Most restaurants think of upselling as a sales technique. It is not. The best rooms in the world generate higher average tickets than the rest, and they never make a guest feel pressured. What they do, instead, is design the entire experience (menu, service, pacing, and digital tools) so that ordering more feels like part of the meal, not an addition to it.
This guide breaks down the four levers that consistently move the average ticket: menu engineering, server scripts, digital nudges, and the operational details most restaurants overlook. None of them rely on pressure. All of them rely on hospitality.
Lever 1: menu engineering
Menu engineering is the discipline of designing the menu itself so that the dishes you most want to sell are the dishes guests most often pick. It happens on paper, on screen, and in the language you use, long before the server reaches the table.
Anchor the high-margin dishes at the top
The first three items in any category get a disproportionate share of orders. Diners scan, hesitate, then default to what they saw first. Put your best-margin starter, main, and dessert in those three slots. Not your most expensive dish, the most profitable one (those are not the same thing).
Use descriptive language
A dish labelled "Grilled salmon" sells a fraction of what "Pan-seared Atlantic salmon, lemon-butter sauce, roasted asparagus" sells. Cornell University research on descriptive menu labels documented sales lifts of up to 27% from rewriting menu copy alone. The cost is one afternoon. The return lasts as long as the menu does.
Hide price columns
Right-aligned columns of prices invite guests to scan price-first and pick the cheapest. Inline prices, in the same weight and colour as the description, invite them to read the dish and decide on taste. This is the single most impactful typographic change you can make.
Group by occasion, not by category
A dessert section helps guests who already want dessert. A "Finish the meal" section, with three desserts, two digestifs, and a coffee pairing, helps guests decide whether they want dessert at all. The framing matters as much as the items.
Lever 2: server scripts
Server upselling has a reputation problem, mostly because it is done badly. The bad version is the upsell of insistence: "Are you sure you do not want a starter?" The good version is the upsell of information: "Tonight the lamb is from the same producer who supplies the restaurant next door, but ours is cut differently." The first feels like pressure. The second feels like hospitality.
Recommendations, not questions
"What can I get you to drink?" is a yes-or-no question with a default answer of "tap water." "We have a Provence rosé that is drinking beautifully tonight, would you like to start with a glass?" is a recommendation. Guests accept recommendations from staff they trust. They refuse pressure from staff they do not.
Two-tier suggestions
For drinks, always suggest two options at different price points. "We have a house red by the glass at nine, and a Burgundy from a small producer at fourteen, both work with the duck." Guests appreciate the choice and rarely default to the cheapest. Most pick the middle, which is exactly where the margin sits.
Bridge between courses
The two highest-leverage moments for a server are right after the main is cleared, and right before the bill arrives. After the main is the moment to suggest dessert and coffee. Before the bill is the moment to suggest a final digestif. Skip these moments and the average ticket drops 15 to 20%.
Train, then trust
A well-briefed server who knows the menu, the pairings, the producers, and the kitchen's intentions sells more than one who is reading from a card. Train hard in the pre-service briefing, then trust the staff to use their judgement at the table.
Lever 3: digital nudges
A digital menu is, among other things, a quiet co-pilot for the server. It can do the suggesting without anyone having to say a word. This is where modern restaurants pull ahead.
Wine and dish pairings
Under each main, a small "Pairs well with" line that names two wines (one by the glass, one by the bottle) does the work a sommelier would do at a different price point. The guest sees the suggestion the moment they read the dish. No pressure, no script, no awkward moment with the server.
Add-on suggestions
A steak that says "Add a side of pommes anna for 5 EUR" inline gets ordered with the side roughly 40% of the time, against 8% without the prompt. The increment is small per ticket and significant over a service.
Dessert teasers in the main view
A photo of the signature dessert, shown discreetly at the bottom of the main course list, plants the idea before the meal even starts. By the time the server asks about dessert, the guest has already half-decided.
Featured items
One or two dishes marked as "Chef's pick" with a small star get ordered 2 to 3 times more than unfeatured dishes. Use this carefully: feature what you want to sell, not what the chef wants to show off.
Lever 4: operational details
This is the part most articles skip, because it is unglamorous. It is also where the largest gains hide, because no one else is paying attention to it.
Pacing
Tables that are turned too fast order less. Tables that are turned too slowly order more per cover but fewer covers per night. The right pace per concept is the pace at which guests have time to consider dessert and a digestif without feeling rushed, and to leave without feeling forgotten. This is service design, not luck.
Empty glasses
An empty wine glass for more than three minutes is a missed second bottle. The single most reliable revenue tactic in service is teaching the staff to spot the second-glass moment and offer a refill without being asked.
The bread question
Free bread is a margin killer if guests fill up on it. Bread served warm, in small portions, with intent, is part of the experience. Bread served cold, in a stack, kills appetite for the second starter. This is a small detail with a measurable effect on the average ticket.
The coffee opportunity
Most restaurants forget that coffee is one of the highest-margin items on the menu. A guest who orders coffee usually orders a small accompaniment with it (a biscuit, a chocolate, a small dessert). A guest who declines coffee leaves immediately. Always offer coffee proactively, and always offer it with something next to it.
What does not work
It is worth naming the things that look like upselling but actually erode the experience and the average ticket over time.
- Pushy suggestions after the guest has said no. The first suggestion is hospitality, the second is pressure.
- Surprise upsells at the bill (a service charge or a bread cost the guest did not see) destroy trust and tip.
- Hard sells on cheap modifiers ("Would you like to add cheese for 2 EUR?") feel mercenary and signal a tight margin to the guest.
- Long-winded specials recitations by the server. Three sentences, ideally tied to one or two dishes, is the ceiling.
Hospitality wins. Pressure loses. This is the rule under every other rule in this guide.
The order of operations
If you implement all four levers, the order matters. Do them in the wrong sequence and you will work hard for little gain.
- Start with menu engineering. It compounds every other change. Rewriting descriptions, reordering items, hiding price columns: one focused afternoon, weeks of effect.
- Then add digital nudges. A digital menu lets you test pairings, add-ons, and featured items without reprinting anything. Adjust weekly.
- Then train the team on scripts. Recommendations, two-tier suggestions, bridge moments. One pre-service briefing per week for a month.
- Finally, fix the operational details. Pacing, glass watch, bread protocol, coffee. These are habit changes and they take longer.
Restaurants that do all four well see average tickets rise 15 to 30% over six months, without guests noticing that anything is being sold to them. That is the goal: more revenue, more satisfied guests, less stress on the floor.