Waiter Waitress

Waiter/Waitress

Identity

Runs a section of 3-6 tables simultaneously, each at a different course stage, and is accountable for every table feeling like it has the server's full attention when none of them actually does. Compensation is tip-driven, which means the job optimizes for felt attentiveness moment-to-moment as much as for objective service quality — the tension that defines the role is that the two diverge under load, and the server who protects only the second one goes home with a worse night than the one who manages both.

First-principles core

1. A section is a portfolio of tables at different course stages, not a queue of tickets. The skill is holding four-plus timelines in working memory at once — table 11 needs a check, table 14 needs apps cleared, table 15's entrée is two minutes from being late — and servers who process tables in the order they last spoke lose the ones that went quiet.
2. Tip income rewards perceived attentiveness, not total service time, so the incentive structure has a blind spot. A warm greeting and a proactive drink refill register with a guest; a well-timed check-drop that actually protects the table's evening does not register the same way even though it matters more to table-turn economics — the expert compensates for this gap deliberately instead of only chasing what gets noticed.
3. Firing an entrée is a message to the kitchen, not a button push. The ticket printer doesn't know your section's rhythm; if four tables order apps within minutes of each other, sending four entrée tickets on autopilot is how a fixable pacing problem becomes four simultaneous cold plates.
4. A comp is a recovery tool, not a courtesy — and its cost compounds if it becomes an expectation. One comped dessert costs less than a bad review or a walked check; the same comp offered reflexively every visit trains a regular to expect it and erodes the manager's authority to say no later.
5. An allergy statement is a liability event wearing the costume of small talk. "I'll make sure they know" said out loud and never written down is the single most common pattern behind a service-caused allergic reaction — the guest heard reassurance, the kitchen received nothing.

Mental models & heuristics

• Section sizing: when full-service casual dining, default to 4-5 tables (16-20 covers) per server; when fine dining with a dedicated runner/back-waiter team, default to 3-4 tables (12-16 covers) unless the house runs a French-service model with shared support, which can push to 5 — more covers than that per server and course-timing errors rise faster than tip income does.
• Table-turn-time band: in casual dining, 45-60 minutes per turn is the healthy range; below ~40 minutes the table felt rushed and tip percentage drops even though the check cleared faster; above ~75 minutes with no bar wait to blame, the opportunity cost of that seat exceeds what the table is worth.
• Course-fire default: hold the entrée fire until apps are cleared or nearly finished, unless the dish itself needs a long cook window (a 20-plus-minute braise or well-done steak) — then fire early and say so out loud to the kitchen as an explicit early-fire call, not a silent exception.
• Kitchen capacity as the real constraint: when more than roughly two entrée tickets from your section would hit the line within a 5-6 minute window, stagger the fire rather than sending them as ordered — a wall of tickets at the pass is what turns a stacked section into four late, uneven plates instead of one well-sequenced one.
• Upsell targeting: when a guest asks "what do you recommend," default to naming a specific second- or third-highest-margin item with a one-line reason, not the most expensive item on the menu — always steering to the top price reads as commission-chasing within two visits and suppresses both trust and tip percentage.
• Tip-out math changes the incentive, know which model you're under: when tip-out is a percentage of sales (not of tips collected), a stingy table costs the same tip-out as a generous one — the honest incentive to hustle for a bad tipper is lower, and an expert manages that consciously rather than letting service quality visibly sag at that table.
• Escalation threshold: when a table has gone more than ~8 minutes since the last check-in with no proactive touch, treat that as the priority table over whichever table is currently the loudest.
• The 51% rule (Danny Meyer): technical service skill can be taught in a single shift; the instinct to read a table's mood and warmth cannot be trained as fast — when hiring or self-assessing readiness for a section, weight that instinct over polish.

Decision framework

1. Apply the escalation threshold before responding to anything else. Scan every table's course stage and time-since-last-touch, and address the most-overdue table first.
2. Communicate the fire sequence to the kitchen before it becomes a crisis, not after a ticket is already late — an explicit hold-and-stagger call to the expo beats letting four tickets print at once and hoping the line absorbs it.
3. Reset guest expectations proactively wherever a delay is real. A table told "your entrées are five minutes out" recovers tip percentage far better than the same delay discovered in silence.
4. Protect the highest-value recovery lever first. A stalled or visibly unhappy table gets visible attention before an on-track table gets a refill it doesn't yet need — attention is the scarce resource in a loaded section, not effort.
5. When a delay is genuinely unrecoverable, escalate the comp decision to a manager rather than promise one unilaterally — but give the table an honest heads-up immediately so they aren't guessing while the decision is made.
6. After the rush, audit your own sequencing, not just the guest recovery. Note which stage of the section repeatedly collided (usually app-to-entrée fire) and adjust the habit before the next shift, rather than waiting for a manager to flag the pattern.

Tools & methods

• POS course-fire / hold functions and the kitchen display system (KDS) that timestamps ticket-to-fire and fire-to-ready.
• The reservation book (OpenTable/Resy-style pacing view) to see how tightly the host has stacked a section before the rush starts.
• ServSafe Allergens certification and the house allergen matrix — the reference document, not memory, for cross-contact risk.
• Tip-out settlement sheet (percentage-of-sales or percentage-of-tips model, house-specific) reconciled at shift close.

Communication style

To the kitchen expo: terse, sequenced, and directive — "holding fire on 14 and 15, sending in three minutes" — never a vague "whenever you get a chance." To guests: warm, specific, and proactive about real delays rather than reassuring platitudes. To a manager: factual and numeric — table number, time elapsed, what was already done — because a comp or recovery decision above the server's authority needs the facts, not the emotion.

Common failure modes

• Reading a table's chattiness as satisfaction — a talkative, friendly table can still be running long past its healthy turn time, and a quiet table isn't necessarily a neglected one; the clock and the course stage are the signal, not the guest's mood.
• Firing entrées the instant they're ordered to "keep things moving," which is exactly what causes the pile-up when a section was stacked too tight — and the overcorrection, holding every fire defensively to "control" the section, causes simultaneous empty-plate waits at every table instead.
• Running one upsell script on every table regardless of the read, which reads as scripted within a night and suppresses trust and check average both.
• Treating tip-out as adversarial and shorting support staff — the short-term tip gain is smaller than the cost of bussers and runners deprioritizing that server's tables on the next rush.
• Verbal-only allergy handling — telling the guest "I'll let them know" without a ticket modifier or direct word to the kitchen.

Worked example

Situation. Casual-upscale dinner service. Maria's section is four 4-tops (16 covers): tables 11, 12, 14, 15. The host, working a wait-list, seats all four within a 7-minute window (6:00, 6:02, 6:05, 6:07) instead of the 15-20 minute stagger that lets a server run one table's rhythm ahead of the next. All four order apps within minutes of each other and apps clear at roughly 6:20, 6:24, 6:28, and 6:31.

Naive read. Fire each table's entrée as soon as its apps clear, since that's "as ordered" and keeps every table moving. That sends four entrée tickets to the line within an 11-minute window (6:20-6:31).

Expert reasoning. The sauté/grill station on this line can run about two entrée tickets cleanly per 5-6 minute window before ticket time exceeds 12 minutes and plates start arriving unevenly hot. Four tickets in 11 minutes is roughly 1.8 tickets per 5-minute window against that ~2-per-window ceiling — close enough to the wall that any one slow protein (table 14 orders a well-done steak, a 4-minute-longer cook) pushes the whole run over. Maria calls the sequence to the expo instead of letting the printer decide: fire 11 and 12 immediately at 6:20/6:24 (they cleared first and neither ordered a slow protein), hold 14's fire until 6:26 with an early flag on the steak so it's not the tail ticket behind three others, and hold 15 until 6:31 as originally timed since its order is two salads and a pasta that cook fast regardless of queue position. She also tells 14 and 15 directly: "kitchen's running a few tables at once right now, I've got your order in, you're in good shape" — a 15-second proactive touch that removes the silence a stacked section would otherwise create.

Reconciling the numbers. Each table checks at $160 (4 covers x $40 average check, in line with this restaurant's $28 average entrée plus apps/dessert/beverage). Section total for the one turn: 4 x $160 = $640. At Maria's average tip rate of 19%, that's $121.60 in tips. House tip-out model here is percentage-of-tips: 15% to bussers, 5% to food runners/bar, 20% total — $24.32 out, leaving Maria $97.28 net for the turn. Against the federal tipped-wage math: an 8-hour shift at the $2.13/hr direct cash wage is $17.04 in wages; federal law requires direct wage plus tips to reach at least $7.25/hr ($58.00 for 8 hours), and $17.04 + $97.28 from this one turn alone already clears that floor with the rest of the shift still to come — the tip credit is legal here because tips make up the gap, which is the reconciliation a server (or a manager checking payroll) actually needs to run, not just assume.

The actual deliverable — Maria's call to the expo, quoted: *"Firing 11 and 12 now, two out. Holding 14 four minutes behind — steak's well-done, flag it early so it's not the last ticket in. 15 stays on original timing, it's a fast table."*

Going deeper

• references/playbook.md — section-sizing tables, course-fire timing windows, tip-out settlement worksheets, and an upsell script bank by table-read type.
• references/red-flags.md — smell tests for pacing overload, tip-math discrepancies, and allergen-handling gaps, with the first question and the data to pull.
• references/vocabulary.md — working vocabulary generalists misuse, with practitioner usage and the common misuse for each term.

Sources

• Danny Meyer, *Setting the Table* (HarperCollins, 2006) — the 51%/hospitality-vs-service distinction and hiring for warmth over polish.
• Will Guidara, *Unreasonable Hospitality* (Optimism Press, 2022) — reading the table, proactive service recovery, and the Eleven Madison Park "table 4" story as the template for a generosity moment that isn't scripted.
• Michael Lynn, Cornell School of Hotel Administration — published tipping-behavior research on server actions (introducing by name, repeated touch, suggestive selling) that measurably move tip percentage.
• U.S. Department of Labor, Fair Labor Standards Act §3(m) and the 2021 "80/20/30" dual-jobs final rule — federal tip-credit mechanics, the $2.13/hr federal direct cash wage, and the tip-credit-eligible vs. non-tip-credit-eligible work split; state law frequently overrides the federal minimum (several states require full minimum wage before tips).
• National Restaurant Association and Toast POS restaurant benchmarking reports — check-average, table-turn-time, and upsell-attach-rate industry ranges cited as stated heuristics rather than a single authoritative figure.
• FDA Food Code and the FASTER Act (2021, effective January 2023) — the "Big 9" major food allergens (sesame added to the prior eight) and the basis for ServSafe Allergens certification and documented-disclosure practice.

View SKILL.md source on GitHub · maturity: draft

Jurisdiction: US (baseline)