Case Study 2 — The Dollar Drink: When a Beverage Discount Is Not a Price, and the Independent Who Learned It the Hard Way

This case comes at the chapter's arithmetic from the opposite direction. Case Study 1 showed a restaurant losing its beverage margin to a rule it did not control. This one shows operators giving it away on purpose — sometimes brilliantly, sometimes catastrophically — and the difference between the two outcomes is entirely a question of what the discount was for.

Part 1 — The public case: the dollar margarita

Background

In October 2017, Applebee's — the casual-dining chain operated by Dine Brands Global — began selling a \$1 margarita, promoted as the "Dollarita." The promotion received extensive national press coverage and circulated widely on social media, which is unusual for a drink special at a suburban casual-dining chain and which was plainly part of the point.

The context is a matter of public record. Applebee's was in a difficult period: the brand had been reporting declining comparable-store sales, and the parent company had announced significant restaurant closures. The dollar drink arrived as part of a value-led traffic strategy under new brand leadership. The company subsequently ran a recurring series of \$1 and \$2 "Neighborhood Drink of the Month" promotions — variously a \$1 Long Island iced tea, a \$1 zombie around Halloween, a \$2 spiked tea — turning a one-off stunt into a monthly mechanic.

Company leadership publicly discussed the promotions as a deliberate value play. What was not publicly broken out — and this is the part that matters for our purposes — is what the promotion did to unit-level profitability. Chains do not publish per-item contribution. Neither will your competitors. So the useful exercise is not to guess at Applebee's numbers; it is to work the arithmetic on a constructed drink and see what a dollar price point must mean.

The operating issue: a dollar drink cannot be made up on volume

Here is a constructed casual-dining margarita, at figures typical of the category. All figures illustrative.

Apply §15.8's break-even multiplier and it fails, because the denominator is negative:

$$\text{Multiplier} = \frac{\$6.40}{-\$0.60} \quad \text{— undefined; there is no volume that recovers this.}$$

Every additional dollar drink makes the beverage line worse. That is not a criticism; it is a statement of what the mechanic is. And it tells you immediately that a dollar drink is not a beverage pricing decision at all. It cannot be. It is a marketing expenditure that happens to be denominated in tequila.

Even re-engineering the drink to the bone does not save it as a beverage play. Suppose you rebuild it at \$0.55 — a smaller pour, a well spirit, a batched mix:

• New contribution margin: \$1.00 − \$0.55 = \$0.45
• Break-even multiplier against the \$6.40 standard: 6.40 ÷ 0.45 = 14.2×

You would need fourteen times the volume. Nobody sells fourteen times the margaritas.

What it shows: the discount has to be funded by something else

The only coherent case for a dollar drink is the attachment. Run it, constructed:

THE DOLLAR-DRINK ARITHMETIC, PER GUEST                  [constructed teaching example]

  Guest orders 1.4 dollar drinks             1.4 × (−$0.60)  =  −$0.84
  Guest orders one $11.00 appetizer @ 30%    $11.00 × 0.70   =  +$7.70
  ────────────────────────────────────────────────────────────────────
  Net contribution per guest                                 =  +$6.86

  IF AND ONLY IF:
    (a) the guest is incremental — they were not coming anyway
    (b) the guest buys food
    (c) the guest is not displacing a full-price guest from the same seat
    (d) the drink volume does not create a service or safety problem

  Fail (b) and the guest contributes NEGATIVE $0.84 plus a seat and a server.

Look at what that structure implies about program design, because this is the transferable part. Every architectural choice in a well-built dollar-drink promotion exists to force condition (b):

• Dine-in only. No off-premise, no delivery. The guest has to sit down, which means a server, a menu, and an appetizer.
• One flavor, one price, all month. A single drink of the month is far easier to batch, to train, and to control than a rotating half-off list — and it is not time-compressed, which matters below.
• A limit per guest per visit, where imposed.
• Heavy pairing with food value messaging. The drink is the hook; the appetizer is the business.

An independent operator can steal every one of those design choices without ever running a dollar drink, and should.

The part that is not arithmetic

A promotion that sells alcohol at a price point below its cost has an obvious hazard, and it must be addressed directly rather than left as an implication.

Deep alcohol discounts create over-service risk. Price is one of the frictions that paces consumption. Remove it and you have removed a brake, and the guest who would have had two now has four, in the same amount of time, at the same body weight. That is the mechanism, and it is why several states regulate discount formats rather than discount depths — banning two-for-one, banning unlimited service, banning time-limited pricing, or requiring that a discounted price be available all day rather than in a window. Those rules are not arbitrary; they are aimed at exactly this.

An operator running any deep-discount beverage program takes on a heightened obligation, not a reduced one:

• The same refusal standard applies at \$1 as at \$15, and staff must be told so explicitly and repeatedly, because the price signals otherwise.
• Pace tracking matters more, not less.
• Identification checking gets harder when volume spikes, which is precisely when it must not slip.
• Dram-shop exposure does not scale down with the price of the drink. Chapter 8 owns the liability; Chapters 18 and 25 own the training and certification. Verify your own state's rules on discount formats before designing anything.

The commercial and the ethical point converge here, which is convenient but also true: a promotion built around a single drink at a fixed price available for a long window is both less regulated in most states and materially safer than one built around cheap rounds in a ninety-minute window. Design for the second thing and you usually get the first for free.

Part 2 — The contested decision: an independent that discounted its way out of its own concept

The following is a clearly labeled composite, built from a pattern common enough that most operators will recognize it. It is not a specific restaurant.

The setup

A 70-seat chef-driven neighborhood restaurant, six dinners a week, \$46 average check, opened to good press. Eighteen months in, the 4:00–7:00 window is dead and the owners are paying rent on it. They launch a half-price cocktail-and-wine happy hour, 4:00 to 7:00, six days a week, dining room included.

Year one

It works. The bar fills at 5:00. The room looks alive from the street, which brings in more people. Beverage volume in the window roughly triples. Constructed: the window produces about \$52,000 of additional annual contribution, and everyone agrees the decision was obviously right.

Year two

Three things happen slowly enough that nobody connects them.

The 7:00 seating softens. Guests who used to book 7:00 now arrive at 6:15 to catch the last forty-five minutes of happy hour. Same guests, same night, different price.

The check average falls. Not because anyone ordered less food — because the beverage half of the check is now half price for a growing share of covers. Constructed: the average check drifts from \$46.00 to \$41.00.

The room changes. The 6:00–7:00 period is loud, bar-heavy, and moving fast. It is a good time. It is not the restaurant the owners describe when they talk about their concept, and the guests who wanted that restaurant have started booking at 8:30 or not at all.

The arithmetic

THE SECOND-YEAR RECKONING                                [constructed composite]

  Covers:  100 a night × 6 nights × 52 weeks       =  31,200
  Check erosion:  $46.00 − $41.00                  =  −$5.00 per cover
  Revenue lost:   31,200 × $5.00                   =  −$156,000
  Contribution lost @ 72% (100% − 27.8% COGS)      =  −$112,320

  Happy-hour incremental contribution (year one)   =  +$52,000
  ─────────────────────────────────────────────────────────────────
  NET                                              =   −$60,320

  Plus: a positioning problem that does not appear anywhere in this table
        and that will take two years and a menu change to reverse.

What went wrong, precisely

Not the discount. The scope.

The original problem was a dead 4:00–7:00 window and the goal was to fill seat-hours that were otherwise worthless — which is exactly the shoulder-hour logic §15.8 endorses and Chapter 24 develops. That logic depends on one condition: the discounted guest must be a guest who would not otherwise have paid full price in that seat.

By running the promotion in the dining room, six days, until 7:00 — thirty minutes into the peak seating — the operator converted full-price demand into discounted demand and called the result growth. The happy hour did not fail. It succeeded so thoroughly at the wrong thing that it took a year to notice.

Three design errors, each cheap to avoid:

1. The window overlapped peak demand. A happy hour that ends at 6:00 fills empty seats. One that ends at 7:00 discounts full ones.
2. It ran everywhere. Bar and patio only is a different program from bar, patio, and dining room.
3. Nobody measured the adjacent hours. The only measurement that would have caught this — the 7:00–9:00 window, tracked against a comparable prior period — was never run. §15.8 names it as the variable that decides the answer, and it was the one nobody looked at.

The lesson

A beverage discount is one of three completely different instruments wearing the same clothes, and you must know which one you are holding.

Almost every bad discount decision in this industry is a category error among those three rows — a yield tool applied to peak hours, or a marketing spend judged by its pour cost, or a price set without computing the multiplier.

And one more, which is the thread running through both halves of this case: the mechanic matters more than the depth. A fixed price on one drink, available for a long window, dine-in, with food — is a defensible program at a very low price point. Cheap rounds in a compressed window is a worse program at a higher price point, commercially and in every way that matters more than commerce.

Discussion questions

1. Compute the break-even volume multiplier for a \$14 cocktail costing \$3.50 discounted to \$7.00, and then to \$3.00. Explain in your own words why the multiplier explodes as the discount deepens, and what happens to it below cost.

2. The dollar-drink arithmetic works only if the guest buys food. Design the three specific program rules you would write to make that condition as likely as possible, and name the one you would be most tempted to drop for operational simplicity — and what dropping it would cost.

3. This case argues that a fixed low price available all month is both safer and less regulated than cheap rounds in a compressed window. Explain the mechanism behind the safety claim. Then describe how you would train a bar team to hold the same refusal standard at \$1 that they hold at \$15, given that the price itself is signaling the opposite.

4. In the composite, the operator's happy hour produced \$52,000 of incremental contribution and cost \$112,320 of check erosion. At what end time — 5:30, 6:00, 6:30 — would you have set the window, and what data would you have wanted before deciding? What would you do now, eighteen months in, given that guests have learned the current schedule?

5. The composite's positioning damage "does not appear anywhere in this table." Chapter 2 owns positioning and Chapter 3 owns brand. Argue for a way to make that damage at least partially measurable, and state what you would accept as evidence.

6. A rep offers you a heavily subsidized promotion on a particular spirit: they fund the discount, you feature the brand. Name four things you would verify before agreeing — at least two of them regulatory — and explain how this changes (or does not change) the three-instrument framework above.

7. The hard one. Your happy hour is unambiguously profitable and your dining-room manager tells you the restaurant no longer feels like the place you opened. There is no arithmetic that settles this. Write the decision you would make and the reasoning you would give your partner, in under 200 words.