The Discount Trap: What 2-for-1 Tuesday Teaches Guests

A poké bowl owner I read about recently ran a promo he was proud of. Two premium bowls for EUR 20, normally EUR 33 together. He called it TwoForTuesday and pushed it on every channel he had. Almost nobody booked. The few who came were mostly existing weekday regulars who would have bought one bowl anyway.

He thought the problem was the offer was too small, or the marketing too weak. The commenters who piled on told him something harder to hear: the deal worked exactly as deals work. It filled a few seats at a loss, it trained a small group to expect EUR 10 bowls on Tuesdays, and it gave new guests nothing to come back for once the promo ended.

There's a cleaner way to spend the same discount money. And it isn't louder marketing.

The cost you're not counting

Most restaurant owners think about a promotion as a line item. Thirty percent off a check is a thirty percent hit on that check's margin. That's the math on the receipt, and it's the wrong math.

The real cost is the price your guest now remembers. Daniel Kahneman and Amos Tversky showed in 1974 that people's numerical judgments are heavily biased by whatever number they saw first — even when that number is obviously unrelated to the question. In their well-known study, participants who spun a rigged roulette wheel that landed on 10 later estimated that 25% of UN member countries were African. Participants whose wheel landed on 65 estimated 45%. Same question, different anchor, different answer.

Richard Thaler later built on that foundation with what he called transaction utility — the idea that guests don't judge a price in isolation. They compare it to a reference price in their memory. If they paid less than the reference, they feel they got a deal. If they paid more, it stings. The reference price isn't set by the menu. It's set by whatever price they paid last time, or saw advertised most recently.

When you run 2-for-1 Tuesday, you aren't just giving away margin on the nights it runs. You are writing a new reference price into every guest who takes the deal. The bowl that used to be EUR 16.50 is now, in their memory, a EUR 10 bowl. Full price on any other night will feel like a markup.

That reframing is the cost you're not counting. And once it's set, it's sticky. Research on promotional cannibalization suggests a large share of sales lift during a discount comes not from genuinely new demand but from existing customers shifting when they buy — moving full-price visits onto the discount window.

The math, run two ways

Consider a 35-seat neighbourhood bistro. Average check EUR 38. Tuesday runs at about 40% capacity — 14 covers.

Scenario A: 2-for-1 Tuesday

Drop the Tuesday check from EUR 38 to EUR 19 for anyone who shows up in pairs. The promo pulls capacity up from 14 to 28 covers — a big lift on paper.

Revenue goes from EUR 532 to EUR 532. Exactly the same. The doubled headcount is cancelled by the halved check. Food cost at 30% is now EUR 160 (up from EUR 80). Labour needs to scale for the busier room — add one more server and a kitchen hand, call it EUR 90 extra. You end the night with EUR 282 of contribution, versus EUR 372 on a quiet full-price Tuesday.

The 2-for-1 Tuesday costs the bistro roughly EUR 90 per service in lost contribution, every week it runs. Over a year, that's around EUR 4,700.

But the receipt-level loss isn't the main cost. Of the 14 new guests the promo pulled in, most now hold a EUR 19 reference price in their head. Ask them to come back on Friday at EUR 38 and they'll feel overcharged on a meal they would happily have paid EUR 38 for a month ago. A fraction of them will come back at full price. Most will wait for next Tuesday. Some will stop thinking of the restaurant at all once the promo ends.

Scenario B: Buy three, get the fourth free

Same total discount commitment. Different structure.

Any guest who visits four times inside a rolling three-month window gets a fourth visit free — up to a meal value of EUR 38. No "half off everything." No special Tuesday. Just a commitment reward for repeat visits, tracked by the booking system.

The redemption cost per guest who completes the card is EUR 38 — one full visit's worth of revenue, fully comped. That's more than the EUR 19 per Tuesday cover discount. But look at what the guest had to do to earn it.

They had to visit three times at full price first. Three covers at EUR 38 is EUR 114 of revenue the bistro wouldn't necessarily have captured without the nudge. The fourth visit is the comp. So the promotion costs EUR 38 per completion but drives EUR 114 of full-price revenue per completion — and creates a guest who has now eaten at the bistro four times in a quarter.

A guest who has visited four times is on the path to becoming a regular. Loyalty-curve data from e-commerce shows the probability of a customer returning jumps from around 27% after the first purchase to 62% after the third. The shape of that curve is consistent across industries that rely on repeat visits.

Thirty guests per month completing that card is 30 × EUR 114 = EUR 3,420 in incremental full-price revenue, against EUR 30 × EUR 38 = EUR 1,140 in comps. Net contribution is positive, and the side effect is 30 people who walked in four times in 90 days.

Same total discount money. Completely different reference price. Completely different behaviour on the other side.

Why the punch card works

There's an actual academic reason a card-based reward outperforms a percentage-off. Ran Kivetz, Oleg Urminsky, and Yuhuang Zheng published a 2006 study in the Journal of Marketing Research showing that people accelerate their purchases as they get closer to a reward. They gave one group of coffee drinkers a 10-stamp loyalty card and another group a 12-stamp card with two stamps pre-filled. Both groups needed 10 stamps to earn the reward. The group holding the pre-filled card — closer to the goal at every moment, even though the stamps needed were identical — completed it measurably faster.

They called it the goal-gradient hypothesis. As the finish line gets closer, people speed up. A punch-card structure does two things a 2-for-1 can't. It changes the question in the guest's head from "is tonight a deal night?" to "where am I on my card?" And it turns each visit into progress toward a reward rather than a discrete transaction.

A percentage discount teaches guests to wait. A progress bar teaches them to accelerate.

The effect compounds when visits are spread across different weeknights. Ten guests working a four-visit card over ninety days means forty bookings distributed across months — not fourteen bookings squeezed into one over-promoted Tuesday.

When a discount actually makes sense

I'm not arguing that all discounts are a mistake. They aren't. But the situations where they work are narrower than most operators treat them.

A genuine soft launch. A new restaurant opening with a lower price for the first three weeks is buying reviews and word-of-mouth in a window where the reference price hasn't been set yet. The guest who walks in that first Wednesday doesn't yet have a "normal" price in mind. You're writing the reference price from scratch, not discounting it.

A new-menu launch around a specific experience. A prix fixe tasting at a reduced price for the first month of a seasonal menu is about introducing the dishes, not competing on price. The discount is framed as an invitation, not a deal.

A genuine slow-season programme framed around the experience, not the number. A winter wine series with a set per-head cost reads differently from "20% off January." The guest remembers the wine dinner. They don't remember the discount.

In all three of these, the discount is tethered to something else the guest takes away — a new menu, an experience, a reason to have come — that doesn't collapse back into a percentage the moment the promo ends. The promo isn't the point. It's a way into the point.

What doesn't work: recurring, date-based discounts with no story attached to them. "Tuesdays 20% off" trains guests to associate Tuesday with a lower price. "Wine dinner first Wednesday of the month" trains them to associate the first Wednesday with an experience worth showing up for. Same night-of-the-week problem, very different reference price set in memory. Our piece on giving quiet nights an identity gets into this distinction in more depth.

What a booking system should do to support loyalty

Most of the repeat-visit structures worth running — punch cards, buy-3-get-1, visit-count rewards, birthday comps — fall apart without the infrastructure to track them. You need to know who's walked through the door, how many times, and when.

This is where a booking system earns its keep. Nine Tables builds a guest profile from every reservation, keyed on phone number or email. The guest doesn't have to sign up for anything. The system records each visit automatically, carries preferences forward, and surfaces context at the moment of the next booking. A host sees "fourth visit in 90 days, redeem a free main if they book" when the name pops up on the sheet. No separate loyalty app, no stamp card getting lost in a wallet.

That kind of structure is what makes spread-over-visits promotions operationally feasible. Without it, you're back to a clipboard, or to a percentage-off that lives on one night a week because that's the only promo shape a restaurant can actually run by hand. We've written about why our pricing is flat-rate rather than per-booking — the same logic applies here. When the booking system is the cheapest part of the stack, it can carry the weight of the loyalty structure without the software itself eating the margin you were trying to reinvest in regulars.

A trade-off worth naming: tracking repeat visits requires a guest to book rather than walk in, or for walk-ins to share a phone number or email. You'll miss some. That's real. For most neighbourhood restaurants, the share of guests who'll happily leave a number for a small reward is high enough that the structure pays back.

The question every promotion has to answer

Before you run anything — a 2-for-1, a punch card, a Tuesday deal, a birthday comp — ask what the promotion teaches.

A price-off teaches guests your normal price is negotiable.

A visit-count reward teaches guests their loyalty is worth tracking.

Both cost you roughly the same money on the P&L if you size them carefully. They cost you wildly different amounts in the reference price you've written into guest memory, and in the kind of guest who comes back next.

The poké bowl owner who ran TwoForTuesday wasn't wrong to want more Tuesday traffic. He was just spending his discount dollars in a shape that doesn't generate more Tuesday traffic — it generates fewer full-price Tuesdays, for a while, and trains a small group of deal-hunters to wait.

Same money, spread differently, does different work. What is your next promo teaching?