A restaurant doing 1,500 orders a month at a $24 average check books $36,000. Push that check to $27, a 12.5% lift, and the same 1,500 orders book $40,500. That's an extra $4,500 a month, and most of it lands on the margin line, because a drink and a side cost almost nothing to add and you're not spending a cent more on ads or couriers. Every operator knows the average check matters. Fewer can say which levers actually move it, or by how much. So here's my ranking.
1. Screens out-sell counters: kiosk and app-style upsell (+10–20% in-store)
This is the best-measured AOV effect I know of. When ordering moves off a rushed counter exchange and onto a screen, checks go up 10–20%. The reason is boring. The screen offers the large size and the extra sauce every single time, it never feels awkward doing it, and the customer isn't deciding with a queue breathing down their neck. If you run counter service, self-order screens are the single biggest AOV move you can make. It's the same reason branded ordering apps carry higher checks than phone orders: people spend more when they browse in peace than when they order under pressure.
2. The free-delivery threshold (+8–15% on delivery orders)
Nothing motivates a food customer like a $3.99 fee they could dodge. Set free delivery at roughly 20–25% above your current delivery AOV, so a $24 average check gets a $30 threshold, and watch a chunk of $24 carts climb to $31 as people add a dessert "to make it worth it." You're trading the delivery fee for a bigger basket, so do the math per order. Giving up $3.99 to gain $6–8 of mostly high-margin add-ons is a good trade. A threshold set right at your current average, where most orders already qualify, just hands the fee back for nothing. Check the result by zone in your reports after two weeks and adjust.
3. Bundles and combos (+6–12%)
A "Family Friday" bundle (two pizzas, sides, a drink, priced a little under the sum of the parts) moves the check two ways. It anchors the order around something bigger than a single dish, and it decides dinner for the whole household, which is the help the customer wanted in the first place. Build bundles from high-margin components like dough, pasta, and fountain drinks and you can lift AOV 6–12% while barely touching your margin percentage. There's a delivery bonus too: bundles travel well, and they cut per-item picking errors in the kitchen.
Thresholds, combos, and smart upsells are settings in Dots — not a development project. See them live.
4. Checkout upsell prompts (+5–10%)
This is the digital version of "want fries with that?": one relevant suggestion between cart and payment. The word doing the work is relevant. Pizza gets dips and garlic bread, sushi gets miso and extra ginger, and nobody gets a generic bestseller list. One or two suggestions convert; five feel like a gauntlet and drag down checkout completion. On an ordering site or app this is configuration, not code. Set the item-level pairings once in your ordering site and every order gets the offer a busy cashier forgets. A well-paired prompt sees 15–25% acceptance, which is how the lever nets out at +5–10% across the whole channel.
5. Menu price architecture (+3–7%)
This one is about framing. Offer three sizes, price the large so it's obviously the value pick, and the middle option starts selling itself. Add a single premium variant, a "signature" burger $4 above the standard, and two things happen: some buyers always trade up, and the premium price tag makes everything under it look reasonable. The gains here are smaller, and you need item-level sales data to confirm them, but they cost nothing and they compound with every other lever. The deeper version of the same idea is menu engineering.
What not to do: buy AOV with margin
Two traps look like AOV wins and aren't. Deep "spend $40, get 25% off" promos raise the average check while gutting the profit inside it, so check contribution per order, not revenue per order, before you celebrate (our unit-economics worksheet turns that into a five-minute job). Across-the-board price hikes lift AOV too, right up until order frequency starts to sag. The levers above grow the basket. Blunt discounting just rearranges it.
Stack the levers, but change one thing at a time
The ranges above don't simply add up. A customer who padded the cart to clear the free-delivery threshold is less likely to also take the checkout upsell. For a restaurant pulling levers 2–4 well, a realistic stacked outcome is +10–18% AOV inside a quarter, not the sum of the maximums. Roll them out two weeks apart so you can actually see which lever moved the number; a before-and-after on blended monthly AOV hides more than it shows. And keep one guardrail metric next to AOV: checkout completion. A lever that lifts the average check while dropping completed orders is a net loss wearing a nice KPI.
Do this next
Pull your delivery AOV for last month. Set a free-delivery threshold 20–25% above it, add one paired upsell to your top five sellers, and change nothing else for two weeks. Both levers are pure configuration, both touch every order, and together they give you a clean before-and-after. Then work down the list.
Upsells, bundles, thresholds, and per-channel analytics — configured on your menu in about two weeks.