Suggestive selling restaurant

Suggestive selling in a restaurant — waiter, QR menu, and measurement without the hard sell

A guest orders a burger. The waiter asks “anything to drink?”. The guest says “no, thank you”. Dessert never enters the conversation, and you do not know whether that was budget, rush, or simply a bad question. On the QR menu at the table there is often no suggestion at all — or five pop up at once and look like an ad.

Suggestive selling means offering a specific add-on, drink, or combo at the moment the guest is already choosing a dish — so it is easier to add something alongside, not instead. It is not about raising prices across the whole menu. It is about giving the guest who still has room in their budget one clear offer: from the waiter’s mouth or in the cart on their phone.

Below: how to do this at the table and in the digital channel with the same rule, how to measure whether the guest actually adds the suggested item, and when to consciously not suggest.

Average check without raising the whole menu — what this is about

Average check grows in two ways: either you raise prices, or the guest takes more items from the same visit. The second way is safer when guests already watch their budget.

According to the MADE FOR Restaurant 2025/2026 report (ARC Rynek i Opinia, N=1011), the declared budget averages PLN 65 for dinner and PLN 52.20 for lunch per person. These are guest declarations, not the average from your venue’s receipts — your numbers will differ. From the same study: 55% of respondents skip starters and alcohol, 54% skip dessert to pay less. At the same time 45% want wine or add-on suggestions on the food menu, and 50% — food pairing on the drinks menu.

Read it this way: part of the floor cuts add-ons by default. Part is open to an offer if it is specific and not pushy. Suggestive selling works on the second group. Raising the whole menu squeezes both.

Menu engineering helps decide which add-ons are worth offering. This piece is about something else: how to say it at the table, how to show it in the QR menu, and how to check whether the suggestion works.

Suggestive selling, upselling, and cross-selling — a simple distinction

On the floor these three ideas often blur. One distinction is enough:

In day-to-day venue practice you mostly work on cross-selling and combos: drink, sauce, sharing starter, dessert, coffee. Upselling makes sense when the guest is torn between two variants of the same dish — not when they already said “no” to everything.

One starter rule: one offer at a time. Three questions in a row sound like a sales list, not service.

Why “anything to drink?” does not work

A closed question gives the guest the easiest answer: “no”. Not because they do not want a drink — often because they have to invent what to take themselves, or they dread another decision at a full table.

Open questions or a ready pair from the menu work better. Examples (illustration, not a script to memorize):

• instead of “anything to drink?” → “With the burger we have citrus lemonade or a small craft beer — which do you prefer?”
• instead of “dessert?” → “For dessert we have baked cheesecake or coffee with the house cake of the day — shall we save room?”
• instead of “sauce?” → “With fries people usually take garlic or BBQ — add one?”

The difference: the guest chooses between two concrete options or declines on purpose. They do not have to design the order from scratch. The waiter is not “pushing” — they offer a choice from the menu the venue already has.

If the waiter worries about being pushy, set a team limit: at most one offer per visit stage (drink at the start, add-on with the main, dessert or coffee at the end). The rest is silence and good service pace.

Three moments at the table (drink / add-on with the main / dessert or coffee)

Not every visit can take three offers. Pick the moments that fit your venue’s rhythm.

1. Start — drink
When the guest sits down and does not have a plate yet. A natural moment for water, lemonade, beer, a glass of wine. Office lunch: often one clear non-alcoholic option is enough. Dinner: a glass or a shareable drink if the table is larger.

2. With the main — add-on
Sauce, sharing starter, grill side, bread. A pair written on the menu or held in the waiter’s head works here: burger → sauce; pizza → olives / dip; steak → pepper sauce. If the guest already added three sides on their own, you do not pile on a fourth “because the rule says so”.

3. After the main — dessert or coffee
Not during the last bite of the main and not with the check in hand. A short pause, plates cleared, one sentence. If the guest glances at their watch or talks about the bill — let it go.

For a “quick” table, moment 1 is enough. For a celebratory dinner — 1 and 3. Moment 2 makes sense when the menu has sensible “alongside” add-ons, not only mains.

The same offer in the QR menu and online

The most common mistake: the waiter suggests lemonade with the burger, and on the QR menu for that same burger a random dessert pops up — or nothing. The guest gets two worlds. So does the team.

Set the rule once: dish → specific add-on (or combo). Then that same pair goes into the waiter’s mouth and into the digital cart. It does not have to sound identical word for word — it must be the same logic.

What the guest sees in the cart / modal

In the QR menu and in online orders the guest adds a dish to the cart. Then they may see a clear modal: a matching drink, sauce, or a combo offer with an upcharge. They can add it or close it. No pressure, no five banners at once.

In OrderNow such add-on suggestions (Smart Pairings) and slotted combos (Combos) are set by the manager in Sales Engine — the system does not pick pairs “on its own”. You decide: burger → BBQ sauce / lemonade; pizza → garlic dip; burger + fries + drink combo with a clear upcharge.

What the waiter sees in the panel

The same rule can appear in the waiter panel when taking the order at the table. The waiter does not have to remember every pair from memory on a Friday rush — they see a prompt next to the item they are entering.

Effect for the guest: consistency. Effect for the manager: one list of rules to train and improve, instead of a separate “floor script” and a separate “something in QR”.

How to measure whether the suggestion works

Without measurement you are left with a feeling: “seems better” or “waiters do not want to”. The numbers do not need to be an industry benchmark. They need to be yours and comparable week to week.

Average check and items per check

Count lunch and dinner separately:

• average check = revenue / number of checks (or / number of guests, if that is how you count locally — stick to one definition),
• average items per check — how many lines a typical receipt has.

If average check rises while item count stays flat, more expensive variants (upselling) or price increases may be at work. If item count rises at stable prices, add-ons are being attached more often. Both signals are useful; they say different things.

You can also read average basket and sales structure in reports — that is venue-wide context, not recommendations alone.

How often the guest adds the suggested add-on (attach rate)

How often the guest actually adds the suggested add-on — that is the attach rate metric. You calculate it for a specific rule, not “for the whole restaurant in general”:

attach rate = (orders with the suggested add-on attached / orders where the rule was shown) × 100%

Example (illustration, not a market figure):
the “burger → sauce” rule showed on 100 burger orders; on 28 the guest added sauce. Attach rate = 28%.

What to do with it:

• a rule with a very low attach rate → change the offer, price, moment, or turn it off (maybe the pair does not make sense),
• a rule with a sensible attach rate and revenue → keep it and train waiters on that same pair,
• compare yourself to yourself (same day of week, lunch vs dinner), not to “the industry”.

In Sales Engine, besides attach rate you also see revenue from recommendations — how many zloty actually came from clicked / accepted suggestions and combos. That is not a promise of check growth. It is a measurement of what already works for you.

When NOT to suggest

Suggestive selling without a brake damages reputation faster than it raises the check. Give the team a clear “stop”.

One offer at a time and the right to say “no” without follow-up questions — that is not softness. That is how waiters stay willing to suggest at all.

14 days on the floor: from 2–3 rules to the first correction

Do not build a matrix for the whole menu on day one. Take two weeks and three rules.

Days 1–2 — pick the rules
Choose 2–3 pairs you have in stock and can say in one sentence. Examples: burger → sauce; pizza → drink; after the main → cheesecake or coffee. Write them on a bar card and (if you have it) in the system as the same offers in QR / the panel.

Days 3–4 — short briefing
15 minutes with waiters: what we offer, at which moment, when we let it go. Drill: two question variants instead of “anything to drink?”. No motivational training — concrete sentences and limits.

Days 5–10 — floor work + notes
Waiters apply the rules. The manager checks once a day: whether QR / online rules are on, whether something sold out, whether someone skips the dessert moment in the rush. Note exceptions (complaints, private events, outage).

Days 11–12 — first read
Count average check and — if you have it — attach rate / revenue from recommendations for those 2–3 rules. Compare lunch with dinner separately. One weak rule? Change the offer or the moment, do not “add five new ones”.

Days 13–14 — correction
Keep what works. Turn off or fix what guests skip. Only then add another rule. The goal is a list the team remembers, not a catalog of 40 items.

How this ties to menu engineering (briefly)

Menu engineering answers: which menu items are worth supporting (including add-ons with good margin and sensible demand). Suggestive selling answers: how to offer them at the table and in QR, and whether the guest takes them.

Do not mash these two topics into one checklist. First you know what is worth pushing — that is a separate guide on menu engineering on sales data. Then you take 2–3 such add-ons and build service and digital rules under them, as above.

If you start with scripts without choosing items, waiters will suggest whatever is on the tip of their tongue. If you stop at menu analysis without concrete offers on the floor and in QR, the knowledge stays in a spreadsheet, and the guest still only hears “anything to drink?”.

Rules and measurement in one place

If you want the same add-on suggestions and combos in the QR menu, in online orders, and in the waiter panel — and to see how often the guest adds the add-on and what revenue that brings — see Sales Engine (Pro and Enterprise plans). OrderNow does not promise “+X% average check”. It gives manager-set rules and measurement of what guests actually attach.

Start with two pairs from the menu and one week of honest measurement. The rest is correction, not another team presentation.

Sources

• MADE FOR Restaurant 2025/2026 via lokal360 — declared guest budget, skipping starters/alcohol/dessert, desire for menu suggestions (ARC Rynek i Opinia, N=1011); declarations, not venue receipts
• GastroWiedza: Up-selling and cross-selling — up-selling and cross-selling definitions and tone of offers at the table; without citing % lift figures
• OrderNow: Sales Engine — add-on suggestions, combos, attach rate, revenue from recommendations, QR / online / waiter panel channels
• OrderNow: Menu engineering — which menu items are worth supporting
• OrderNow: Menu QR — ordering at the table from a phone
• OrderNow: Waiter panel — taking orders on the floor