Restaurant Discount Profitability: Promotion Volume Guide

The Promotion Margin Bridge

Use a no-promotion baseline as the left side of the bridge and the promotion period as the right side. The bridge has three order populations:

1. **Baseline orders:** orders you reasonably expect without the offer. These establish the normal price, normal variable cost and normal contribution. 2. **Cannibalised redemptions:** offer users who would probably have ordered at the regular price. Their order still happens, but the restaurant gives away part of its normal contribution. 3. **Incremental orders:** orders caused by the promotion that would not otherwise have occurred in the same period, channel and daypart. Their discounted contribution is the potential economic benefit of the offer.

A redemption is not automatically an incremental order. Treat redemption as an observed activity and incrementality as an assumption to be measured or qualified. If the offer is used during a period that was already full, an incremental order may also displace a full-price order or create extra labour, delay, packaging or refund exposure.

Core formulas.

Define the cost boundary before doing the arithmetic. Include food and beverage cost, order-specific packaging, payment processing, delivery fees or commissions, and any other cost that changes because the order exists. Add promotion-caused incremental labour to the discounted order when it is genuinely attributable to the offer. If labour is a shift-level step cost, keep it as a period cost instead of pretending it is a per-order amount.

Use the same time period throughout. If discounted contribution is zero or negative, the formula does not produce a useful volume target from those order economics; change the offer, cost base or channel before relying on extra volume. Keep taxes, fixed occupancy costs and other period expenses outside the order bridge unless they truly change because of the campaign.

Promotion Margin Bridge worksheet.

Copy the structure below into a spreadsheet or operating note. Enter one currency and one comparable test period. The three channel rows are deliberately separate: a dine-in offer should not inherit pickup or delivery economics.

| Channel | Normal price | Discounted price | Food/beverage cost | Packaging | Payment or delivery fee | Incremental labour per promo order | Baseline orders | Redemptions | Estimated cannibalisation | Incremental orders | |---|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:| | Dine-in | enter | enter | enter | enter | enter | enter | enter | enter | enter | enter | | Pickup | enter | enter | enter | enter | enter | enter | enter | enter | enter | enter | | Delivery | enter | enter | enter | enter | enter | enter | enter | enter | enter | enter |

Add these shared fields below the channel table:

For each channel, calculate normal contribution and discounted contribution before applying order counts. Then calculate cannibalised orders as redemptions multiplied by the estimated cannibalisation rate, unless the operator has a better documented allocation. Calculate incremental orders separately where possible; use redemptions minus cannibalised orders only when the offer population and measurement method make that relationship valid. A baseline order count is context, not proof that every extra redemption came from the promotion.

• **Normal contribution per order** = normal selling price − normal variable order costs.
• **Discounted contribution per order** = discounted selling price − discounted variable order costs.
• **Margin sacrificed per cannibalised order** = normal contribution − discounted contribution.
• **Required incremental orders** = (cannibalised orders × margin sacrificed per order + fixed campaign cost) ÷ discounted contribution per incremental order.
• **Net promotion contribution** = (incremental orders × discounted contribution) − (cannibalised orders × sacrificed margin) − fixed campaign cost.
• **Fixed campaign cost for the period:** enter the actual paid media, platform, creative, code or other campaign cost that is not already in the order rows.
• **Other variable order costs:** record any channel-specific refund allowance, disposable item, third-party fee or service cost that changes between normal and discounted orders.
• **Incremental-order method:** state whether the number is derived from redemptions and a cannibalisation estimate, a comparison period, a holdout group, ordering history, a survey, or a combination. Label an estimate as an estimate.
• **Capacity note:** record the daypart, available production or seating capacity, staffing change and any constraint that could make an extra order unlike a normal order.

Shape the Offer Around the Economics

The bridge should influence the offer design, not merely score a finished coupon. Compare these choices using the same channel worksheet:

The useful question is not “Which offer looks most attractive?” It is “Which offer changes the right orders, in the right channel and period, without making the operation more expensive or less reliable?”

• A percentage reduction can remove more contribution from a high-value order than its headline rate suggests.
• A fixed amount can be easier to model when order values vary, but it still needs a minimum order or item boundary if packaging and fees rise with basket size.
• A minimum-spend offer can support add-ons only when the added items carry enough contribution after their own food, labour and service costs.
• An item-specific offer can steer demand towards items with stronger contribution, provided the item can be prepared consistently and the guest proposition remains clear.
• A time-limited offer can target a slow daypart, but it needs a capacity and baseline check so that it does not weaken a period that already attracts full-price demand.
• Bundles should be costed at order level. The bundle price is not evidence of profitability unless every component, add-on, packaging item and fee is included.

Example: a Hypothetical Promotion Margin Bridge

The following is a transparent, hypothetical worked example using arbitrary currency units (CU). It is a demonstration of the method, not a market price, benchmark or expected result.

A dine-in promotion has a normal selling price of 30 CU and a discounted price of 24 CU. Food and beverage cost is 9 CU per order, packaging is 0 CU, and payment processing is 0.90 CU. The promotion adds 1.10 CU of labour per promotional order. The operator enters 100 baseline orders for the period, observes 40 redemptions, estimates that 50% would have happened without the offer, and assigns 40 CU of fixed campaign cost.

| Input | Hypothetical value | |---|---:| | Baseline orders | 100 | | Redemptions | 40 | | Estimated cannibalisation | 50% | | Cannibalised orders | 20 | | Incremental orders used in the model | 20 | | Fixed campaign cost | 40 CU |

The order economics are:

The arithmetic break-even point is (142 + 40) ÷ 13.00 = **14 incremental orders**. Under the stated assumptions, the model uses 20 incremental orders, so it clears that arithmetic threshold. That is not a decision by itself: the estimate of 50% cannibalisation, the labour assumption and the capacity check still need evidence.

Sensitivity makes the uncertainty visible. If the same 40 redemptions were instead 75% cannibalised, there would be 30 cannibalised orders and 10 incremental orders. With every other input unchanged, the bridge would be 130 − 213 − 40 = **−123 CU**. The promotion would need redesign or a different channel, even though the redemption count stayed at 40.

Do not transfer this dine-in result to pickup or delivery. A delivery row with a marketplace fee, extra packaging and dispatch labour can have a much smaller discounted contribution even when the customer-facing discount is identical.

• Normal contribution = 30 − 9 − 0 − 0.90 = **20.10 CU**.
• Discounted contribution = 24 − 9 − 0 − 0.90 − 1.10 = **13.00 CU**.
• Sacrificed margin per cannibalised order = 20.10 − 13.00 = **7.10 CU**.
• Contribution from incremental orders = 20 × 13.00 = **260 CU**.
• Contribution sacrificed on cannibalised orders = 20 × 7.10 = **142 CU**.
• Net bridge before any other period costs = 260 − 142 − 40 = **78 CU**.

Stop-or-Go Decision Table

Write the rule before the offer starts. “Go” means proceed to the next controlled review, not that the outcome is settled.

| Gate | Proceed when | Modify when | Pause or stop when | |---|---|---|---| | Arithmetic bridge | The conservative contribution bridge covers sacrificed margin and campaign cost. | It clears only under an optimistic cost or volume assumption. | Discounted contribution is zero or negative, or the bridge remains negative. | | Incrementality | A documented comparison, holdout, ordering pattern or survey supports the incremental-order estimate. | The evidence is mixed; narrow the audience, daypart or item set. | Most redemptions appear to replace full-price demand. | | Capacity | Seating, kitchen, dispatch and service capacity can absorb the additional orders. | Extra labour or slower service is likely; price that effect into the model. | The offer displaces full-price demand or damages a defined service control. | | Channel economics | Fees, packaging, refunds and labour are recorded for the actual channel. | Rebuild the offer by channel or exclude weak items. | A channel row is missing material costs or cannot be reconciled. | | Measurement | Baseline period, offer window, codes and order categories are defined. | Use a smaller scope or improve data capture first. | The result cannot distinguish redemptions from genuinely additional orders. | | After-offer behaviour | Repeat ordering and full-price behaviour are tracked as separate observations. | Keep the offer narrower while the pattern is unclear. | Continued discounting is the only reason demand appears to persist. |

Implementation: From Offer Idea to Review

1. **Define the decision.** State the purpose, channel, daypart, eligible items, discount mechanic, test window, capacity limit and owner. A promotion designed for a quiet pickup period is a different decision from a broad peak-time discount. 2. **Freeze the baseline.** Choose a comparable no-offer period and record orders, average order value, item mix, channel, daypart and contribution inputs. Note unusual events rather than silently blending them into normal demand. 3. **Cost the order.** Use current recipe and beverage costs, packaging, payment or delivery fees, refunds and promotion-caused labour. Confirm whether the platform funds any part of the offer or passes the full reduction to the restaurant before entering the fee row. 4. **Separate the populations.** Count redemptions, identify likely baseline customers, and estimate cannibalisation. If direct incrementality is uncertain, show low, central and high cases rather than presenting one estimate as observed fact. 5. **Set stop rules.** Choose the minimum contribution bridge, maximum capacity load, data-quality requirement and review date before launch. Include a rule for pausing if the offer pulls demand into an already constrained period. 6. **Measure by channel and daypart.** Use unique codes or loyalty identifiers where available, and compare like with like. A holdout group, a similar non-promotional period or a customer survey can inform the estimate; none should be treated as perfect attribution. 7. **Reconcile actuals.** Replace assumptions with actual fees, order mix, refunds, labour hours, packaging use and observed orders. Explain every material difference between the worksheet and the operating record. 8. **Choose one action.** Continue only with a documented contribution and capacity case; modify the audience, timing, threshold, product mix or channel when the demand is useful but the economics are weak; pause when the offer mainly transfers value from regular orders.

For item-level inputs before choosing which dishes or bundles can carry an offer, use the [free RestaurantMargin margin calculator and menu-profit dashboard](https://restaurantmargin.com/). The workspace lets you enter ingredient cost, selling price and weekly sales, then view food-cost percentage, contribution margin, weekly contribution, pricing targets and menu-engineering position. The site also distinguishes contribution from net profit, so keep fixed costs, tax and other period expenses in the wider review.

Limitations

This bridge is a decision aid, not a complete profit-and-loss statement. Cannibalisation and incrementality are counterfactuals: the same customer cannot be observed both using the offer and not using it in the same moment. Comparison periods, holdouts, ordering patterns and surveys can improve the estimate, but each can be distorted by weather, events, menu changes, competitor activity, seasonality or data gaps.

Contribution margin also stops short of net profit. Fixed occupancy costs, salaried management, tax, debt service and other period expenses still need separate review. A promotion can have a positive order bridge and still be the wrong use of a scarce kitchen, dining room or management team.

Cost behaviour can change at thresholds. One extra order may use existing capacity; many extra orders may require a person, a courier handoff, a second prep batch or additional packaging. Model those step costs explicitly. Refunds, comps, substitutions, waste and poor item mix can also make actual contribution differ from the worksheet.

Finally, repeat behaviour is difficult to attribute to one offer. A later full-price order is useful evidence to track, not proof that the original discount caused it. Keep assumptions visible, replace them with reconciled observations, and narrow the scope whenever the data cannot support a clean decision.