How much does bundling really lift sales in food & drink?

Bundling is the busiest promotional lever in the sector — the meal deal and the set menu, the bottomless brunch and the sharing board, the day-delegate conference rate and the festival package. Run a group and it sits in every trading calendar. It is popular for one reason: it lifts revenue almost without fail. But revenue was never the test. A bundle earns its place only if it leaves more contribution — the cash a sale leaves after the cost of the food and drink in it, the money that pays the rent, the wages and the profit — than selling the same items apart. Whether it does comes down to two things the headline discount hides: the margin mix of what's inside it, and how much of the take-up is genuinely new. This is what the evidence says about both, for a single venue or a multi-site estate.

The uplift everyone quotes — and why it's fiction

Most bundling advice opens with a number: bundles lift sales by 20–30%, often "according to McKinsey". That figure does not measure bundling, and it does not come from where it claims to.

Strip the folklore away and you're left with a harder, more useful question: what does a bundle actually do, and what does it cost to do it? That starts with seeing a bundle for what it is — not one decision, but two.

A bundle is two decisions, not one

Discounting a single line is one decision: how deep to go. A bundle is two. You set a price — how much cheaper the deal is than buying its parts — and you set a mix: which items go in, in what quantities, at what individual margins. The price decides how much extra volume you need simply to stand still. The mix decides how much each of those extra sales actually earns.

That second decision is what makes bundling its own discipline rather than discounting under another name. Two deals at an identical 20% off can have opposite economics depending only on what sits inside them. Anchor a bundle with a high-margin line — the soft drink, the side, the coffee — and the blend absorbs the discount comfortably; build the same deal from two thin-margin mains (the centre-plate) and that discount guts it. Get the mix right and you can afford to be generous on the headline price. Get it wrong and even a shallow discount loses money on every deal sold. Everything below — and the calculator this article links to — turns on that interaction between the price and the mix.

"How much more will it sell?" is the wrong first question

It is the question everyone reaches for — and the honest answer is that no public figure can settle it. There is no reliable, hospitality-specific number for how much extra a meal deal shifts; the headline percentages are folklore, and a category-wide elasticity borrowed from grocery tells you next to nothing about your deal in your room. Chasing a lift forecast is precisely how bundles get priced on hope. The two questions that can be answered — and that actually decide whether the deal makes money — are the ones the discount hides: does the mix earn enough to carry the giveaway, and is the take-up genuinely new spend rather than business you already had? The rest of this article takes them in turn.

The lever that's unique to bundling: attach, not price

Here is where bundling parts company with plain discounting. A bundle doesn't only lower a price; it simplifies a decision and pulls extra items onto the bill. Controlled experiments find people buy more when items are bundled even with no discount at all: presented as a single unit, a bundle reads as one choice, and shoppers build around it — basket sizes rose by roughly a third among the shoppers who actually chose a bundle (the average across all shoppers was much smaller) in a 2024 retail study, with no price cut involved.[2] A live convenience-store field experiment found an add-on bundle lifted profit by close to a quarter, with more than half of customers taking it.[3]

For an operator, that is the prize worth chasing — a higher attach rate: the side, the drink or the dessert that would not otherwise have reached the table, rather than a blanket markdown on demand you already had. But the same studies carry two warnings worth holding onto. The effect leans almost entirely on take-up — a bundle nobody chooses changes nothing, however well designed — and it does not persist: in the field experiment, behaviour snapped back the moment the deal was withdrawn. Bundling builds a habit only while it runs.

All of which leads to the question that decides the whole exercise: incrementality — the share of bundle sales that are genuinely extra, rather than sales that would have happened anyway. Every bundle handed to a guest who would have bought those items at full price regardless is pure give-back: contribution surrendered on a sale you already had. The costlier version isn't even that — it's the guest who trades down, swapping a £24 à la carte spend for a £16 set menu. That shows up as a healthy cover count and a quietly falling average spend per head, which is why a bundle has to be judged on contribution per cover, not on redemptions — how many guests used the deal. Measuring that is the discipline most bundling programmes quietly skip — and because it's the hinge the whole deal swings on, it gets its own article: how high incrementality realistically runs, what lifts it, and how to prove it.

Why "more covers" can still mean less contribution

This is the part the headline uplift conceals, and it is arithmetic rather than opinion. A bundle works against your margin twice over. The discount comes out of contribution — the cash left after each item's own cost — across every line in the deal; and because a bundle sells more units, cost of sales rises with the volume. Revenue climbs, and the cost beneath it climbs too.

Put numbers on it. Three items retail for £19 together. Strip out the 20% VAT and £15.83 is yours; take off £5.30 of product cost and the set leaves about £10.50 of contribution sold apart. Bundle them at £15.95 — a 16% discount — and the net is £13.29, so the same set now leaves about £8.00. You keep roughly a quarter less per deal, so you need about a third more deals — near +30% volume (£10.50 ÷ £8.00 ≈ 1.3× the deals) — just to stand still. And that is only the break-even: it assumes every one of those extra deals is genuinely new business. The moment some of them are guests who would have bought anyway, the bar climbs higher again. (The example assumes 20% VAT on every line; a deal that mixes food with alcohol, or eat-in with takeaway, carries different VAT by item, so model it line by line — which is what the tool does.) That structural gap, not weak execution, is why so many bundles quietly erode contribution, and it widens fast on thinner-margin lines.

Bundling at scale: portfolio, not promotion

For a multi-site operator the stakes shift again, because a bundle is never run once. It lands in every unit at the same time, across sites and dayparts that behave nothing alike, and it has to survive all of them. That turns three things from store-level details into board-level questions:

• Menu architecture. The mix that protects margin in a city-centre site can sink it in a roadside one with a different cost base and average spend. A deal built around a genuinely high-margin anchor travels across the estate; one built on a hero main dish doesn't.
• Operational drag. Every deal adds prep, training, till logic and waste risk at each site. A few points of margin can disappear into kitchen complexity that never shows up in the promotional plan.
• Who's funding it. Many group deals are part-paid by a supplier — a brewer's wet-led deal, a soft-drinks pour contract, a category marketing fund. That money can reverse the contribution maths above, and proof of incrementality is exactly what a supplier wants before they pay — so model the deal both gross and net of any funding.
• The numbers underneath. Royalties and commissions are charged on the discounted take, not the full price; like-for-like comparisons — the same sites against the same period last year — are distorted by the give-back; and high redemption reads as success when the measure that matters is incremental contribution. The groups that win at bundling cap the discount, anchor the margin and track incrementality, not redemptions.

How to read this — and what it means for the business

Three caveats keep it honest:

• The attach evidence is broad, not bundle-bespoke. No public, line-level dataset for hospitality bundling exists; the basket-size and add-on studies come from shops and convenience stores, so read them as direction rather than a number to bank — your own mix can land either side of them.
• It leans on take-up, and it doesn't persist. A bundle nobody chooses changes nothing, and the field evidence shows behaviour snaps back the moment the deal is withdrawn — bundling builds a habit only while it runs.
• Volume up is not profit up. The one thing that always holds is the arithmetic: break-even is fixed by your margin and your discount, whatever demand does. Judge the deal on contribution and on how much of the take-up is new — never on redemptions.

The takeaway is the same whether you run one venue or a hundred. Don't price a bundle off a hoped-for uplift — there isn't a credible one to lean on. Decide the mix first and anchor it with margin; model the contribution the deal leaves before and after; then judge it on the one thing that actually settles it — whether the margin on the genuinely new items it brings in outweighs the discount you hand back across the whole deal. Where it does, the bundle earns its place. Where it doesn't, fix the mix, the price or the cost before it goes anywhere near the calendar. And either way, prove it with a small, controlled test rather than a hope, because incrementality is the one figure no spreadsheet can settle on its own.