Restaurant discounts are often judged by the most visible numbers: revenue increased, more orders were placed, or guest traffic was higher during the promotion. None of these results, on its own, proves that the promotion improved the economics of the restaurant.
Some guests may have visited anyway. Existing customers may simply have paid less for the same purchase. A discounted item may have replaced another menu item sold at full price. Orders may have shifted from one sales channel or trading period to another rather than creating genuinely incremental demand.
A restaurant promotion should be evaluated against what would probably have happened without it. The central management question is not how much was sold under the promotion, but how much incremental economic value the promotion created after accounting for discount cost, sales mix, cannibalisation and variable costs.
This changes the analytical sequence. Instead of moving directly from “sales increased” to “the promotion worked”, management needs to move from the result to the factors behind it:
metric → factor → cause → controllable factor → decision → plan → control.
This approach is particularly important for restaurants operating across several outlets, sales channels or markets, where differences in seasonality, customer mix, delivery penetration and trading patterns can make simple period-to-period comparisons misleading.
How Restaurant Discounts Affect Revenue: The Factor Tree
A discount is not an economic result in itself. It is a management variable intended to change customer behaviour and, through that behaviour, influence sales and profit.
A promotion may affect:
- guest traffic;
- the number of transactions or covers;
- conversion from demand to purchase;
- items per transaction;
- the realised selling price;
- menu sales mix;
- average check;
- channel mix;
- repeat purchasing;
- contribution margin.
The analysis should therefore start with the business result the restaurant is trying to change. A promotion designed to increase lunch traffic has to be assessed differently from one designed to stimulate delivery orders, introduce a new menu category or increase the number of items purchased per transaction.
At the first level, restaurant revenue can be expressed as:
Revenue = Number of checks × Average check
This immediately identifies two very different routes to revenue growth. A promotion can generate more transactions, or it can change the value of each transaction.
The number of checks can then be broken down further:
Number of checks = Relevant traffic × Conversion rate
Traffic may be affected by restaurant visibility, location, marketing activity, seasonality, day of week, trading period, local events, tourism patterns and the availability of particular sales channels.
Conversion can be affected by price, the attractiveness of the offer, availability of menu items, the mechanics of the promotion, service speed, assortment and the relevance of the offer to a particular guest segment.
The second branch of the factor tree is average check:
Average check = Average number of items per check × Average realised price per item
The realised price is not simply the menu price. It reflects the combined effect of:
- listed menu prices;
- discount depth;
- the share of discounted items;
- the mix of high- and low-priced products;
- bundles and set offers;
- sales channel;
- other price adjustments.
The mechanics of the restaurant average check therefore need to be understood before management can explain why a promotion changed it.
The resulting factor tree can be viewed as:
Revenue
→ checks
→ traffic
→ conversion
and:
Revenue
→ average check
→ items per check
→ realised price
→ menu mix
→ discounts
Each branch should then be analysed by relevant dimensions such as restaurant, channel, meal period, day of week, menu category, guest segment and promotion type.
This is why the statement “the promotion increased revenue” is analytically incomplete. Management needs to establish which component of revenue changed and what caused that component to move.
Separate the metric from its cause
Suppose the restaurant records an increase in the number of checks during a promotional period. That is a change in a metric, not yet an explanation.
The next question is whether traffic increased, conversion improved, or both.
If traffic increased, the underlying cause could be the promotion, but it could also be seasonality, a public event, a stronger delivery platform presence, changes in opening hours or other marketing activity.
If conversion improved, the cause might be the promotional price, the offer structure, product availability, a change in menu presentation or a different mix of customers.
The analytical discipline is important: two events occurring at the same time do not by themselves demonstrate that one caused the other.
Restaurant pricing and promotions should therefore sit within a broader restaurant pricing and sales-management framework, rather than being managed as isolated marketing activities. Cornell’s restaurant revenue-management guidance likewise treats pricing as a decision that needs to be aligned with demand and restaurant operating conditions through a structured restaurant revenue management approach.
Baseline, Uplift and Cannibalisation: Measuring Incremental Demand
The economic analysis of a promotion requires a baseline: an estimate of the sales that would reasonably have occurred during the same period without the promotion.
Let:
Q₀ = expected sales volume without the promotion
Q₁ = actual sales volume during the promotion
The absolute uplift is:
Uplift = Q₁ − Q₀
The percentage uplift is:
Uplift % = (Q₁ − Q₀) / Q₀ × 100%
This difference is much more useful than the total number of promotional sales. If a customer would have purchased anyway, giving that customer a discount does not create an incremental transaction. It reduces the realised price of demand the restaurant already had.
Building a credible baseline
A baseline can take account of:
- comparable weekdays;
- meal periods and trading hours;
- seasonal patterns;
- recent underlying sales trends;
- comparable outlets in a multi-unit operation;
- sales channel;
- menu category and item;
- known menu-price changes;
- other marketing activity;
- significant external demand changes.
The appropriate method depends on the restaurant and the quality of available data. A busy city-centre venue, a seasonal coastal restaurant, a hotel restaurant in the Gulf and a suburban delivery-led concept may all require different comparison periods.
A simple “week before versus promotion week” comparison can be particularly misleading where demand varies materially by weekday, season, tourism flow, public holidays, weather or local events.
Why the previous period is not necessarily the baseline
The difference between current sales and a previous period can contain several effects at the same time:
- underlying demand growth or decline;
- seasonality;
- calendar effects;
- price changes;
- menu changes;
- marketing activity;
- the promotional effect;
- random variation.
The purpose of the baseline is to separate the promotional effect as far as the available data reasonably allows.
Cannibalisation: additional sales or redistributed sales?
Uplift also needs to be tested for cannibalisation. An increase in sales of the promoted item does not necessarily represent an increase in total restaurant demand.
Several forms of cannibalisation should be considered.
Menu cannibalisation. A guest chooses a discounted product instead of another item that would have been purchased at full price. The relevant result is therefore the change in contribution from the affected menu category, not simply the sales of the promoted SKU.
Channel cannibalisation. Orders increase in the promoted channel while decreasing elsewhere. Delivery, takeaway and dine-in results should therefore be examined both individually and in aggregate.
Time-period cannibalisation. A promotion moves demand into a particular day or meal period rather than creating new demand. Sales rise during the promotion but fall immediately before or afterwards.
Full-price cannibalisation. Existing customers who would have paid the standard menu price instead use the promotion. Transaction volume may remain unchanged while realised price and contribution margin decline.
Restaurant promotion research reported by Cornell University has also highlighted cannibalisation as an issue when promotional coupons are used by existing frequent customers rather than generating wholly incremental demand.
This is why promotional sales should not be equated automatically with incremental sales.
From Discount Depth to Contribution Margin
Once incremental demand has been estimated, the next step is to determine what that demand contributed economically.
Measure the value of the discount
The financial value transferred to customers through a promotion can be calculated as:
Discount amount = Revenue at standard prices − Actual net revenue after discounts
The effective discount rate can be expressed as:
Discount rate = Discount amount / Revenue at standard prices × 100%
This shows the value of the price concession, but it does not show whether the promotion was successful.
The same discount amount could generate substantial incremental traffic, almost no incremental demand, a favourable change in sales mix, or a large reduction in margin on customers who would have purchased anyway.
Move from revenue to contribution margin
For management purposes, revenue needs to be connected to the variable costs generated by the sale.
A simplified calculation is:
Contribution Margin = Net Revenue − Variable Costs
Variable costs should include only costs that actually vary with the volume being analysed under the restaurant’s management-accounting model. Depending on the operation, these may include food and beverage cost, packaging, transaction-dependent channel charges and other costs incurred directly because the incremental sale took place.
The distinction between cost behaviour categories is covered in more detail in the guide to variable and fixed restaurant costs.
A cost that does not change as a result of a modest increase in promotional orders should not automatically be treated as an incremental variable cost simply because it appears in the restaurant P&L.
At item level:
Contribution margin per item = Selling price after discount − Variable cost per item
A discount reduces selling price, while the ingredient cost of the dish does not normally fall simply because the selling price has been reduced. The promotion therefore has to compensate for the lower unit contribution through incremental volume, a more favourable basket, a better sales mix or another measurable economic effect.
Calculate incremental contribution margin
Let:
- Q₀ = expected quantity without the promotion;
- P₀ = standard realised selling price without the promotion;
- VC₀ = variable cost per unit without the promotion;
- Q₁ = actual quantity during the promotion;
- P₁ = realised selling price during the promotion;
- VC₁ = variable cost per unit during the promotion.
The expected contribution margin without the promotion is:
CM₀ = Q₀ × (P₀ − VC₀)
The contribution margin during the promotion is:
CM₁ = Q₁ × (P₁ − VC₁)
The incremental contribution is:
Incremental CM = CM₁ − CM₀
If the promotion generates additional costs that are not already included in the calculation:
Incremental CM = CM₁ − CM₀ − Additional promotional costs
| Measure |
What it answers |
| Promotional sales |
How much was sold while the offer was active? |
| Baseline |
What would probably have been sold without the promotion? |
| Uplift |
How much additional volume was generated? |
| Discount amount |
How much revenue was given up through the price concession? |
| Contribution margin |
What remained after the relevant variable costs? |
| Incremental contribution margin |
Did the promotion improve economic contribution versus the baseline? |
This distinction matters because revenue and profit contribution can move in opposite directions. A promotion can produce higher sales but lower contribution margin.
The cost of discounting existing demand
Promotional transactions should conceptually be separated into two groups:
- sales that would probably have occurred without the promotion;
- genuinely incremental sales generated because of the promotion.
Discounting the first group reduces the margin earned on existing demand.
The economics of a promotion can therefore be viewed as:
contribution from incremental sales
minus
margin sacrificed on customers who would have purchased anyway
minus
contribution lost through cannibalisation
minus
additional promotional costs.
This is a stronger basis for decision-making than transaction growth or revenue growth alone.
Check the effect on average check and menu mix
Promotions can change several components of average check simultaneously. A discounted price may reduce average realised price while a bundle or threshold offer increases the number of items ordered.
As a result, average check may increase, remain broadly stable or decrease.
The metric should therefore be decomposed:
Average check = Items per check × Average realised price per item
The realised price is then influenced by standard pricing, discount depth and sales mix.
Average values can hide significant redistribution between food and beverages, premium and entry-level items, bundles and individual products, or dine-in and off-premise orders. For that reason, promotional analysis should move from total sales down to category and item economics:
Sales → category → item → quantity → realised price → discount → variable cost → contribution margin.
Data and Analytical Dimensions for Promotion Analysis
A useful promotion analysis does not require every possible data point. It requires enough information to reconstruct the relevant cause-and-effect chain.
Sales data
Depending on the restaurant’s systems and analytical model, useful transaction-level fields include:
- date;
- time or meal period;
- restaurant or outlet;
- check or order;
- sales channel;
- menu item;
- menu category;
- quantity;
- standard price;
- realised selling price;
- discount value;
- promotion type;
- net sales.
Economic data
To move from sales volume to profitability, the analysis also needs the relevant cost information:
- variable food and beverage cost;
- packaging where relevant;
- variable channel-related costs;
- other incremental costs associated with the promotion.
Demand data
Where the information is available and reliable, additional demand indicators may include:
- footfall or digital traffic;
- covers or guests;
- conversion rate;
- customer segment;
- acquisition source;
- new versus returning customers.
The purpose is not to build the largest possible dataset. Each field should help management identify a factor, investigate its cause or measure the economic consequence of a decision.
Use the right analytical cuts
A monthly total can hide most of the information needed to understand a promotion. The same offer may perform differently by:
- restaurant location;
- country or market;
- weekday;
- meal period;
- dine-in, takeaway or delivery channel;
- menu category;
- individual item;
- promotion mechanic;
- customer segment;
- period before, during and after the promotion.
For multi-unit operators in Europe or the Middle East, aggregation deserves particular attention. A group-wide improvement can hide underperformance in individual locations, while local promotional success may simply reflect differences in demand patterns, guest mix or seasonality.
Channels should also be evaluated both separately and in combination. An increase in delivery orders has limited incremental value if those same guests have simply moved from another channel with a stronger contribution margin.
Separate controllable and external factors
Management can directly change many elements of a promotion, including:
- discount depth;
- eligible menu items;
- promotion dates;
- days of week and trading hours;
- eligible channels;
- minimum transaction value;
- bundle structure;
- customer segment;
- redemption conditions;
- availability limits;
- interaction with other offers;
- menu assortment;
- standard price;
- communication of the offer.
Other drivers cannot be controlled directly. These may include seasonality, weather, tourism patterns, local events, changes in consumer demand or competitive activity.
External factors still matter because they affect the baseline. Management cannot change the external event itself, but it can change how the restaurant responds through pricing, promotion timing, channel selection, assortment and resource planning.
How to Evaluate a Restaurant Promotion and Turn Analysis into Action
A consistent analysis process prevents teams from selecting the most favourable KPI after a campaign has finished. The expected economic logic should be defined before the promotion starts whenever possible.
1. Define the intended result
Specify which business outcome the promotion is expected to change. “Increase sales” is too broad. A more useful objective would be to increase transactions during a specific low-demand period without reducing total contribution margin.
2. Establish the baseline
Estimate expected checks, revenue, average check, sales mix and contribution margin without the promotion. Use comparable periods and adjust the comparison for known demand differences where the available data allows.
3. Measure the discount
Calculate:
- sales at standard prices;
- actual net sales;
- total discount value;
- effective discount rate.
4. Decompose the revenue change
Start with:
Revenue → checks × average check
Then investigate:
Checks → traffic × conversion
and:
Average check → items per check × average realised price
This identifies where the measurable change actually occurred.
5. Analyse sales mix
Identify the categories and items responsible for the change. Check whether additional promotional items were genuinely incremental or replaced standard-price products.
6. Calculate uplift
Compare actual performance with the baseline:
Uplift = Actual result − Baseline
Where practical, calculate uplift at the same level at which the promotion operates: outlet, channel, meal period, menu category or customer segment.
7. Calculate contribution margin
Compare expected contribution without the promotion with actual contribution during the promotion. Include relevant incremental promotional expenses where necessary.
8. Test for cannibalisation
Check adjacent menu items, categories, channels and trading periods. A gain in the promoted area should be considered alongside any corresponding decline elsewhere.
9. Identify the cause behind the factor
If transactions increased, determine whether this came from greater traffic, improved conversion or another factor. If average check changed, separate quantity, price, discount and mix effects. Only then move to a management action.
10. Recheck the same metrics after the action
Use the same factor tree to determine whether changing the promotion mechanic actually improved the targeted economic result.
Convert findings into a management decision
The output of the analysis should not be “the promotion was successful” or “the discount was too high”. It should connect the result to a controllable factor.
For example:
Result: promotional contribution margin was below plan.
Factor: a large share of discounted transactions came from existing demand.
Cause: customers could obtain the discount without changing their purchasing behaviour.
Controllable factor: promotion eligibility and mechanics.
Decision: redesign the mechanic so that the benefit is linked to the incremental behaviour the restaurant actually wants to stimulate.
Another analysis might show:
Result: transaction volume increased but total contribution changed very little.
Factor: promotional items replaced full-price alternatives.
Cause: the offer mainly redistributed existing menu demand.
Controllable factor: the products included in the promotion.
Decision: revise the promotional assortment and then remeasure category mix and contribution.
Use plan-versus-actual control
For material promotions, management can define an expected model before launch containing:
- baseline sales;
- planned number of checks;
- planned uplift;
- average check;
- expected discount rate;
- expected sales mix;
- variable costs;
- contribution margin;
- additional promotional expenditure.
After the period closes, the analysis follows a consistent sequence:
plan → actual → variance → factor → cause → action.
If actual uplift is below plan, the management question is not simply why sales were lower than expected. The team should determine whether the variance came from traffic, conversion, customer response, assortment, channel, promotion timing, discount depth or external demand.
The same indicators then need to be measured after the corrective action. Otherwise, the restaurant records that a decision was made but cannot establish whether the decision improved the result.
From individual promotions to a repeatable management system
Restaurants that use promotions frequently should build a comparable history of their mechanics and economic outcomes:
promotion mechanic → period → segment → channel → discount → baseline → uplift → mix → contribution margin → incremental economic result.
Over time, this allows management to compare promotional methods by their ability to influence specific business factors rather than by gross promotional sales alone.
RestoFactor provides the methodology for defining metrics, building factor models, diagnosing causes and designing management control. Once the model has been defined, Finoko can be used to automate data collection, calculations, management reporting, budgets, plan-versus-actual analysis and recurring factor control. Automation does not replace the economic model: management first needs to decide what should be measured and which relationships should be tested.
The essential management question remains straightforward: what incremental economic result did the restaurant receive in return for the discount it gave away?
A rigorous answer requires the full chain:
baseline → actual sales → uplift → discount → sales mix → cannibalisation → variable costs → contribution margin → management action → control.
When promotions are managed through this framework, discounts become a measurable factor in restaurant economics rather than simply a tool for generating short-term sales activity. Broader restaurant performance metrics can then be connected through a consistent restaurant KPI framework.
