checks run higher than counter orders. Published benchmarks sit in the low to mid $30s for digital orders (Paytronix first-party web AOV of $35.97 for non-loyalty guests; DoorDash Q2 2026 at about $34 of GOV per order including taxes, tips, and fees), but your own trailing 90 days is the benchmark that counts. The levers that move it most are add-on prompts at checkout, combos and bundles, a free delivery threshold set just above today's average, and one scripted upsell per order. At 3,000 orders a month, a $2 lift is $72,000 a year in revenue and a $4 lift is $144,000. Pick two tactics, test for 30 days, and measure by channel.
TLDR
Average order value (AOV) is net sales divided by order count, tracked per channel, because online checks run higher than counter orders. Published benchmarks sit in the low to mid $30s for digital orders (Paytronix first-party web AOV $35.97; DoorDash Q2 2026 about $34 of GOV per order including taxes, tips, and fees), but your own trailing 90 days is the benchmark that counts. The levers that move it most: add-on prompts at checkout, combos and bundles, a free delivery threshold, and one scripted upsell per order. At 3,000 orders a month, a $2 lift is $72,000 a year and a $4 lift is $144,000.
To increase average order value, add a specific add-on prompt at checkout, sell combos and bundles, set a free delivery threshold above today's average, and give staff one scripted upsell per order.
This is for an independent owner who wants a bigger check from the customers already ordering. More customers is the direct orders playbook; more visits from the same people is customer retention; which dishes to push is menu engineering. This page covers the formula, the benchmarks, the tactics by channel, a staff script, a worked example, and a 30-day test.
What is average order value and how do you calculate it?
Average order value is net sales divided by the number of orders in the same period: $84,000 in net sales across 3,000 orders is a $28 average order value. Use net sales after discounts and refunds, before sales tax, tips, and delivery fees, and calculate it separately for each channel.
What to include. Count orders, not customers, so a table of four on one check is one order. Strip tax, tips, and pass-through delivery fees; none of that is money you keep. If you mark up your marketplace menu, pick one basis (what the customer paid, or what you keep after commission) and stay consistent.
Where to find it. Toast Web: Reports, then the Benchmarking drop-down; the overview cards show net sales, total orders, and average check size. Square Dashboard: Reports, Sales, Sales trends, metric "average total sale"; the Sales summary report filters by channel and source. Any other POS: export the orders report, sum net sales, count rows, divide.
Why split it by channel. A blended number hides where the money is. Paytronix's April 2024 report found guests ordering from a brand's own site or app put 35% more items on the check and spent 30% more per transaction than guests on third-party marketplaces. Each channel has its own baseline and its own fix.
| Channel | Orders per month | Net sales | Average order value |
|---|---|---|---|
| Dine-in | 1,200 | $32,400 | $27.00 |
| Phone | 300 | $7,800 | $26.00 |
| Own website | 900 | $27,900 | $31.00 |
| Delivery apps | 600 | $15,900 | $26.50 |
| All channels | 3,000 | $84,000 | $28.00 |
Example numbers. The blended $28 looks fine, but phone and app orders are dragging. Test there first.
What is a typical restaurant average order value?
There is no single published benchmark for restaurant average order value; the figures that exist land in the low to mid $30s for digital orders and measure different things, so the only benchmark worth managing to is your own trailing 90 days, split by channel.
| Source | Figure | What it measures |
|---|---|---|
| Paytronix report, October 2025 | $35.97 AOV for non-loyalty guests, $34.79 for loyalty members | First-party web orders across the brands on the Paytronix platform, 12 months of data |
| DoorDash Q2 2026 results, August 2026 | About $34 per order ($33.1 billion Marketplace GOV, 970 million orders) | Includes taxes, tips, consumer and membership fees; mixes restaurant, grocery, and international orders |
| Toast press release, August 2016 | Online checks 23% larger than in-store | Toast online ordering clients, a decade-old snapshot |
| Paytronix report, April 2024 | First-party orders carry 35% more items and 30% more spend than third-party | 2023 data, 1,800+ restaurant and convenience-store brands |
Read those as direction, not targets. A taqueria doing $14 bowls and a steakhouse doing $95 covers will never share a benchmark, and vendor figures skew toward multi-unit brands with apps and loyalty programs. Every tactic below is measured against your own trailing 90 days.
Menu design: anchoring, combos, bundles, and sizes
Menu design changes the check without anyone asking for more money. Which items deserve the spotlight is the job of menu engineering; this is how the menu is built around them.
Anchoring. Put one clearly premium item at the top of each category: the $34 ribeye above the $22 chicken. The anchor does not need to sell; it makes the item under it look reasonable. Choice research has documented the related decoy effect since 1982 (Huber, Payne, and Puto), but nobody publishes a clean restaurant-wide effect size for anchoring, so treat it as a cheap test. On price format, a Cornell study at one upscale casual restaurant found guests spent $5.55 more, about 8%, when the menu dropped the dollar sign.
Combos. A combo turns a $12 entree into a $16 order by default. Price it below the a la carte sum and above the entree alone, and judge the margin in dollars, not percent: $4 of price for $1.20 of food cost is a win even if the food cost percentage ticks up.
Bundles for groups. Name them by headcount: Dinner for Two, Family Meal for Four. DoorDash's own menu guidance tells merchants to build "dinner for two" and "a full family-style meal for 6" bundles because they are faster to order and easier on the kitchen. A bundle is the biggest single jump available: one $28 order becomes one $50-plus order.
Sizes. Offer three sizes and price the large so the step up from the medium is small relative to the extra portion. The medium is where you want most orders to land; the large makes it look sensible and catches the hungry.
Online checkout: add-on prompts, complete the meal, delivery thresholds, and fee framing
Online is where you control every step, so it is where the check moves fastest. Four levers, in order of payoff.
Add-on prompts (modifiers). When someone adds an entree, offer the side, the drink, and the dessert on the same screen, each priced, one tap. Paytronix's October 2025 report measured recommendation acceptance of about 3.6% on mobile and 4.1% on web, and says suggested add-ons and "rolled-by" category prompts can lift digital ticket sizes by 14% to 18%. The acceptance rate is for one recommendation slot; the lift is the vendor's claim for the whole add-on layer. Plan on the small number. DoorDash tells merchants the same about modifiers, calling sides, toppings, extra protein, and drinks "an opportunity to increase ticket size"; french fries were the number one item ordered on DoorDash in the US in its 2025 trends report. People skip the fries because nobody asked.
Complete the meal. One prompt after the entree, three suggestions maximum, priced. "Add a drink for $3?" beats a page of 40 add-ons. Fit the suggestion to the item: chips and salsa with tacos, garlic knots with pizza.
A free delivery threshold. Set it a few dollars above your website AOV and show it on the menu page, not at payment. DashPass and Uber One both gate $0 delivery behind a per-merchant minimum basket shown on the storefront, so customers already know the deal: add one item, skip the fee. If your website AOV is $31, "free delivery over $35" turns some $31 orders into $36 orders without touching a price.
Tip and fee framing. Surprise costs at the last step kill orders: in the Baymard Institute's checkout research, the top reason shoppers abandon is extra costs (shipping, tax, fees), cited by 40%. That is e-commerce data, but the behavior carries over. Show the delivery fee and any minimum before the cart, skip the "service fee" line that only appears at payment, and present tips as percentage buttons with a sensible default.
What should staff say to upsell without annoying guests?
Staff upselling works when the server or phone-taker names one specific item at a specific price at the moment it fits, and stops after one no; "Anything else?" is not an upsell, it is a yes/no question that gets a no.
| Situation | What to say | What not to say |
|---|---|---|
| Taking the drink order | "Can I start you with a fresh lemonade or one of the draft beers?" | "Just water?" |
| Guest orders an entree | "Want to add the garlic fries to that? They are $4 and they go with it." | "Anything else?" |
| Phone order for one | "That comes with rice. Want a side of plantains for $3?" | "Is that everything?" |
| Guest picks a size | "Make it a large for $2 more? It is almost half again as much." | "Small or large?" |
| Plates cleared | "We have two desserts tonight, the flan and the churros. Want one to share?" | "Any dessert?" |
| Guest says no | "Sounds good." Move on. | A second pitch. |
Two rules make it measurable: one suggestion per order, chosen in advance for each menu item, so the script is the same at 12:15 and at 8:45; and track the attach rate (orders with a side or drink divided by all orders) per server or shift from your POS item sales by employee. There is no reliable public number for what scripts add, because it depends on whether they are used, so measure your own.
Delivery-specific levers: minimums and bundles for two
Delivery orders are the most expensive orders you fulfill, so small ones hurt twice.
A minimum order for delivery. The marketplaces already do this: DoorDash and Uber Eats both charge customers a small order fee when the subtotal falls below a minimum. On your own site, set a delivery minimum above your smallest profitable order (work it out from your driver's cost per drop, or your Uber Direct or DoorDash Drive fee) and leave pickup with no minimum. You lose a few tiny orders and stop paying a driver to carry a $9 burrito.
Bundle for two. A Dinner for Two (two entrees, a shared side, two drinks) turns two possible $26 orders into one $50 order and one trip. Price it a couple of dollars under the sum of its parts and photograph it: DoorDash's data says items with photos see up to 44% more monthly sales and items with descriptions up to 18% more.
On the apps. You do not control the marketplace checkout, so the only levers there are menu structure: bundles, modifiers, photos, descriptions. Every dollar you add pays commission first; a $4 lift on a 25% plan is $3 in your pocket. Run your channel mix through the commission calculator.
Which tactics move average order value the most?
Add-on prompts at online checkout and combos or bundles move average order value the most for the least effort, staff scripts and a free delivery threshold sit in the middle, and anchoring, price formatting, and size ladders are small, cheap, and worth running anyway.
| Tactic | Expected effect | Effort | Where it applies |
|---|---|---|---|
| Add-on prompts and modifiers | Medium to large: Paytronix measured 3.6% to 4.1% acceptance per slot and claims 14% to 18% overall | Low: an afternoon of menu setup | Online, phone (if staff or Voice AI ask) |
| Combos and group bundles | Medium for single-diner combos; large for Dinner for Two and family bundles | Low to medium: pricing and photos | All channels |
| Threshold for free delivery | Medium on delivery orders just under the line; zero on pickup | Low | Online delivery |
| Staff script, one upsell per order | Medium, only as consistent as the training; no clean public number | Medium: training plus a weekly attach-rate review | Dine-in, phone |
| Delivery minimum | Small on AOV; removes money-losing orders, may cost a few | Low | Online delivery |
| Anchoring, decoy sizes, price format | Small: decoy effect documented (Huber, Payne, and Puto, 1982); Cornell found $5.55 (about 8%) without dollar signs | Low | Menus in every channel |
| Photos and descriptions | Indirect: DoorDash reports up to 44% more monthly sales with photos and up to 18% with descriptions | Medium: photograph the whole menu once | Online, delivery apps |
Worked example: what a $2 or $4 lift is worth
Example, round numbers: 3,000 orders a month at a $28 average order value, or $84,000 in monthly net sales.
| Line | Today | $2 lift ($30 AOV) | $4 lift ($32 AOV) |
|---|---|---|---|
| Monthly net sales | $84,000 | $90,000 | $96,000 |
| Extra revenue per month | $0 | $6,000 | $12,000 |
| Extra revenue per year | $0 | $72,000 | $144,000 |
| Extra gross margin per year at 30% food cost on the add-ons | $0 | $50,400 | $100,800 |
Same kitchen, same customers, same order count. The $2 lift is one $6 side or drink accepted on every third order; the $4 lift is the same attach rate on a $12 Dinner for Two upgrade. Two adjustments: add-ons carry food cost, so at 30% the $72,000 is about $50,400 of gross margin, and card processing takes roughly 3% of the increment. And if 600 of the orders come through a marketplace on a 25% plan, the lift on those pays $0.50 of every $2 to the platform, about $300 a month.
A 30-day average order value test plan
Pick two tactics, one online and one in-person or phone, and change nothing else for 30 days.
| When | What to do | What to write down |
|---|---|---|
| Day 0 | Pull the trailing 90 days from your POS: orders, net sales, AOV, items per order, by channel | Baseline AOV per channel |
| Day 0 | Pick one online tactic (add-on prompt or delivery threshold) and one in-person tactic (the script for one item) | Exact wording, price, or threshold |
| Days 1 to 3 | Build the modifiers or threshold, print the script, brief every shift; no price changes or promotions during the test | Launch date |
| Days 4 to 30 | Run it. Every Monday, pull AOV, order count, and items per order per channel | Four weekly rows per channel |
| Day 30 | Compare each channel to its own baseline | Keep, tweak, or drop |
Decision rule: keep a tactic if channel AOV is up at least $1 and order count is within 5% of baseline; tweak it if AOV is up but orders fell more than that (the threshold is too high, or the prompt is in the way); drop it if AOV moved less than $1. Then pick the next two.
Where DirectOrders fits
Most of the online levers above need a checkout you control. DirectOrders gives independent restaurants commission-free online ordering on their own branded website, with modifier groups so every entree carries its add-on prompts, and the $2 you add to a check is $2 you keep rather than $1.50 after a 25% commission. Voice AI phone ordering can ask the add-on question on every call. Plans are a flat $249 a month (Pro) or $349 a month (Pro + Voice, 500 AI phone minutes included) with a 14-day free trial, and you own the customer data; details on the online ordering feature page.
Sources
All checked September 2026.
- Paytronix, 2025 Online Ordering and Catering Report, press release, October 2025
- Paytronix, 2024 Online Ordering Report, press release, April 2024
- DoorDash, second quarter 2026 financial results, August 2026
- DoorDash Merchant Learning Center, strategies for increased order size
- DoorDash Merchant Learning Center, 11 tips to design a high performing menu
- DoorDash Help, what fees do I pay
- DoorDash Help, what is DashPass
- Uber Help, what fees may apply to my order
- Uber Help, changes to Uber One
- Cornell Chronicle, diners spend more when menus don't use dollar signs, December 2009
- Huber, Payne, and Puto, Journal of Consumer Research, 1982
- Baymard Institute, cart abandonment rate statistics
- Toast Central, Toast Benchmarking overview
- Square Support, sales summary and sales trends reports
- Toast press release via Newswire, August 2016
Bottom line
Average order value is net sales divided by orders, and it only becomes useful once you split it by channel. Published figures sit in the low to mid $30s for digital orders, but your own trailing 90 days is the benchmark that counts. The best return for the least effort comes from add-on prompts at checkout, combos and bundles, a free delivery threshold just above today's average, and one scripted upsell per order. At 3,000 orders a month every $1 of lift is $36,000 a year, so run two tactics for 30 days and keep what moves the check by a dollar.
Frequently Asked Questions
Same math, different labels. Toast calls it average check size, Square calls it average total sale, and online ordering platforms call it average order value. All three divide sales by the number of orders or checks in a period. The only thing to watch is whether a report uses gross or net sales and whether it includes tax and tips, so read the definition once and then compare like with like.
Related resources
Related Articles
Food used ÷ food sales
$15,300 ÷ $46,000
= 33.3%
Restaurant Food Cost Percentage: Formula & 2026 Benchmarks
What is a good food cost percentage for a restaurant? The formula, how to cost a menu item step by step, and 2026 benchmarks by type, from pizza to steakhouse.
Pankaj Avhad
3 to 5%
typical net margin
Delivery apps take 15 to 30%
Restaurant Profit Margin Benchmarks (2026 Data Study)
The 2026 benchmark study: net margin, prime cost, food and labor cost, and revenue per seat by segment, plus how a 30% commission collapses a thin margin toward zero.
Pankaj Avhad
"I'd like to order a large pepperoni pizza"
"Perfect! That's $18.99. Pickup in 20 min?"
How to Take Phone Orders at a Restaurant: Script and Checklist
A restaurant phone order script, a training checklist for new staff, the safe way to take card payments by phone, and how to count the calls you miss.
Pankaj Avhad