Gift card revenue arrives in a compressed window. Between early November and December 24, gift cards become the default answer for shoppers who have run out of time, and the days immediately before Christmas are the reliable peak. That concentration is why so many restaurant operators treat the program as a December problem.

The revenue is a December event. The margin is a September decision. By the time the peak-week traffic arrives, every variable that determines what the program earns — bonus structure, corporate order deadlines, digital delivery, redemption terms — has already been fixed. Operators who start in November are not running a Q4 campaign; they are accepting whatever their existing setup produces.

The corporate order is the part that expires first

Consumer gift card purchases can be captured late. Corporate and bulk orders cannot. Companies building holiday gifting programs complete recipient lists and budget allocation in September, with the strategy work starting in July or August. A restaurant that opens its corporate ordering conversation in November is pitching into a budget that closed two months earlier.

This is the single largest avoidable loss in most restaurant gift card programs, and it is invisible in the reporting. A corporate order that never happens does not show up as a declined sale. It shows up as a flat year-over-year number that gets attributed to consumer softness.

The September work is unglamorous: a named contact list of local employers, a one-page bulk order sheet with volume tiers, a stated order deadline, and a delivery method that does not require the buyer to visit. None of it requires new technology. All of it requires lead time.

What the bonus structure actually costs

The dominant promotional format is “buy $X, get $Y” — commonly a $10 bonus on $50, or $20 to $25 on $100. Operators frequently set this rate by copying a competitor, which means the discount rate gets chosen without anyone calculating it.

Run the arithmetic before committing. A $20 bonus on a $100 purchase is a 16.7% discount on the total $120 of face value issued, not 20%. Against a food and labor cost structure, the relevant question is what that discount does to contribution margin on the redeemed portion — and whether the bonus card pulls a visit that would not otherwise have happened.

Two factors work in the operator’s favor. First, redemption is rarely exact: guests spend over the card value often enough that average check on a gift-card visit typically exceeds the card denomination. Second, a share of value issued is never redeemed at all. Industry estimates put unredeemed holiday gift card value at roughly 3% of the roughly $100 billion spent annually on them. Breakage should be treated as a rounding factor that improves the math, never as the reason the program exists — a program designed around non-redemption is a program designed to disappoint guests.

The decision that matters more than the bonus rate is the redemption window. Making bonus value redeemable in January and February moves the cost into the slowest weeks of the year, when an incremental cover is worth considerably more than it is in December. That single term converts a discount into traffic where traffic is scarce.

The measurement problem to solve before the campaign, not after

Gift card ROI is routinely mismeasured because the sale and the redemption land in different periods and often in different systems. Sales are recorded as a liability, not revenue. Redemptions post as tender. If those two records are not linked to the campaign that generated them, the program’s return cannot be calculated at all.

Three things need to be in place before the first promotion runs:

  • A campaign identifier that survives to redemption. Cards issued under the Q4 bonus promotion need to be distinguishable at redemption from cards sold at face value.
  • Average check on gift-card-tendered visits, tracked separately from overall average check. This is the number that tells you whether the program is buying incremental spend or subsidizing existing spend.
  • A redemption curve by month. Knowing what share of December-issued value comes back in January versus June changes both the cash flow forecast and the accounting treatment.

Configuring this in the POS takes an afternoon in September. It is functionally impossible to reconstruct in February.

Digital delivery is now the deadline-extender

The one part of the program that can genuinely be sold late is digital. A physical card requires stock on hand, a display, and a guest who walks in. An e-gift can be purchased at 11 p.m. on December 24 and delivered instantly, which is precisely when the highest-urgency demand exists.

Fiserv’s Q3 2025 Gift Card Gauge, surveying more than 1,000 U.S. consumers, found 53% taking advantage of off-season sales and 38% deliberately spreading gift spending across the year — behavior that rewards operators who sell continuously rather than only in December. Meanwhile 39% of total holiday gift spending is concentrated in the five days from Thanksgiving through Cyber Monday, with 62% of consumers still buying into December.

Those two patterns point in opposite directions, and a program built for only one of them leaves money on the table. The early-spreader needs an offer live in October. The last-minute buyer needs digital delivery working flawlessly on December 23. Both requirements have to be specified now, because the second one in particular tends to fail in ways that are only discoverable through testing.

The September checklist

Against a restaurant and foodservice industry the National Restaurant Association projects at $1.55 trillion in 2026 sales, gift card programs are a small line. They are also one of the few levers that generates cash in advance of the cost of goods, which makes their timing disproportionately valuable.

Before September ends: fix the bonus structure and the redemption window, set and publish the corporate order deadline, confirm e-gift delivery works on mobile, and configure the POS to distinguish promotional cards at redemption. The campaign itself can be built in October. What cannot be built in October is the ability to measure whether it worked.