Restaurant Loyalty Rewards: Free Item or Discount?

Choose a free item when you can offer something guests genuinely want at a predictable incremental cost. Choose a discount when flexibility matters more to guests and you can afford the contribution given up across eligible baskets. Neither format wins on its face: the answer changes with basket size, what guests would have bought anyway, and the rules around redemption.

The fairest comparison starts with rewards that look equally valuable to a guest at a typical basket size. Then check what happens on smaller and larger orders. A percentage discount grows with the bill; a single free item usually has a more predictable cost, provided it does not replace an item the guest was already going to buy.

Start with comparable guest-facing value

Consider this hypothetical restaurant. Its free-item reward is one side with a regular menu price of $6 and an incremental food and packaging cost of $1.80. Its alternative is 20% off eligible food. On a $30 eligible basket, that discount is also worth $6 to the guest.

Those rewards are equivalent in stated menu value at $30, not necessarily in what every guest would choose. A guest who wants the side can redeem it as a $6 item. A guest who does not want it cannot turn that stated value into a different $6 reward unless the program allows a substitution. The discount, by contrast, can reduce the price of any eligible items in the basket.

Before comparing costs, write down what each reward actually permits:

• Free item: Which item or items qualify? Are size upgrades, substitutions, or modifiers included?
• Discount: Which food and ordering channels qualify? Is there a maximum discount or minimum purchase?
• Both: How is the reward earned, when does it expire, and can it be combined with another offer?

Keep the earning requirement the same for this comparison. If one reward requires more purchases or points, you are comparing both the reward and the effort needed to reach it. That may be useful later, but it makes the first decision harder to read.

Also check how the offer appears where guests order. A short description such as “one eligible side” may be easy to scan, but guests still need to know which sides qualify before they commit. For the discount, show whether it applies to the whole food basket or only selected items. A reward that needs a long explanation at checkout can create extra work for the team even if its numbers look attractive.

Compare contribution at two basket sizes

For the same hypothetical comparison, assume eligible food costs 30% of its regular selling price. The guest adds the free side to an order they would otherwise place and would not have bought that side without the reward. The discount applies to food the guest would otherwise buy at full price, with no change in food quantity. These assumptions isolate the reward cost; real behavior may differ.

The table shows the reduction in contribution against the same order without a reward. For the free side, that is its $1.80 incremental cost. For the discount, food cost stays the same, so the contribution reduction equals the discount. The table also shows contribution after the reward on the eligible food basket.

At $30, regular food cost is $9 and contribution before a reward is $21. The free side adds $1.80 in cost, leaving $19.20. The discount takes $6 off sales while the original $9 food cost remains, leaving $15. At $50, regular food cost is $15. The free side leaves $33.20, while the $10 discount leaves $25.

This worksheet includes variable food cost on the original basket and the free side's variable food and packaging cost. It excludes fixed costs such as rent and salaried management, as well as other variable costs such as payment processing and any delivery charges. It also excludes exception costs, including remakes, substitutions and refunds. Add those from your own operating records before treating the figures as a budget.

Under these assumptions, the free side protects more contribution at both basket sizes. It also becomes relatively more attractive as the basket grows because the 20% discount grows with it. That is not a prediction of total profit: the comparison does not yet account for a guest changing the order, visiting more often, or leaving the reward unused.

A discount cap can limit exposure on large orders, but it also changes the guest-facing offer. For example, a 20% discount capped at $6 still takes $6 off a $30 eligible basket, but it would take only $6 off a $50 basket. Show the cap clearly rather than letting a guest discover it at payment.

Adjust for redemption mix and menu constraints

The free item becomes less attractive if guests redeem it for something they would have purchased anyway. In the hypothetical $30 order, suppose the guest would have bought the $6 side at full price instead of receiving it as an extra. The restaurant then gives up $6 in sales while still incurring the side's $1.80 cost. Compared with the original order including that side, contribution falls by $6, not $1.80.

That outcome matches the contribution reduction from the $6 discount on the $30 basket. At that point, the discount may be the better guest-choice option if both offers are equally easy to redeem and the restaurant can accept the larger exposure on bigger baskets. The free side still has value as a recognizable reward, but its cost advantage depends on whether it adds an item rather than replacing a sale.

Build a simple redemption mix before making the decision. Estimate the share of free-item redemptions that are truly additional and the share that replace a full-price purchase. For an illustrative mix of 70 additional sides and 30 displaced purchases per 100 redemptions, the contribution reduction is $126 for the additional sides and $180 for the displaced purchases: $306 total, or $3.06 per redemption. This is a separate scenario from the two basket rows, not a forecast of actual guest behavior.

Compare that average with the discount at the basket sizes you expect. If every one of those 100 discounts were redeemed on a $30 eligible basket, the reduction would be $600, or $6 each. If basket sizes differ, calculate each discount at its actual eligible basket value rather than assuming the average guest receives exactly $6.

Then check the menu and service realities behind the estimate:

• Is the free item available during every eligible ordering period, or can it sell out?
• Would a choice among two or three items improve appeal without making costs unpredictable?
• How are paid modifiers handled? A free base item with a paid upgrade needs a clear price at checkout.
• Can employees explain exclusions quickly, including whether the reward works on pickup, dine-in, or another channel?
• Can the system apply the exact restriction you want, or will staff need a manual workaround?

Ask vendors to demonstrate the actual rules, not just the reward label. Check whether an eligible order earns points, whether a reward can be limited to the intended items, and what staff see when a guest disputes eligibility. Fleksa's restaurant marketing automation page describes loyalty points for eligible orders, promo codes, and campaign tools; confirm the setup and channels that fit your program.

Check refunds and partial redemption before launch

Reward economics can change after an order is placed. Decide in advance what happens if an eligible item is refunded, an order is canceled, or a guest uses only part of an available reward. The answer should be visible to staff and understandable to the guest.

For a free item, ask whether a canceled order restores the reward and whether a refund for the free item has any cash value. For a percentage discount, ask how the discount is allocated when one of several items is refunded. Also ask whether a reward can be used again after a partial refund. Do not assume the system will follow your preferred policy without a demonstration.

Use a separate hypothetical exception to inspect the arithmetic. A guest orders two eligible $15 dishes and receives 20% off, paying $24 for the $30 food basket. If one dish is refunded and the discount was allocated evenly, that dish's paid amount is $12. Refunding $12 leaves $12 in food sales for the dish kept. If its food cost was $4.50, the kept dish contributes $7.50 before other costs. The refund amount is not a second discount; it reverses part of the original sale.

That example includes the kept dish's variable food cost and excludes fixed costs, other variable costs and exception costs. If the refunded dish was already made and cannot be reused, its $4.50 food cost remains. The transaction then contributes $3.00 before the excluded costs. Your actual result depends on the allocation and refund rules, so inspect those rules in the system and consider the cost of food already prepared.

Run these checks in a simulation or training environment using test records only. Do not charge a real card, send a message to a guest, or dispatch a courier. Include an unavailable free item, a paid modifier, a canceled order and a partial refund. Record the expected outcome before the check, then compare it with what staff would see. A simulated check confirms the workflow you can observe; it does not show how guests will behave during service.

Run a limited live pilot and decide

Once the rules work in simulation, choose a limited live service pilot with a defined period and a manageable group of eligible guests. Brief the team on the offer, exclusions, recovery steps and who can resolve an exception. Use normal production controls and real transactions only when the restaurant is ready to serve those guests. Monitor the pilot and be prepared to pause it if the offer causes confusion or an unexpected cost.

Track eligible orders, redemptions, basket value before the reward, the reward applied, refunds and the variable costs you can reasonably associate with the offer. Separate the number of rewards issued from the number actually redeemed. A low redemption rate reduces immediate reward expense, but it can also mean the offer is hard to understand or unappealing.

Incrementality is the harder question: did the reward lead to a visit or purchase that would not otherwise have happened? A larger basket on a redeemed order is not proof by itself. Compare with a suitable group or period without the offer where practical, and note differences in season, promotions and guest mix. Ask staff what guests say, but do not treat a few comments as a measurement of the whole program.

Check reporting before you commit. Can you export redemption and refund records, distinguish eligible food from excluded items, and see the basket before and after the reward? If you plan to contact guests about the offer, ask how audience selection and send controls work. Access to contact details is separate from permission to market to a guest.

Use the free item when the pilot supports the additional-item assumption and guests value the available choice. Consider the discount when guests prefer flexibility and its observed basket mix stays within the contribution reduction you are prepared to accept. If neither meets both conditions, narrow the eligible menu, adjust the cap, or revisit the earning requirement and run the comparison again.

For a deeper look at the cost of one design, use the loyalty reward cost guide. If online redemption is part of the plan, map the guest and staff steps against your restaurant ordering system. Next, compare the reward rules, reporting and exception workflow against your operating requirements with Fleksa, using the two reward designs you are considering.

Frequently asked questions

Is a free item always cheaper than a discount?

No. An additional free item may cost only its incremental food and packaging cost, but a reward that replaces a full-price purchase gives up that sale. Compare both cases with the discount on the baskets where guests will use it.

How can I make the rewards comparable?

Give them the same earning requirement and compare their stated value at a typical eligible basket. Then check other basket sizes and whether guests can actually redeem the free item they want.

What should I check before a loyalty pilot?

Simulate eligibility, substitutions, cancellations and partial refunds without real payments, guest messages or courier dispatches. Then run a limited live pilot and review redemptions, contribution and guest understanding before expanding it.