Key Takeaways: Beverage and Bar

The one thing

A bar is a rate against a duration, not a budget line.

The queue is demand per minute against service per minute. Running out is consumption per hour against inventory. The over-served guest is drinks per hour against food and time. The invoice shock is pours per guest against a per-pour price.

Every one of them is invisible in a total and obvious in a rate, and modelling the rate takes the same ten minutes as writing the total.

The nine claims

1. The four models are three pricings of the same generosity, plus one that is a different thing. Package is predictable; consumption is unbounded without a written cap; cash is a licensing question before it is a social one. A hosted limited bar — beer, wine, one or two batched cocktails — is the right answer most of the time, because it is cheaper, roughly twice as fast, and carries the couple's mark better than forty bottles nobody notices.

2. Without a forecast, prefer the package. Not because it is cheaper — it may not be. Because being wrong on a package costs a known few hundred and being wrong on consumption is unbounded and arrives three weeks later.

3. Forecast from the guest list, not from a rule.

Drinkers × (2.0 in hour one + 1.0 per hour after) ± age, transport, weather, community, dancing + 10% shrinkage

A rule of thumb cannot be corrected because it cannot be argued with. A decomposed forecast has seven lines and every one of them can be challenged and fixed.

4. Convert drinks into things you can buy. 750 ml of still wine = 5 glasses; sparkling = 6 flutes; spirits = 16 pours; a half-barrel = 165. Ice is 2.5 lb per guest at a normal venue and 4 lb per guest at a raw site, because at a raw site everything that has to be cold must be made cold with ice.

5. Buy 15% over on wine and exactly to forecast on chilled beer, because the marginal bottle of wine comes back and the marginal case of beer does not. Get the returns policy in writing before you buy anything.

6. The queue is a design problem before it is a staffing problem. Passed drinks remove 40–60% of demand. A batched cocktail pours in 8 seconds instead of 45. A self-serve beer tub doubles throughput for that share. Two bars halve the queue length for free. A third bartender is the most expensive of the five and the one clients ask for first.

7. Five alcohol questions, and the fifth is the one nobody asks. Who holds the licence · who may pour · sale or furnishing · who carries liquor liability · and who has authority to refuse service. Name them in writing at −21 days. Not the planner, and never a family member.

8. Responsible service is seven items written before the day. A named refuser · food early and continuously · water everywhere, unasked · non-alcoholic options that are not a punishment · a flattened rate rather than a hard stop · a departure plan · and somebody sober who is watching.

9. Ten to twenty per cent of the room is not drinking, and they are routinely served worse. A non-alcoholic version of the signature cocktail — same glass, same garnish, same batch minus the spirit — is the highest-return item in the chapter. Do not label it. A "mocktail menu" separates the people ordering from it.

The arithmetic to internalise

Forecast: drinkers × (2 + 1×(hours−1)), adjusted, +10%

Break-even, package vs. consumption: package per head ÷ price per drink = drinks per guest

Cocktail-hour spike: ~45% of first-hour demand in the first 20 minutes

Bartender throughput: 30–45/hr mixed · 60–80/hr beer and wine only

Ice: 2.5 lb/guest normally · 4 lb/guest at a raw site

Threshold concept

🚪 A bar is a rate against a duration.

"$3,100 for the bar" tells you nothing. "650 drinks over 320 minutes through two pairs of hands" tells you the inventory, the queue, the staffing, the ice, and the closing plan — and it takes the same ten minutes to write.

What this chapter added to the project

The Reyes–Whitfield bar, complete at $2,732.31 against a $3,100 allocation — $367.69 under. The first line in the project to come in under, and the case study is explicit that three of the five reasons were decisions and two were accidents of an allocation set at 118 guests and never re-based.

Forecast: 650 drinks. 88 drinkers across 5 hours 20 minutes, adjusted −12% for a rural site with no shuttle and +18% for a warm afternoon and a dancing-heavy reception. The two largest adjustments nearly cancel, which no rule of thumb could have told anyone.

Two batched signature cocktails — a paloma and an apple-and-bourbon — at about $180 of spirits total, chosen because they are simultaneously the fastest, the cheapest, and the most personal thing on the drinks list. The paloma exists without the tequila, in the same glass, on the same unlabelled list.

The ice finding: 400 lb, not 250. At a raw site with no refrigeration, 218 units of wine, sparkling, and beer must be chilled in ice water for four hours before the bar opens. Two deliveries — 220 lb at 1:00 p.m. and 180 lb at 4:30 — plus $85 of tubs that appeared in nobody's order.

The corkage clause that did not apply. Sundown's proposal carried $9/bottle on alcohol served by their staff. At 64 bottles that is $576. Confirmed in writing that it was not engaged. Nine minutes.

And the cocktail-hour crunch solved entirely demand-side: 205 drinks demanded against 89 servable, fixed with two trays of passed drinks, batched cocktails, and a self-serve beer tub. Zero additional bartenders.

What is still open

Spaced review

From Chapter 11: the queue arithmetic, and that four of the five fixes are free. This chapter tested it against a real event and the free fixes were sufficient.

From Chapter 9: liquor liability is carried by a business that sells or serves and is frequently excluded from general liability. A certificate showing $1,000,000 general liability tells you nothing about whether the bar service is covered for what the bar service does.

From Chapter 14: the loaded rate. Sundown's 1.2775 does not transfer to the bar — alcohol is taxed differently in most jurisdictions, and client-supplied alcohol is taxed at the point of retail purchase instead. Read the tax line in each contract separately.

From Chapter 5: fully variable costs. The bar is one, which is why a guest reduction released real money here where the tent released none.

Before Chapter 16

Chapter 16 leaves arithmetic behind and starts on design.

Bring two things from here. The batched paloma is a design object as much as a drink — it is one of the few places a couple can put something specific into the room without buying anything. And the bar is the largest piece of furniture at most weddings: the thing guests approach most often and stand at longest, and in most weddings a folding table with a cloth on it.

Chapter 16 will ask why.