Daily Operations

Pick Toast Core ($69/mo + 2.49%+15c processing) or Lightspeed Restaurant Essentials ($59/mo + 2.6%+10c) — those two split the NYC mid-market 50-200 seat segment. Toast wins if you need the deepest US ecosystem (Resy native, DoorDash native, 7shifts native, Toast Payroll bundled) and 24-72hr deployment; pick it for bars, pickup-heavy, and fast-casual. Lightspeed wins if you want hardware portability (BYO iPad, no proprietary lock-in) and stronger coursing/sommelier features — pick it for wine programs and hotel F&B. Square for Restaurants Plus is the budget alt at 2.6%+10c but lacks router-redundancy at base tier. Skip TouchBistro unless you're committed to iPad-only fine-dining; skip Clover (Fiserv-locked hardware swaps when you change banks). Budget $5-10K hardware for 3 terminals + KDS + 2 handhelds.

Budget three buckets. Hardware $5-10K upfront for a mid-market full-service buildout — 3 Toast Flex terminals at $799-1199 each, 2 Toast Go 2 handhelds at $409-699, 1 KDS at $450, printer $200-500, cash drawer $80-200. Software $69-300/mo for the base subscription, climbing to $1,000+/mo once online ordering, loyalty, payroll, and marketing add-ons stack. Processing 2.49-2.6% + 10-15c card-present is the flat-rate norm; cross $50K/mo card volume and interchange-plus (Heartland, Aloha) typically beats flat by 20-40 bps. Toast hiked rates ~$0.10 per $100 in card sales on Mar 20 2026 — multi-unit operators saw $200-500/mo per location creep. NY GBL §518 prohibits surcharges above your actual processing cost; cash-discount programs are legal, pure card surcharges are not.

Add KDS the day you open if you have a multi-station kitchen — cold/hot/grill/expo — because ticket-misread errors drop 60-80% and per-station ticket-time analytics only exist on KDS, never on paper. Single-station coffee shops or 30-seat counter spots can stay on a kitchen printer until covers cross ~50/turn. The math: $25-79/mo per screen + $400-800 hardware pays back in 6-12 months on faster expo (~30 seconds/ticket on multi-station) and reduced comp/refire costs. Toast KDS ($25-50/screen/mo + $700 hardware) is the default for Toast operators; Fresh KDS ($19-79/mo unlimited screens) is the POS-agnostic pick on Square/Clover/Lavu. Use commercial-grade Elo or kitchen-rated screens (IP54+) — consumer iPads near a fryer or salamander fail in 12-18 months.

Use Toast Online Ordering if you're on Toast — it's the deepest commission-free pickup/delivery flow tied to your menu, with Toast Delivery Services (Doordash drivers via Toast) at flat $0.99/order vs DoorDash Storefront at 15-30% commission. ChowNow ($149-399/mo flat, no per-order fee) wins if you want a true commission-free white-label site and you're processing 200+ orders/wk — break-even vs Toast Online Ordering is around $30K monthly online sales. DoorDash Storefront is free to enable but charges 15% delivery + 6% pickup commission — only use it as a top-funnel acquisition channel, never your primary site. The right stack for most NYC independents: native POS online ordering (Toast/Square) for direct + DoorDash/UberEats/Grubhub via Otter ($99-499/mo) for marketplace exposure. Never run three DSP tablets — single aggregator or you'll lose $30-50/wk to missed/duplicate orders.

Yes — your POS must talk to reservations or your hosts retype guest data and your service team has zero guest context tableside. The Feb 2026 Resy-Toast partnership made Resy bookable inside Local by Toast and surfaces Resy guestbook notes on Toast handhelds during service — that's the upscale default for Toast operators. SevenRooms has 100+ POS integrations including Toast/Square/Lightspeed/Aloha/MICROS with real-time bidirectional sync — pick it for hotel F&B and 4+ unit groups. OpenTable syncs cleanly only on Core/Pro tiers ($299-499/mo + per-cover fees) — Basic gets nothing. TouchBistro Reservations and Lightspeed Reservations are POS-bundled (cheap, but lock you to that POS). Toast Tables ($50-75/mo) is the cheapest Toast-native option but has zero marketplace — fine for direct-bookings-only neighborhood spots.

Configure three things in setup. (1) Tip-prompt screen — set 18/20/22% defaults on the customer-facing display; CFD adds 8-15% to average tip percentages per Toast/Square case studies. (2) Tip pool — Toast and 7shifts are the FLSA-compliant gold standard with weighted-points distribution; NY Labor Law §196-d restricts tip pools to customarily-tipped staff (servers, runners, bussers, bartenders) — back-of-house cooks may NOT participate under NY law (stricter than the 2018 federal amendment), exposure is triple damages plus attorney fees. (3) Service charges — mandatory auto-grats on parties of 6+ are taxable as part of the bill AND subject to 8.875% NYC sales tax, and per NYS DOL may NOT be classified as tips for tip-credit purposes; post them to a separate GL account. NYS Wage Theft Prevention Act §195 requires written tip-credit notice signed at hire and retained 6 years — failure invalidates the tip credit retroactively.

Build modifier groups not modifier lists — group by required vs optional, then by single-select vs multi-select. Required-single (e.g., "choose your protein") forces the screen and prevents send-without-selection errors. Optional-multi (e.g., "add-ons") lets the server skip and move. Cap any single screen at 8 visible modifiers — past that, servers slow 2-3 seconds per ticket and tap the wrong row. Use modifier prefixes for 86 communication ("NO onions", "SUB gf bun +$3") so kitchen reads it instantly. Price the modifier in the POS config — never let staff key custom prices, that's the #1 audit risk. xtraCHEF's drag-and-drop visual recipe builder is best-in-class for tying modifier costs to plate margin so you can see when a $0.75 avocado modifier on a $14 plate kills the margin. Re-test modifier flow weekly during the first 60 days; cut anything that adds >2 seconds to ticket entry.

Pull seven reports every Monday morning. (1) Sales by category (food/beverage/delivery commission split) — if your POS posts everything to one "Sales" account, fix the chart-of-accounts mapping today. (2) Labor as % of sales by day-part — target sub-30% full-service, sub-25% counter; flag any shift over 35%. (3) Overtime hours by employee — target <5% of total payroll. (4) Comps and voids by server — anything over 2% of that server's sales is a coaching conversation. (5) Ticket time by station from KDS — flag any station running 8+ minutes average at peak. (6) Top 20 SKUs by mix — kill the bottom 10 quarterly. (7) Modifier attach rate — high-margin modifiers should hit 30%+ attach. Your Daily Sales Recap (Section 33 below) is the input — these weekly reports are the action layer. Toast, Lightspeed, Square Plus, and Restaurant365 all generate these natively; xtraCHEF and MarginEdge layer the recipe-cost overlay.

Run them in parallel for 14 days minimum, never a hard cutover. Day 1: stand up the new POS at one terminal, train one shift; Days 2-14: run both systems with the new POS taking 25/50/75/100% of orders progressively. Export from the old POS at end-of-day daily — Toast/Square/Lightspeed all support raw transaction CSV export, but menu/modifier/combo configs rarely export cleanly so plan a full menu rebuild ($3-7K migration cost typical). Retain read-only access to the old POS for 3 years for IRS lookback (or print monthly summaries to PDF). Reconcile sales tax under both systems — NYS ST-100 quarterly filings need consolidated sales by month, so switch at quarter boundary (end of March, June, Sept, Dec). Toast hardware is non-transferable and worthless when you leave — budget the writedown. Square is most-portable; Clover hardware re-locks to whatever Fiserv acquirer the new POS uses.

Default to Resy for NYC chef-driven 1-3 units — Amex announced the Tock merger Feb 24 2026 collapsing the upscale tier into a single 25,000+ venue Amex platform, and the Resy-Toast partnership delivers guestbook on Toast handhelds. Pick OpenTable Core ($299/mo + $1.00 network covers) only if you're discovery-dependent in a tourist Manhattan corridor — at 250+ covers/night × 6 nights × 50% network mix you'll pay ~$60K/yr but the marketplace fills seats you couldn't fill direct. Pick SevenRooms ($500-2K/mo, no cover fees) for hotel F&B (PMS integration), 4+ unit chains, or any operator with a marketing person who'll actually use the automation — DoorDash-owned now and powers DoorDash Reservations as a new demand channel. Tock is sunsetting summer 2026 — sign Resy directly. Always enable Reserve with Google as a free overlay regardless of platform.

OpenTable: Basic $149/mo + $1.50 network/$0.25 direct per cover; Core $299/mo + $1.00 network/$0.25 direct; Pro $499/mo + same per-cover plus a new 2% service fee on Experiences/prepaid (introduced Jan 2026). Resy: quote-based flat $249-899/mo, zero per-cover fees. Tock: Essential $269/mo + 3% prepay; Premium $399/mo + 2% prepay (consumer app sunsetting summer 2026). SevenRooms: quote-based $500-2K/mo, no cover fees, $1,500-5,000 enterprise onboarding. Yelp Guest Manager: Basic $99/mo (iPad included), Plus $299/mo. Toast Tables: $50-75/mo Toast bundle. Run the math: a 200-cover/night × 6 nights × 50% network OpenTable Core operator pays ~$40K/yr; the same operator on Resy flat pays ~$5-11K/yr but loses 10-20% of network discovery short-term per TableLink April 2026 modeling.

Open 30 days for neighborhood spots, 60 days for mid-market FSR, 90 days for fine-dining tasting menus and destination restaurants — match the window to how far ahead your guests actually plan. Resy/OpenTable/SevenRooms all let you set rolling windows by daypart and tier (e.g., 30-day general, 60-day for VIP guestbook tagged, 90-day for chef's counter tasting). Open the window at the same hour every day — 9am ET is the NYC industry standard so guests can train themselves. Avoid 365-day windows: they fill with no-shows, distort no-show baselines, and lock out walk-ins your forecast actually wants. For FIFA 2026 (Jun 11-Jul 19) and America250 windows, push your window to 90 days minimum and require credit-card holds or deposits on every booking — Mamdani's permit-denial period overlaps and tourist demand spikes 2-3x.

Require a credit-card hold on every booking of 6+ from day one and charge $25-50/seat for casual and $100-250/seat for fine-dining no-shows. For tasting menus and prix fixe $150+, go full prepay via Tock (now) or Resy (post-summer-2026 merger) — Tock's 2-3% prepay fee is cheaper than no-show losses at any restaurant doing $150+ tasting menus and drops no-show rate from 12-18% (no holds) or 3-7% (with holds) to <1%. Industry-standard cancellation window is 24 hours; same-day cancellation forfeits the deposit (or 50% credit-to-future-visit for fine-dining). Group reservations >6 typically require 50% deposit + 50% no-show charge. Disclose the policy in the booking confirmation SMS and email — open rate is 95%+ — plus on the booking page itself; without the disclosure the dispute chargeback wins on any contested no-show fee.

NYC neighborhood spots run 60-70% reservation / 30-40% walk-in as the baseline; bars and brunch spots flip to 30/70 walk-in-heavy. The ratio you actually want depends on covers per turn and dwell time — a 60-seat bistro doing 1.8 turns at 90 minutes/cover wants ~70% booked because walk-ins fill the no-shows and 6:00pm/7:30pm peak slots. Hold back two 4-tops and one 2-top per service for walk-ins — that's the sweet spot to fill no-shows without sacrificing booked covers. Track "walk-in waitlist clear time" via Waitlist Me (free) or Yelp Guest Manager Plus ($299/mo) — if walk-ins consistently wait >25 min you're under-allocating, if <8 min you're over-allocating. Hudson Yards and Times Square corridors flip to 80/20 reservation-heavy because of tourist booking habits; West Village and LES neighborhood scenes push 50/50 because of late-decision diners.

Set a 24-hour cancellation window with tiered fees: $0 if cancelled 24+ hours out, $25-50/seat same-day for casual, $100-250/seat same-day for fine-dining, full forfeiture for tasting-menu prepay (or 50% credit-to-future-visit). Enforce via credit-card hold (Resy/OpenTable/SevenRooms all support; Tock invented true prepay). Document the policy on the booking page, in the SMS confirmation, and in the 24-hour reminder SMS — the SMS reminder alone reduces no-shows 30-40%. Process the charge within 48 hours of the no-show with a clear receipt — anything older than 7 days loses chargeback disputes. For 6+ tops, require 50% deposit at booking and refund only if cancelled 48+ hours out. NYC consumer-protection note: pure surcharges above your processing cost are restricted under NY GBL §518, but legitimate no-show fees disclosed at booking are enforceable. Track no-show rate weekly — anything above 7% with holds in place means your hold isn't actually charging.

Full count every Sunday night before close (or Monday at open if you don't run Sundays) — that's the period boundary your accounting closes against. Spot-count the top 10 highest-value SKUs daily — proteins, premium liquor, wine by the bottle, cheese — those are 70% of your COGS variance. Bar runs a separate full liquor count weekly (Wisk AI, Partender, BarVision automate this in 15-30 min vs 2-4 hr manual); high-volume bars count daily. Walk-in fridge gets a daily quick visual sweep at open for what needs to come out for the day's prep. Use mobile counting apps (MarketMan, MarginEdge, xtraCHEF all have iPad/iPhone apps with Bluetooth barcode scanners) — paper counts re-introduce 5-10% transcription errors. Reconcile theoretical-to-actual variance every Monday — if you're not investigating anything over 2%, you'll bleed 2-5% of COGS to invisible loss every quarter. NYC operators add a monthly 20% sample audit by an outside party (your bookkeeper or owner) for theft deterrence.

Pick MarketMan ($199-249/mo Starter/Growth) for POS-agnostic flexibility and ghost kitchen multi-brand tagging — works with Toast/Square/Lightspeed/Aloha/Clover. Pick xtraCHEF if you're committed to Toast forever — bundled with Toast Plus/Premium tiers, $149-299/mo for the recipe-costing layer plus $1,049 onboarding, with the best visual recipe builder in the category but Toast lock-in (xtraCHEF requires Toast payment processing — typically $200-400/mo more vs negotiated rates). Pick MarginEdge ($297-330/mo flat regardless of unit count) for 5-50 unit mid-market chains keeping QuickBooks underneath — auto-updates recipe costs daily from invoice OCR with claimed 95%+ line-item accuracy. Pick Restaurant365 ($469-689/mo per location) only at 5+ units when you're ready to replace QuickBooks entirely — 60-90 day implementation, $10-30K setup. Pick BlueCart (free for buyer) as the ordering layer alongside any of the above — vendors pay subscription, you order from 125K-restaurant marketplace.

Industry-standard NYC pour cost is 18-22% blended; below 18% means you're either over-pricing (tourist-trap risk) or under-pouring (review risk), above 22% means theft, over-pour, comps drift, or wrong cocktail recipe costing. By category: house liquor 16-20%, premium liquor 20-24%, craft cocktails 18-22% (the spec is the lever — 2.5oz vs 2oz on a $16 cocktail moves pour cost 3 points), wine by the bottle 25-32% (wine carries the highest pour cost because guests pay for label not pour, so 2.5x markup is the floor and 3x is healthy), beer 18-22% (draft 16-18%, bottle 22-26%). Bar specialty tools — Wisk AI ($165+/mo), Partender ($249-749/mo iPhone tilt-detection), BarVision ($300-1500/mo RFID pour spouts) — typically pay back in 60 days at 25%+ pour cost. Hotel bars run higher 24-28% because of complimentary pours and event service. Track variance weekly, never quarterly — pour cost erodes invisibly.

Industry-standard NYC food cost is 28-32% blended for full-service; fine-dining tasting menus run 30-38% (high ingredient quality + low yield trim factors), casual full-service 28-32%, fast-casual 25-30%, pizza 20-25% (cheese and dough are cheap, the labor is the hidden cost), coffee 18-22%. Above 35% on full-service means recipes aren't costed correctly, portion drift, theft, or you're running specials at zero margin. Below 25% means you're either over-priced for the market or under-portioning (Yelp/Google review risk). Prime cost (food + beverage + labor) is the master KPI — target sub-60% NYC full-service, sub-65% fine-dining; over 65% means you're losing money on every plate. 2026 volatility makes the benchmark a moving target: coffee at $4.30/lb (BLS CPI shock 2025), HPAI-driven egg shortage, cocoa peak $12,500/MT, NYC foie gras ban reinstated March 2026 — recost menus quarterly minimum, monthly for high-volatility ingredients.

Build recipes in xtraCHEF, MarketMan, or MarginEdge with three layers: (1) ingredient-level cost from invoice OCR — auto-updates daily as invoices land in MarginEdge or xtraCHEF, never trust a static spreadsheet (5-15% line-item OCR errors at setup, declining to 2-5% steady state once item dictionaries train). (2) Sub-recipes for components used in 2+ menu items (house aioli, demi-glace, simple syrup) — sub-recipes must be costed before parent recipes or margin reports silently break. (3) Yield/trim factors per protein — verify per kitchen (a 70% trim yield on a tenderloin in your kitchen may actually be 64% — that's a 9% recipe cost error before you even plate). Cost in cents not dollars — $0.04 per garnish times 60 covers/night times 365 days is $876/yr per item. Re-cost every recipe quarterly minimum and any time a key ingredient moves >5% in price. Cross-check theoretical food cost against actual depletion weekly — gap should be <2%.

(1) Recipe accuracy first — sub-recipe missing or stale, yield factor off, or a modifier added without recipe update. xtraCHEF and MarginEdge both flag recipes that haven't been updated in 90 days; that's typically the answer. (2) Theft and over-pour second — pull the bar's pour-cost variance separately from kitchen variance (Wisk/Partender/BarVision will show you per-bottle variance), pull comps and voids by server (anything over 2% of that server's sales is a flag), and check the back-door receiving log against invoices for receiving fraud. (3) Waste and spoilage third — pull the waste log; if it's blank that IS the answer (your team isn't logging, so 2% is hiding in the trash). NYC kitchens systematically under-log spoilage by 30-50% vs reality. Other suspects in order: portion drift (re-weigh 5 random plates of your top sellers tonight), receiving errors (vendor short-shipped or substituted), and POS modifier mapping (a free modifier that costs you $0.75 unrecorded). Investigate within 48 hours of close — cold trail is unsolvable.

Log every waste event in your inventory platform's waste module — MarketMan, MarginEdge, R365 all have native waste logs with reason codes (spoilage, prep error, comp, breakage, expired). Make it mandatory at end-of-shift with manager sign-off; if it's optional, your team logs nothing and you bleed 2-5% to invisible loss. Categorize tightly: prep-error waste means recipe training, spoilage means par-level or rotation problem, comps means service or quality problem, breakage means equipment or training. NYC compliance: LL 199/2023 requires source separation of organics for businesses with 7+ employees (effective in tiers through 2024-2025) — all full-stack inventory platforms satisfy the recordkeeping side. NYS Food Donation & Food Scraps Recycling Law currently sets 2 tons/wk threshold for designated generators; expands to 1 ton/wk Jan 1 2027 — start logging waste now to size your composting hauler contract. 86 lists go in your KDS or scheduling-app team chat (7shifts has photo-attached shift swap) so FOH sees them in real-time, not on a paper next to the host stand.

An item that 86s more than once a week means one of four things: par level too low, shelf life shorter than your usage cycle, vendor unreliability, or recipe mix shifted and you didn't update pars. Pull the 86 log against the inventory platform's reorder report — if you're ordering enough product but still 86ing, your par calculation is wrong (typical fix: par = lead time × average daily use × 1.5 safety factor, not the eyeball estimate from open). If you're consistently 86ing fresh fish or specialty produce, your vendor's delivery cadence doesn't match your menu cadence — switch to a 5-day vendor (Baldor, Greenleaf) or move that SKU to a 2x/wk delivery schedule. If 86s cluster on one prep cook's days, it's a yield problem (over-trimming) or theft. Set a hard rule: any item that 86s 3+ times in a month gets a recipe + par + vendor review the next Monday. Operators who tolerate frequent 86s lose 8-12% of potential revenue on covers walking out, plus Yelp/Google reviews drop 0.2-0.4 stars within 60 days.

Pick 7shifts The Works ($76.99/loc/mo) for 95% of NYC independent and mid-market full-service operators — best-in-class FLSA-compliant tip pool with weighted points, deepest Toast/Square/Lightspeed/Aloha API integration, and 4.7 Capterra rating. Pick HotSchedules/Fourth ($40-100/mo per location, custom at scale) for 30+ location fast-food chains needing NYC Fair Workweek automation — purpose-built FWW module with 14-day scheduling, predictability pay, and clopen detection. Pick Sling (Toast-owned, free or $1.70-3.40/user) only if you're a Toast operator wanting bundled simplicity and you have no tip pool complexity. Pick Toast Tap-In (bundled with Toast POS) for the same reason — newer, feature-light vs 7shifts but zero integration friction. Skip Homebase Free for NYC FSR — fine for sub-25 staff coffee shops but no FWW automation and weak tip pool. Skip When I Work entirely if you have a tip pool — no tip module.

Feed three years of POS sales data into your scheduler's forecasting engine — 7shifts AI Auto-Scheduler (launched 2024), HotSchedules Forecasting (5+ year POS data, the gold standard for enterprise), or Restaurant365 native forecasting all work. The model needs day-of-week + week-of-year + weather + holiday/event flags + reservation count from Resy/OpenTable. Build labor budget as % of forecasted sales — target 28-32% full-service, 20-25% counter, 18-22% fine-dining (BoH-heavy). NYC-specific overlays for 2026: FIFA 2026 (Jun 11-Jul 19) and America250 push midtown corridors +30-50%; Mamdani's permit-denial overlap shuts down street activations during that window. Build the schedule by daypart (lunch/dinner/late) not by full shift — cover-per-labor-hour is the real KPI (target 4-6 covers/labor-hour mid-market, 2-3 fine-dining). Re-forecast every Wednesday for the following week using 7-day actuals; lock the schedule by Thursday for 14-day NYC FWW posting (if you're a 30+ location chain).

Hard-cap weekly hours per employee at 38 in your scheduling app — 7shifts, HotSchedules, Homebase Plus all auto-flag projected OT before publish; never let a manager publish a schedule that projects any employee over 40. Cross-train so you have 2-3 people qualified for every station — single points of failure drive emergency OT. Track OT weekly by manager (not just by employee) — managers who hit 7%+ OT consistently are scheduling lazy or hoarding favorite employees on shifts. Use shift-swap apps (7shifts, Homebase, HotSchedules) with manager approval required for any swap that creates OT. NYS Spread of Hours rule (12 NYCRR 142-2.4) requires extra hour at minimum wage if workday spans 10+ hours — that's not OT but it's an invisible $16.50 hit per occurrence; auto-flag in time clock. NYC FWW chains (30+ national locs) face $10-75 premium per <14-day schedule change plus $100/clopen — those penalties dwarf OT cost so the FWW operator's primary OT control is the 14-day post.

Allow swaps via 7shifts, HotSchedules, or Homebase but require three controls. (1) Manager approval mandatory on every swap — auto-approve creates uncontrolled OT, mismatched skill levels, and FLSA violations. (2) Skill-tag enforcement — bartender swaps only with bartenders, server swaps only with servers, and minor-flag (16-17 yo) swaps blocked from school-night shifts (NYS Labor Law §142.2 caps 4hr school-night, 8hr non-school day; $1K-3K civil penalty per violation). (3) OT projection check — the app must reject any swap that projects either employee over 40 hours that week. NYC FWW chains: any swap within 14 days of the shift triggers $10-75 predictability pay unless the employee initiated the swap and signed a waiver in the app. 7shifts wins on swap UX (photo support for menu change handoff); Homebase is fine for sub-25 staff. Audit swap patterns monthly — same two people swapping every week is a scheduling-mismatch signal.

Run all requests through your scheduler's request queue — 7shifts, HotSchedules, Homebase all have native time-off forms. Set a hard rule: 14-day advance notice for any single-day request, 30-day for 3+ consecutive days, 60-day for any week-plus block. Auto-deny on Friday/Saturday peak service unless approved by GM (front-load this in the request form so staff know before submitting). Cap simultaneous time-off at 2 servers, 1 bartender, 1 line cook, 1 dishwasher per service — bake the cap into the scheduler. NYC Earned Safe and Sick Time Act (LL 46/2017) accrual must be tracked separately and is non-deniable for qualifying use; all major schedulers track accrual. Approve in writing in the app, never verbal — verbal approvals are the #1 source of NY DOL Wage Theft Prevention complaints. Build a 90-day rolling time-off calendar visible to staff so they can self-coordinate. For peak periods (Restaurant Week, holidays, FIFA 2026 Jun-Jul) post a blackout list 60+ days ahead.

Document a written progressive-discipline policy in your employee handbook signed at hire (NYS Wage Theft Prevention Act §195 requires written wage notice retained 6 years — extend the same retention to discipline records). Standard tiers: 1st late = verbal warning logged in scheduler; 2nd within 90 days = written warning; 3rd = final written; 4th = termination. No-shows: 1st = final written warning; 2nd within 90 days = termination. Define "late" as 5+ minutes past scheduled clock-in (give the buffer or you'll lose at unemployment hearings). NY is at-will but NYC LL 1/2021 requires just-cause for fast-food terminations — document, document, document. Pay all hours worked even on the termination shift — withholding final wages is wage theft and exposes you to NYS DOL $1K-$20K penalty plus liquidated damages. NYS Call-In Pay (12 NYCRR 142-2.3) requires 4hr min pay if employee reports and is sent home; if you send a chronically-late employee home as discipline, you owe the 4hr. Auto-track late/no-show in 7shifts/HotSchedules/Homebase time clock — paper trails win audits.

Hire hospitality-specialized once you cross $1M revenue or 2 locations — generic CPAs miss the FICA Tip Credit (IRC §45B, $3-15K/loc/yr left on the table), miscode service charges as tips, and wreck your COGS chart-of-accounts. NYC hospitality specialists: Paperchase Hospitality (35+ years, $1.5-7.5K/mo, LAVO/Jean-Georges client base), Presti & Naegele ($2.5-10K/mo CFO), DCG (Clutch's #1 NY fractional CFO Apr 2026). Cloud-native SMB tier: Bench, Pilot, Acuity at $300-2K/mo for single-location. The build-vs-buy math: in-house bookkeeper at $55-75K salary fully loaded is $83-121K/yr (benefits 25-30% + payroll tax 7.65% + software/equipment $3-5K + management time); outsourced at $2,500/mo = $30K/yr — typically 50-70% cheaper for sub-5-location groups. Internal bookkeepers turn over every 18-24 months in NYC restaurants — each turnover costs 4-8 weeks of book continuity. Hire the firm before you need them; switching mid-crisis costs double.

Restaurant365 makes sense at 5+ locations OR when you've outgrown QuickBooks at 3-4 units within 24-month growth horizon — it's a full GL replacement, not a layer, with consolidated multi-location P&L, 13-period (4-4-5) fiscal calendar, EDI to Sysco/USFoods, native AP automation, and inventory + recipe + scheduling + payroll bundled. Pricing: Core $249/mo/loc, Essential $469-499/mo/loc (quarterly billed), Professional $635-689/mo/loc, plus 60-90 day implementation at $10-30K. All-in for a 4-loc group: $30-50K/yr. Compeat is end-of-life — R365 acquired April 2021 and is migrating customers; new Compeat sales effectively stopped. Stay on QuickBooks Online Plus ($115/mo) + MarginEdge ($330/mo flat) under 5 locations — you keep your CPA, period-close compresses to 7-10 days vs R365's 4-7 days, and total stack is $20-30K/yr. Don't migrate to R365 without a controller-level person on staff — stalled implementations are common when GMs underestimate the chart-of-accounts and 13-period close workload.

Look at the daily sales recap every morning, the labor + food cost variance every Monday, and the full P&L every Wednesday for the prior week. Monthly closes are too slow — by the time you see a 4-point food cost spike on a 14-day-old P&L, you've already burned 10-14 service days at the wrong margin. Period-close benchmarks: QuickBooks alone closes 14-21 days post-period (too slow); QuickBooks + MarginEdge compresses to 7-10 days; Restaurant365 native to 4-7 days; Sage Intacct + R365 to 3-5 days. Operators who can't close within 10 days systematically make decisions on stale data. Beyond weekly, run a 13-period (4-4-5) fiscal calendar instead of monthly — restaurants run on weekly cycles and 13-period calendars give you comparable Mondays and Saturdays across periods (a calendar-month comparison gets distorted by 4 vs 5 Saturdays). Quarterly: full balance sheet review with your CPA. Annually: tax-strategy session before fiscal year-end (FICA Tip Credit, Section 179 equipment depreciation, R&D credit if applicable).

The DSR has nine lines and is emailed to ownership by 11am the next morning. (1) Net sales by category — food, beverage (split bar/wine/beer), retail, gift card, third-party delivery (DoorDash/UberEats/Grubhub net of commission). (2) Covers and average check by daypart — lunch/dinner/late. (3) Comps and voids in dollars and as % of sales (flag anything over 2%). (4) Tip total — cash tips + credit card tips + service charges (separate column — service charges are taxable revenue per NYS DTF, not tips). (5) Labor hours and labor % of sales (target 28-32% full-service). (6) Sales tax collected (8.875% NYC — 4% NYS + 4.5% NYC + 0.375% MCTD per NYS DTF). (7) Cash deposit total (and any over/short variance). (8) Top 5 sellers and bottom 5 by mix. (9) Reservations — booked vs walked vs no-show count. Toast, Square Plus, Lightspeed, Restaurant365, MarginEdge all generate this natively; if yours doesn't, build it as a memorized report. Review every morning before 11am — anomalies cool fast.

Run dual-count on every drawer at open and close — both counters sign the count sheet. Mid-shift drops every $300 cash or every 4 hours, whichever comes first — drop into a slotted safe that managers can't open without the GM (or weekly armored pickup via Loomis/GardaWorld at $40-80/pickup). Daily bank deposit by noon next morning, never overnight in the safe past 24 hours (insurance policy cash limits typically cap at $5-10K on premises). Surveillance: 4K cameras with audio over every register and the safe, 90-day cloud retention minimum (NYC SLA-licensed venues require this anyway), positioned to capture both the drawer face and the cashier's hands. Reconcile cash to POS every shift — over/short variance over $10 gets investigated, over $50 documented and signed. NYC sales tax 8.875% must be remitted via NYS ST-100 quarterly (or monthly if $300K+ tax due) — automate via DAVO ($54.99/mo/loc, sweeps daily and auto-files) to eliminate the #1 single-location filing error. Rotate counters monthly so no one person has month-long drawer access.

Monthly: NYS ST-100 sales tax if you're $300K+ tax due (most NYC restaurants are quarterly, but high-volume venues hit monthly), federal payroll tax deposit (Form 941 quarterly but deposits semi-weekly or monthly via EFTPS depending on liability), NYS NYS-45 quarterly is per-quarter but withholding deposits are often monthly. Quarterly: NYS Form ST-100 (due 20th day after quarter — Apr 20, Jul 20, Oct 20, Jan 20; e-file mandatory $100K+ tax due), Form 941 federal payroll, NYS NYS-45 unemployment + withholding, NYC CR-Q commercial rent tax (6% on commercial tenants south of 96th St Manhattan paying base rent over $250K/yr after 35% adjustment). Annual: federal Form 1120/1120-S/1065 income tax due Mar 15 (S-corp/partnership) or Apr 15 (C-corp), NYS CT-3-S corporate franchise, NYS IT-204-LL LLC fee ($25-$4,500 graduated by NY-source gross income, due Mar 15 calendar-year), NYC UBT 4% if partnership/LLC over $95K unincorporated income, federal Form 8846 FICA Tip Credit (recovers $3-15K/loc), W-2s by Jan 31, 1099-NEC by Jan 31. NY S-Corp election Form CT-6 due by Mar 15 of election year. Late ST-100 penalty 10% of tax due (min $50) plus 1%/mo interest — automate via DAVO.

Match processor to monthly volume, not brand recognition — Toast and Square hiked rates in 2026 (Square jumped online to 3.3%+30¢ on the free plan Jan 13 2026; Toast tripled fees on incumbent contracts). For a single-loc independent under $60K/mo, Square free (2.6%+10¢ in-person, 3.3%+30¢ online) or Helcim ($0 monthly, IC+0.4%+$0.08) is the right answer. For mid-market FSR $80K-$250K/mo or fine-dining, you want interchange-plus — Heartland negotiated to ~0.30% markup beats Toast flat by $7,800-$30K/yr. Stripe is for catering platforms and ghost kitchens (CNP-heavy, Connect for marketplaces). Clover (Fiserv) is portable hardware but the all-in ~3.0% rarely wins on math; default to it only if your bank channel forces it.

Run interchange-plus math, not flat-rate marketing copy — restaurant blended interchange runs ~1.8% (food MCC favorable) so any all-in over 2.5% means you're funding the processor, not the network. Realistic 2026 NYC effective-rate benchmarks: $50K/mo volume on Toast all-in ~3.4% vs Heartland IC+ ~2.5%; at $250K/mo (busy NYC bar) Toast ~2.9% vs Heartland ~2.2% — that's $15K-$30K/yr on the same throughput. Add 25-35% bar tip mix on top of the bill subtotal — both Toast and Square charge their percentage on the tip-inclusive amount, which is why bar/nightclub operators pay disproportionately. Pull your last three monthly statements, isolate effective rate (total fees ÷ gross processed), and if you're north of 2.7% at >$80K/mo you're leaving money on the table. Premium-rewards-card mix matters — Visa Signature interchange runs 2.10% IP, so Amex-heavy fine-dining clientele needs IC+ visibility or the markup hides.

File SAQ-A annually if you're on Toast, Square, Clover, Lightspeed, or any PCI-validated cloud POS — that covers 95%+ of NYC restaurants and is ~22 controls vs. SAQ-D's ~329. Buy SecurityMetrics or ControlScan ($99-$499/yr) for the SAQ wizard, quarterly ASV scan, and attestation; total all-in ~$300-$500/yr. PCI DSS v4.0.1 is fully enforceable as of March 31 2025 — the biggest operational change is MFA on ALL access to cardholder systems, not just admin (Req 8.4), plus quarterly ASV scans now strict. Self-hosted Oracle MICROS, legacy on-prem POS, or any operator-built ordering portal that touches card data drops you into SAQ-D (full QSA assessment $15K-$50K/yr) — avoid. Level 1 (>6M Visa transactions/yr) requires a QSA-led ROC; that's 50+ location chains, not single-loc operators. Non-compliance fees from acquirers: $5K-$100K/month.

Respond inside the processor's window — typically 7-10 days from chargeback notice — or you auto-lose. The evidence stack that wins: signed receipt or EMV-chip authorization log (chip-and-PIN beats signature-only), itemized check showing dates and items, surveillance video clip from your camera system tied to the transaction timestamp (Verkada/Solink/Envysion will pull this with POS overlay), reservation record with phone/email match, server's contemporaneous notes, and any text/email confirmation from the guest. For card-not-present (catering invoices, deposits), AVS+CVV match plus signed contract is mandatory. Chargeback fees run $15-$50 per dispute and a chargeback ratio above 1% triggers Visa's Dispute Monitoring Program (VDMP) — fines, then termination. If you're a bar/nightclub running MCC 5813 with chronic chargebacks, you'll get flagged off mainstream processors and have to land at Corepay or PaymentCloud high-risk.

Yes — PCI DSS v4.0 explicitly requires the cardholder data environment (CDE) be segmented from non-CDE networks, and failure to segment expands PCI scope to every connected system, turning a SAQ-A operator into SAQ-D overnight. Stand up at minimum three VLANs on a Cisco Meraki MX or Fortinet FortiGate: VLAN 10 POS/CDE (terminals + KDS + receipt printers, no internet egress except processor), VLAN 20 BoH (office laptops, scheduling, accounting), VLAN 30 guest WiFi (captive portal, rate-limited, isolated from everything). IoT (cameras, kitchen displays, thermostats) goes on a fourth VLAN — never bridge cameras and POS. Cyber-insurance carriers (Travelers, Coalition, AIG) now run periodic policyholder scans and will deny claims or non-renew if guest WiFi shares a subnet with POS. Budget $1,500-$5K for proper Meraki/Fortinet install via an MSP; this is not the line item to cheap out on.

The dominant attack on NYC restaurants is fake-vendor-invoice business email compromise (BEC) — attackers spoof Sysco, US Foods, or beverage distributors and email the operator a 'updated wire instructions' notice the week of a real invoice. Second most common: the GM-impersonation gift-card scam targeting hosts/managers. Defenses: (1) phishing-resistant MFA with FIDO2/passkeys on all email + back-office accounts (PCI 4.0 Req 8.4), (2) email gateway — Microsoft Defender for Office 365 if you're on M365, otherwise Mimecast or Proofpoint, (3) staff training via KnowBe4 ($3-$5/user/mo), Hoxhunt, or Microsoft Attack Simulation — for 25 staff that's $1K-$2K/yr, (4) a hard rule that ANY change of payment instructions requires a phone call to a known-good number, never a reply-to-email confirmation. Coalition and Travelers discount premiums for proof of quarterly phishing tests.

Architect three VLANs minimum on cloud-managed gear — Cisco Meraki MX/MR or Ubiquiti UniFi are the two NYC standards. VLAN 1 POS (terminals, KDS, kitchen printers, receipt printers — no internet except processor allowlist), VLAN 2 BoH (office laptop, scheduling, accounting, MFA-protected), VLAN 3 guest WiFi (captive portal, client isolation on, rate-limit 10-25 Mbps/user, blocked from all internal subnets), and a fourth VLAN for IoT (cameras, thermostats, music). Wi-Fi 6/6E APs minimum (2.4/5 GHz congested in dense NYC blocks); plan one AP per ~1,500 sqft, more in 100+ seat dining rooms or basement BoH. Cisco Meraki dashboard is the operator default for multi-site visibility — Ubiquiti is cheaper ($200-$500 vs $800-$1,500 per AP) and works fine single-loc with a competent MSP. PCI segmentation is non-negotiable; budget $5K-$25K install for a 100-seat new build, and pre-opening engagements like BlueSpire, Hospitality Tech Corp, or Olmec do this end-to-end.

200/100 Mbps fiber is fine for 90% of 100-seat NYC restaurants — POS transactions are tiny, KDS is local, the load is staff phones plus guest streaming. Where it falls short: live-streaming events (chef demo, weddings), cloud video upload from 8+ Verkada/Rhombus cameras (each 4K cam pushes 4-8 Mbps continuous), or hotel F&B with simultaneous voice + video conferencing. Real spec: 200-500 Mbps symmetric fiber primary (Verizon Fios Business, Spectrum Business, Pilot Fiber NYC), bonded LTE/5G failover via Cradlepoint or Peplink ($50-$150/mo) so a fiber cut doesn't kill Friday dinner. For events with 500+ attendees needing dedicated bandwidth, that's a Trade Show Internet or Made By WiFi overlay — $1K-$3K for the day, separate from house circuit. Reserve 30% headroom on guest VLAN; rate-limit per-client to 25 Mbps so one Netflix stream doesn't dominate.

Keep a retainer the moment you cross 2 locations or 60+ seats with a Toast/Lightspeed POS, KDS, and kitchen printers — break-fix at $75-$200/hr looks cheap until your POS goes dark at 7:45pm Friday and the freelancer doesn't pick up. NYC retainer benchmarks: per-device $25-$100/device/mo, per-user $99-$200/user/mo (tier-1 NYC spec $185-$275), or fully-managed flat-rate $1,500-$10K+/mo for chains and hotels. Single-loc full-service 40-80 seats is the sweet spot for $1,500-$2,500/mo small-MSP retainer (Heroic Technology, Setton, TechPulse) or a $2K block-time prepaid bucket. Below that scale — 30-seat coffee shop, single Toast terminal — Toast Care + Geek Squad emergency call is the right answer; total IT under $100/mo. Above $5M revenue or 3+ locations, expect bundled EDR, email security, MFA, and Datto BCDR backup baked in; MSPs that don't bundle are losing 2026 RFPs.

POS is backed up by your processor — Toast, Square, Lightspeed all retain transaction data cloud-side, you don't manage that. What YOU back up daily: (1) accounting (QuickBooks Online, Restaurant365 — both cloud-native, just verify retention settings), (2) scheduling and labor (7shifts, Harri, Crunchtime — cloud), (3) email and Drive/SharePoint files via Microsoft 365 Backup or Datto SaaS Protection ($3-$8/user/mo — native M365 retention is NOT a backup, it's 30-90 days of recycle-bin), (4) any local files on the back-office PC via Datto BCDR with immutable storage, (5) photography/marketing assets on Google Drive or Dropbox Business with version history. RTO/RPO targets: 4-hour RTO, 1-hour RPO is the modern standard. Cyber-insurance carriers (Coalition, Travelers, AIG) deny claims when backups aren't tested — schedule a quarterly restore drill and document it. Average ransomware recovery without strong BCDR is 7-14 days; with Datto + tested restore, sub-24 hours.

Match phone system to operator size, not brand. Solo or 1-3 person team — OpenPhone (now Quo) at $15/user/mo or Phone.com at $13/user/mo covers auto-attendant, SMS, voicemail-to-email. 5-30 employee single-loc — Dialpad $15/user/mo (AI included), Zoom Phone $10-$15/user, or Ooma Office. Multi-unit chain or fine-dining with reservation desk — RingCentral $20-$50/user/mo wins on Resy/OpenTable/SevenRooms integrations and AI Receptionist. Hotel with PMS integration — RingCentral or Mitel MiVoice Cloud. Toast Phone is fine if you're already deep on Toast and want one bill, but the AI receptionist quality lags RingCentral and Dialpad. Compliance non-negotiables: Kari's Law and RAY BAUM's Act §506 (47 CFR §9.10) require dispatchable-location 911 on every multi-line system installed since Jan 6 2021 (fixed) / Jan 6 2022 (softphone), and 10DLC carrier registration via The Campaign Registry is mandatory for any A2P SMS — reservation confirmations included — or the carriers filter you. Expect $1.50-$4/user/mo in 911/regulatory recovery fees on top of the listed rate.

Spec 6-12 cameras for a 100-seat single-loc covering entry/exit, bar, host stand, cash drawer, kitchen pass, back door, dry storage, walk-in cooler entry, employee entrance/lockers — and never put audio or any camera inside a bathroom (NY law). 30-day retention is the floor; 90 days is operator-grade for slip-and-fall and chargeback defense. Cloud-managed (Verkada $499-$5,299/cam hardware + $150-$500/cam/yr license, Rhombus $700-$1,500/cam + $200-$400/yr, Eagle Eye $50-$150/cam/mo bring-your-own ONVIF) is the modern default — single-pane multi-loc dashboard plus AI analytics (person/vehicle/face detection, POS overlay for refund/void/comp forensics). On-prem NVR (UniFi Protect, Hikvision, Dahua) is cheaper at $2K-$5K all-in zero subscription but Hikvision/Dahua are NDAA Section 889-flagged — cyber carriers increasingly exclude them and chain franchisors mandate NDAA-compliant brands. Total spend for a 100-seat: $8K-$18K install + $1.5K-$5K/yr cloud licensing. Bar/nightclub low-light needs Verkada CD53/CD63 zoom or Axis Q-series — 4K + IR + WDR, no consumer-tier.

Match CRM to operator persona, not feature list. Fine-dining / reservation-driven (60-200 covers, $$$$, allergens + anniversaries + server notes) — SevenRooms is the spine, $750-$2K/loc/mo plus $2K-$10K setup; DoorDash bought it for ~$1.2B June 13 2025 so roadmap clarity is high. Mid-market FSR with Toast POS and a marketing manager — Thanx or Marsello layered on Toast Loyalty. Single-loc independent under $1.5M revenue — Toast Loyalty bundled at $25-$185/mo (the $185 tier includes Loyalty + Email + Gift Cards) or Square Loyalty $45/mo, both POS-native and zero-friction. Bar/nightclub walk-in heavy — SevenRooms with door-host workflow or Thanx with Wi-Fi capture; SMS-first like Postscript or Attentive sometimes wins. Resy CRM v2 (Amex-owned) caught up on segmentation in 2024 if you're already on Resy reservations. Hotel F&B — SevenRooms with PMS connector to Opera/Mews. Avoid Salesforce/HubSpot — horizontal CRMs can't ingest POS transaction-level data the way restaurant-purpose-built systems do.

Tag at the point of recognition, not the point of arrival — train hosts to tag during reservation creation in SevenRooms/Resy/OpenTable, and train servers to push tags back into the profile after each visit. Standard tag taxonomy: VIP (top 5% spend or relationship), Regular (4+ visits/yr), Influencer (verified reach 10K+, comp-and-content protocol), Press, Industry (chef/sommelier comp), VIP-Birthday/Anniversary-this-week, Allergen-X (auto-fires kitchen alert), Wine-Whale ($200+ wine ticket). Triggers that auto-route: VIP arrival pings GM phone (SevenRooms VIP Alerts), Allergen tag fires kitchen ticket banner in Toast/SevenRooms-KDS integration, lapsed-VIP (90+ days no visit) drops into a winback automation. Influencer protocol needs a comp policy in writing — who approves comps, what the content ask is, who tracks deliverables. SevenRooms is deepest on auto-tags and tier-status; Thanx and Punchh handle this on the loyalty side; Toast Loyalty is shallowest and works only if you're committed to Toast end-to-end.

Treat allergens as a single source of truth that flows reservation -> POS -> kitchen ticket — never rely on a server's verbal handoff. In SevenRooms or Resy, capture the allergen at booking (online form + host re-confirms on phone-in), pin it as a permanent profile tag, and auto-print to the kitchen ticket via the POS integration (Toast, Square, Lightspeed all support). On Toast specifically, build allergen modifiers as required prompts on every entree so the line cook physically can't fire the dish without the modifier line. Train every server on the FDA Big-9 (milk, eggs, fish, shellfish, tree nuts, peanuts, wheat, soy, sesame — sesame added 2023) and the kitchen on the dedicated allergen pan/utensil/glove protocol. Post the NYS food allergy notice in both prep area and on menus — required since May 2025 under Chapter 480 of the Laws of 2024, $125 per violation. ServSafe Allergens certification ($22 online) for FoH leads is the documentation insurance carriers and plaintiff attorneys look for after an anaphylaxis claim — the legal exposure for a missed allergen is six- to seven-figure.

Run the LAST acronym at the table — Listen (no interruption, no excuses), Apologize (sincere, not 'sorry you feel'), Solve (comp the dish, remake, manager visit, next-visit credit), Thank (validate them speaking up). Empower the GM with a comp ceiling ($50-$150 single-table) and require Logbook entry within the shift — date, table, server, complaint type, resolution, cost. Recovery escalation: dish issue -> remake + dessert comp; service issue -> manager apology + drink comp; allergen near-miss -> full comp + GM call next day + incident memo to insurance. Follow up within 24 hours by email or SevenRooms message — 'we want you back, dinner on us' beats a generic gift card 3:1 on return rate. Detractors flagged in your CRM (Wisely sentiment, SevenRooms NPS 0-6) should auto-route to GM email BEFORE they post on Yelp/Google — that's the whole point of NPS routing. If the complaint is foodborne-illness, immediately preserve the meal record (POS line items, who else ordered the dish), call your insurance carrier, and notify NYC DOH if multiple cases — Article 81 reporting obligation.

GM owns the response queue, owner reviews and signs off on tone — never outsource to PR for a single-loc operator. Cadence: respond to 100% of negative reviews within 24-48 hours, 100% of detailed positive reviews within 7 days. Tone rules: thank, acknowledge specific item (proves a human read it), apologize without admitting legal liability, offer a private channel (DM, email) to make it right, never argue facts publicly. Use Wisely (Olo Engage), SevenRooms NPS, or Thanx to intercept detractors (NPS 0-6) BEFORE they post — auto-route to GM email so service recovery happens privately. Promoters (NPS 9-10) get auto-prompted to leave a Google or Yelp review — that's the highest-leverage organic-review pump in 2026. Yelp's filter is opaque; don't pay for Yelp ads thinking it'll lift filtered reviews — it won't. Google Business Profile is now the more important SERP surface for restaurant discovery; respond to every Google review and keep posts/photos updated weekly. Influencer/journalist reviews go to ownership only; never engage publicly with a hit piece, address it through PR or counsel.

Schedule NFPA 96 hood and duct cleaning quarterly for solid-fuel/wood/charcoal kitchens, semi-annually for most full-service kitchens, annually for low-volume cafés — and tag it on the calendar 60 days in advance because this is the single most-slipped maintenance item in NYC. Hire an NFPA 96-certified hood cleaner (Done Right HFS, Cintas, AirX, NYC City Wide); contracts run $400-$2K per visit depending on hood length, fan count, and roof access. Separately, the kitchen hood SUPPRESSION system (Ansul R-102, Kidde) needs semi-annual inspection by an FDNY S-17 certified technician — fusible links replaced every visit, wet chemical agent verified — at $350-$750/system. Expired hood-suppression tag is the #1 FDNY violation trigger for NYC restaurants and a $1K-$5K fine plus shutdown risk. Class K extinguisher within 30 feet of any cooking appliance, annual extinguisher inspection by certified tech. Insurance carriers will deny a kitchen-fire claim if the cleaning log isn't current — keep the dated certificate posted in the back-of-house.

Sign a monthly IPM (integrated pest management) contract — never a quarterly spray-and-pray — with an NYS DEC Category 7A certified applicator and a NYC DCWP exterminator license under §20-307. NYC monthly benchmarks: QSR/fast-casual 1,500-3,000 sqft = 1-2 visits/mo at $200-$450; full-service restaurant 3K-8K sqft = 2-4 visits/mo at $400-$1,200; large bar/nightclub 4 visits/mo at $5K-$15K (late-night service). Initial clean-out for a heavy infestation is $500-$2,500 over 2-4 visits before stepping down to maintenance. Vendors that win NYC accounts: Ecolab Pest Elimination, Magic Exterminating, M&M Pest Control, Bell Environmental, Assured Environments — all carry NYS Cat 7A, IPM-led, and produce monthly trend reports. A live mouse or rat in a food prep area during a DOH inspection is 28+ violation points = automatic C grade or closure (Health Code §81.25, 24 RCNY §23-04); Local Law 13 of 2024 plus Mamdani's 2026 Rat Action Plan expansion put RMZ blocks on a higher inspection cadence. Demand: monthly visit log with catch counts, sanitation feedback, and the Cat 7A cert of every tech servicing your account.

Sign with a BIC-licensed commercial waste hauler under NYC Commercial Waste Zones (CWZ) — your hauler is now zone-assigned, not a free market choice as of CWZ rollout. The DSNY/BIC maximum rate cap is $26.87/cubic yard (2026 ceiling) — verify your contract is at or below that; hauler markup above cap is a BIC complaint. Restaurant typical: 2-4 yards/week pickup at $400-$1,200/mo for a 100-seat single-loc, depending on volume, recycling/organics streams, and zone. NYC Local Law 146 commercial organics rule mandates separate organics pickup for any food service establishment generating organic waste — tri-stream (landfill / recycling / compost) is now baseline, not optional. Used cooking oil/grease pickup is a separate licensed hauler — Darling Ingredients, Mahoney Environmental, Rendertech — usually free or paying you per gallon (yellow grease is a commodity market). Grease trap pumping by a BIC-licensed hauler at 30-60 day cycle for high-volume kitchens, never exceeding 90 days; $300-$600/pump, with manifest retained for DEP. DEP requires interceptor cleaning before grease/solids hit 25% of liquid depth — that's the citation standard.

Run a quarterly preventive HVAC service contract — NYC restaurants with rooftop RTUs, walk-in refrigeration, and kitchen makeup-air units cannot survive on break-fix. Cadence: filters every 4-8 weeks (more in heavy fryer/grill kitchens, weekly in extreme cases), full RTU service quarterly (coil clean, belt check, refrigerant pressure, electrical contactor inspection), refrigeration system semi-annual including condenser-coil clean and refrigerant log. Annual EPA Section 608 refrigerant compliance check is mandatory for any system over 50 lb charge — and the EPA AIM Act ER&R rule effective Jan 1 2026 dropped the HFC threshold from 50 lb to 15 lb, so smaller walk-ins are now in scope and need quarterly leak inspections + log. Quarterly preventive contract for a 100-seat single-loc runs $300-$800/mo (RTU + walk-ins); add $150-$300/mo for filter program. Vendors: Arista Air, Hart's Service, P.E. Stone — all NYC-tier with 24/7 emergency. Skip preventive and one summer compressor failure on a Saturday at 4pm is $5K-$15K in emergency labor + lost revenue.

Move to a service contract once a single equipment failure can shut the line — that's the hood suppression (semi-annual S-17 inspection mandatory regardless), walk-in refrigeration, ice machine, dish machine, fryers, combi ovens, and rooftop HVAC. Break-fix is fine for low-stakes gear: prep tables, holding cabinets, hand mixers, anything you can substitute for a shift. Service-contract economics: a Hobart/Ali Group dish machine PM is $400-$800/yr and prevents the $2K-$4K mid-Saturday-rush emergency call; a True/Continental walk-in compressor PM is $600-$1,200/yr and prevents a $5K-$10K compressor swap plus $3K-$8K of spoiled inventory. Vendors: Ecolab and Auto-Chlor for warewashing, Hart's Service and Arista for refrigeration/HVAC, Pine Tree Food Equipment for cooking line. Get every tech's certifications in writing — S-17 for hood suppression, EPA 608 for refrigerant work, NYC DEP backflow tester for any RPZ. Insurance carriers and DOH inspectors both look for service logs as evidence of operator diligence; binder them by equipment, retain 3 years.

Default to nightly contracted crew unless you're under 60 seats with a tight closing team. NYC nightly crew benchmarks: 100-seat full-service ~$2K-$4K/mo for 4-5 nights/week, 2-person crew, 3-4 hours per visit covering FoH floors, restrooms, glass, BoH degreasing of cook line and hood exterior, mop kitchen floor. In-house cleaning loaded to closing crew runs $14-$22/hr loaded × 2 staff × 3 hrs × 30 days = ~$2.5K-$4K/mo PLUS the chemical and equipment program ($200-$500/mo) PLUS the supervision overhead and the reality that exhausted closers do a worse job. Contracted crew (Servmaster, ABM, Anago, USA Cleaning Services) bring their own chemicals, microfiber, autoscrubber, and accountability — and they bond/insure their staff so theft and slip-and-fall sit on their COI, not yours. Hire only contractors who carry their own GL + WC at $1M minimum and are E-Verify enrolled. Hood and duct deep clean stays separate (NFPA 96 specialist, $400-$2K/visit quarterly or semi-annual). For a 60-seat coffee shop, in-house closing-team cleaning is the right answer; below that scale a contractor is overhead you don't need.

Run the DOH self-inspection worksheet WEEKLY — that single discipline is the highest-ROI compliance practice in NYC food safety per Letter Grade Consulting. NYC Health Code Article 81 governs every food prep step; <14 violation points = A, 14-27 = B, 28+ = C/closure. Monthly drill: cold holding ≤41°F (probe-thermometer log, not the thermostat), hot holding ≥140°F, handwash stations unblocked and stocked, food contact surfaces, pest evidence, employee hygiene, allergen posting (NYS Ch. 480 since May 2025, $125/violation), Food Protection Certificate holder on premises during all hours of operation (NYC §81.15, 5-yr cert, $24 online or $114 in-person). Quarterly hire a mock-inspection firm — Letter Grade Consulting, EHA Consulting Group, Restaurant Professional Services — at $500-$2K to walk the kitchen as DOH would; the ROI is $5K-$20K in avoided fines and grade damage. FSE permit renews annually at $280 plus $25 frozen desserts. Required postings: letter grade card (front window), allergen notice (prep + menu), DOH permit, Smoke-Free Air Act 'No Smoking' at entrances. A B grade drops covers 5-9%, a C drops 20-30% — this is operator existential.

Calendar the FDNY Place of Assembly inspection 60 days before expiration and walk the building like an FDNY inspector first. PA threshold is 75+ occupancy (Place of Assembly classification under NYC FC §403). What FDNY checks: posted occupancy sign matching certificate, all egress paths clear and unlocked (push-bar hardware functional), fire alarm annunciator history, sprinkler and standpipe quarterly inspection tags current (NFPA 25), kitchen hood suppression semi-annual tag current (NFPA 17A — single most-cited fail), Class K extinguisher within 30 feet of cooking, all extinguishers annually-tagged and gauge-in-green, emergency lighting and exit signs operational (monthly 30-second test, annual 90-min load test), F-03 Indoor Place of Assembly Safety Personnel certificate holder on duty during all hours of public occupancy, fire safety/evacuation plan on file. Most common fails: blocked egress ($2,500-$10K+ first violation), expired hood-suppression tag ($1K-$5K), missing F-03 holder ($500-$2K + immediate shutdown), over-capacity. Renewal fee $200-$500 occupancy-dependent; total inspection prep through Milrose or in-house GM walk should run $500-$2K. File 60 days before expiration.

On-premises liquor licenses renew biennially (every 2 years) — start the renewal 90 days out, not 30. T-90: pull the license certificate, verify expiration date in SLA portal, confirm Method of Operation (MO) on file matches actual operations (food service model, entertainment, hours, capacity, principal use) — operating outside the MO is a 'failure to conform' violation. T-75: reconcile any officer/stockholder/LLC-member changes — must have been reported within 10 days of change or you're sitting on an 'availing' violation (one of the harshest SLA charges). T-60: pull all violation history from the SLA portal, brief your SLA attorney (Davidoff Hutcher & Citron, Pierce & Kwok, Stacy Weiss, Korngut Paleudis, Kearney Law Group) on any open issues. T-45: file renewal in the SLA portal with renewal fee, updated corporate filings, current Community Board stipulations check (closing hours, soundproofing, capacity) — operating in violation of CB stipulations is also failure to conform. T-30: confirm Form NY-1 disorderly-premises log is current (3+ 911 calls/arrests = disorderly premises charge). Operating without a valid renewed license is a criminal offense — $15K first-offense fine for unlicensed operation per Evans Fox LLP analysis. SLA hearings run 3-12 months and a 30-day suspension on a $500K/mo bar is $500K of direct revenue.

File NYS sales tax quarterly for most NYC restaurants — quarterly returns are due the 20th of the month following each sales-tax quarter (Mar/Jun/Sep/Dec quarter ends). Heavy hitters above the $300K/quarter taxable-sales threshold get bumped to monthly Part-Quarterly (PrompTax) filings, with electronic funds transfer required. NYC combined sales tax is 8.875% (4% NYS + 4.5% NYC + 0.375% MCTD) — this is what your POS collects and what you remit. File via NY State Department of Taxation and Finance Online Services (Form ST-100 quarterly, ST-101 annual reconciliation due March 20). Annual filers (under $3K/yr in tax) are essentially nobody in restaurants. Penalties: 10% of tax due plus interest for late filing, doubled for repeat offenders, plus the Tax Department aggressively pursues responsible-person liability against owners and officers personally. Toast, Square, and Lightspeed all auto-calc sales tax by location but YOU are responsible for filing — tie this to your bookkeeper or Restaurant365 monthly close. Sales tax is a trust-fund tax — late or short remittance is the #1 trigger for a NYS responsible-person audit and personal liability under Tax Law §1133.

Commercial Rent Tax (CRT) applies ONLY to commercial tenants in Manhattan south of 96th Street with annualized base rent of $250K+ (the small-business exemption threshold raised from $250K to higher in recent years — verify current via NYC DOF). Outside that geography (Bronx, Brooklyn, Queens, Staten Island, or Manhattan above 96th) — zero CRT exposure. The headline rate is 6% of base rent, but a 35% reduction brings the effective rate to 3.9%, and the small-business credit graduates the bill so tenants in the $250K-$300K rent band pay reduced amounts. Quarterly returns CR-Q (Form CR-Q1/Q2/Q3) plus annual CR-A — quarterlies due 20 days after quarter-end, annual due June 20. Includes base rent + escalations + tenant payment of landlord's real estate taxes + percentage rent — escalations and pass-throughs are the most-missed items in CRT calcs. If you're Manhattan-sub-96th and signing a lease above $250K base, build CRT into your 5-year financial model — it's typically 0.5-1.0% of revenue at restaurant levels, NOT trivial. Default vendor: your CPA (Anchin, Withum, Marks Paneth) or your hospitality bookkeeper. Operators in outer-boroughs sometimes mistakenly file CRT — don't, you're not subject.

W-2s to employees and to SSA/IRS by January 31; 1099-NEC to contractors and to IRS by January 31; 1099-MISC (rents, royalties, attorney fees) to recipient by January 31, to IRS by Feb 28 paper / March 31 e-file. Every contractor receiving $600+ in a calendar year gets a 1099-NEC — this is where independent musicians, fitness instructors, bookkeepers, social media contractors get caught. Late-filing penalty $60-$310 per form per IRS §6721, scales to $660+ if intentional disregard. The big audit triggers: (1) misclassifying employees as 1099 contractors — NYS DOL and the IRS use the ABC test plus economic-realities test; W-2 minimum-wage workers misclassified as 1099 contractors is the #1 NYS DOL audit, with back-wages, OT, spread-of-hours, plus penalties stacking, (2) cash tip underreporting — IRS Form 8027 large-food-or-beverage establishment reporting (10+ employees on a typical day) flags any operator below the 8% gross-receipts tip allocation floor, (3) NYS sales tax mismatch vs reported gross receipts, (4) withholding mismatch between Form 941 quarterly and W-3 annual. Use a hospitality payroll provider (Toast Payroll, Restaurant365 Workforce, Paychex, ADP, Gusto) — they file federally + NYS + NYC and produce the year-end forms. Hospitality Industry Wage Order Part 146 audit focuses on tip credit, spread-of-hours, and Fair Workweek (predictability pay).

NYC Noise Code §24-231 et seq. caps amplified commercial sound at 42 dB(A) measured INSIDE the nearest affected residential dwelling unit — that's the threshold, and 42 dB(A) is roughly quiet conversation, so any decent sound system at the residential boundary will trip it. 311 noise complaints are the #1 trigger; DEP responds with a sound-meter inspection, fines run $800-$3,200 per violation, and pattern complaints feed a 'disorderly premises' SLA charge that can suspend the liquor license. Common operator unforced errors: rooftop bar amplification after 10pm, open windows/doors during DJ sets, basement bass coupling through structural wood-frame buildings, late-night sidewalk smoking-section noise. Mitigation: hire an acoustic consultant pre-opening (Cerami, Acoustic Distinctions, Charles M Salter NYC) — sound isolation upgrade is $50-$300/sqft for a residential-adjacent buildout but cheaper than the SLA paper trail. Sound-limiter on the DJ booth wired to the amplifier (Formula Sound, dB-keeper), close windows by 10pm, run a noise-management protocol for outdoor dining. Mamdani-era enforcement crackdown April 21 2026 increased DEP precinct sweeps in Williamsburg, LES, Greenpoint, Astoria — sustained 311 complaints from a single residential receiver are the SLA-renewal kill shot, far worse than the DEP fine itself.

Pull eight numbers every morning by 11am off the prior day's POS Z-report and reservation tail: net sales vs forecast, covers vs forecast, average check, food cost % (theoretical from POS recipe BOM), labor % (hours × blended wage / net sales), comp+void $ and % of sales, no-show + cancellation count, and OpenTable/Resy/Tock seat-fill ratio. Layer a daily cash recon — drawer count vs POS cash declared, with any variance >$10 flagged to the GM that day. Track overtime hours daily, not weekly — a GM running 4%+ OT consistently has a scheduling failure per the GM bible benchmark. Bar-led venues add pour-cost variance (target 18-22%) and bottle-count delta on high-velocity SKUs. Tools: Toast/Square daily flash report + MarginEdge daily summary email — the manual-spreadsheet version takes 90+ min and gets abandoned by week 3.

Run the weekly close every Monday AM — this is the NYC GM standard cadence. Eight numbers: weekly net sales vs plan, prime cost % (labor + COGS, target 58-63%), food cost % (target 28-32%), pour cost % (target 18-22%), labor % (target 28-32%, alarm at 33-36%), management vs hourly labor split (target 5-8% mgmt / 20-25% hourly), comp expense as % of gross sales (alarm at >2%), and OT as % of total labor (alarm >2%). Add same-store-sales WoW + YoY, reservation conversion %, and channel mix (Resy/Tock/Walk-in/3PD). Always report variance against plan, not absolutes — a labor % at 31% misses by 200 bps if plan was 29%. Pull from MarginEdge ($330/mo flat) or R365 — Excel breaks at multi-week trending.

Lock these target bands into your operating plan: full-service restaurant 28-32% total labor (5-8% mgmt + 20-25% hourly); bar/nightclub 18-25% — drink velocity carries it lower; nightclub adds 4-8% security/SOP labor on event nights. Fast-casual (Sweetgreen, Cava, Chipotle, Dig) lives at 26-30% but is gated by DCWP Fair Workweek — 14-day advance schedules + $0.10-0.75/hr premium pay for inside-window changes per NYC Admin Code §20-1201; back-claims hit $500-$2,500/employee. Fine-dining caps at 32% — anything above is a service-choreography or scheduling failure. Hotel F&B runs 30-35%. Above 36% on any concept is the 'call it a problem' zone; 37%+ is unsustainable. NYS spread-of-hours (DOL §146 — extra hour at min wage on shifts >10 hrs) and call-in pay quietly inflate labor 1-2 pts in NYC vs national comps.

Hold prime cost (labor + COGS) inside 58-63% of revenue for an NYC restaurant — 65% is the alarm; 68%+ is bleeding. Hotel rooms-side runs 40-50%. The split: food cost 28-32% + pour 18-22% blended COGS ~30%, plus labor 28-32% = 58-62% target. NYC inflates this 2-4 pts vs national: §146 spread-of-hours + uniform allowance + WTPA §195 compliance + 2025 wage floor + post-RTO weekday lunch softness all compound. Review weekly — Monday AM P&L close is the NYC GM standard. If you cross 65% for two consecutive weeks, freeze hires, cut one shift, and re-spec the highest-cost three menu items before the third week.

Build the proforma off these 2026 NYC anchors: neighborhood casual $45-75 average check (food + bev + tip-out base), upscale casual $85-130, fine-dining $150-250, Michelin tasting-menu $250-450+. RevPASH is the better number — covers × avg check / (seats × open hours) — with NYC fine-dining anchoring at ~$89 RevPASH (80 seats × 2.4 turns × $185 / 5 hrs) per the revenue-management bible. Casual NYC targets 2.5-3.5 turns/seat at dinner; fine-dining 1.8-2.4 turns. A 4% lift in average check + 10% reduction in turn time compounds to ~14% RevPASH. Bar-led venues should track average spend/seat-hour and tables-sold/night separately — bottle service segments use spend-per-table, not per-cover.

Tag every cover in Resy/Tock/SevenRooms with first-visit vs return — most NYC operators never turn this on, then can't measure. Healthy NYC neighborhood casual targets: month 3 ~15-20% repeat weekly covers, month 6 ~30-35%, month 12 ~40-50%. Fine-dining runs lower in absolute count — 20-25% at month 12 — but each repeat is a higher-margin VIP cover and Tock VIP-conversion rate is the headline metric per the GM bible. Bar/nightclub targets are noisier — 30-40% regulars at month 6 is the floor; below 25% the venue is a tourist trap with a closure clock. Below 25% repeat at month 6 for a neighborhood spot signals concept-or-execution failure, not a marketing fix.

Run a theoretical-vs-actual variance report weekly off the POS recipe BOM — variance >2%/month is almost never theft, it's portion drift, over-pour, untracked spoilage, and untracked staff meal. Cooks plate 0.5oz over spec on 200 covers/wk = ~12 lb/wk extra protein at $14/lb = $9K/yr leak per item. Bartenders free-pour 1.5oz vs 1.25oz spec on 800 drinks/wk = pour cost climbs from 18% to 23% — Wisk AI ($165/mo) or Partender ($249) catches this in 60 days. Untracked staff meal is the silent 1-2% — log it as a separate comp account, not COGS. Fix sequence: install scales, post recipe cards at every station, do blind line-checks twice/week, review variance Monday AM with the chef line-by-line per the GM monthly-P&L drill.

Single-drawer assignment is the control — the moment two servers share a drawer, accountability breaks and skim becomes untraceable. The NYC bar/nightclub segment fails this more often than any other operational area per the GM bible — split drawers, no manager-key voids, no end-of-shift count-vs-Z reconciliation. Install: (1) one drawer per shift per bartender/cashier, (2) blind drop — server counts before pulling Z, (3) manager PIN required for voids/comps/no-sales, (4) daily deposit logged same-day with armored pickup (Loomis/Brinks) for >$2K, (5) weekly Z-vs-deposit-vs-bank match by the bookkeeper. Cash discount programs — allowed under NY GBL §518 if structured correctly — also concentrate cash exposure; tighten controls before launching one. Variance >$10 on a single shift is investigable; >$25 is documented; recurring is termination.

Cap comp authority by tier in writing on day 1 — line staff $50-100 (drink/course, no approval), shift manager $300-500 (full check, logged), GM unlimited but reviewed weekly. The 3-tier service-recovery ladder is the NYC GM standard per the GM bible. Without the cap, comps drift from 0.5% to 2-3% of gross sales in 90 days — that's $3K/wk on a $150K-revenue venue, vaporizing the EBITDA bonus pool. Post comps to a separate expense account in QBO/R365, not netted against revenue — this is mandatory per the accounting bible — so the weekly P&L surfaces the trend on Monday AM. Audit the comp log weekly: same server comping the same table = sweetheart pattern; same time-of-day comp spike = unposted shift meal. Tighten authority back to written tiers and require manager-PIN entry in Toast/Aloha for any comp >$50.

Audit your AP last quarter — operators routinely run 30-60 vendors with sub-$500/mo spend each, and the bookkeeper spends 5-10 hrs/wk just typing invoices. Manual AP exceeds the SaaS cost of OCR within 60-90 days at NYC wage rates per the accounting bible. Consolidation sequence: (1) move ordering to BlueCart Endless Aisle — free, single-cart, multi-vendor, ~125K restaurants on it; (2) layer MarginEdge ($330/mo flat regardless of unit count) for OCR invoice digitization + daily food-cost report; (3) collapse the produce/protein/dry-goods long tail into a primary broadliner (Sysco SOS or US Foods Moxe) plus 2-3 specialty vendors with EDI; (4) DAVO ($49/mo per location) automates NYS sales tax — late filing = 10% penalty + 1% per month per the accounting bible. Goal: cut back-office hours 50%+ and get to a 7-10 day period close (vs 14-21 manual).

Kill the spreadsheet — install 7Shifts, HotSchedules, Harri, or Dayforce (all on the GM tech-stack list) within 30 days of opening. Manual scheduling has no labor-forecast tie-in, no sales-per-labor-hour budget, no real-time clock-in vs schedule alert — labor % runs 200-400 bps above plan and OT exceeds the 2% target band uncontrolled. For NYC fast-food (statutorily defined: limited-service, 30+ national locations) you also need DCWP Fair Workweek compliance baked in — 14-day advance schedules, $0.10-0.75/hr premium pay for inside-window changes per NYC Admin Code §20-1201, plus call-in pay (4 hrs at min wage for canceled scheduled shift). Premium-pay back-claims reach $500-2,500/employee per the GM bible failure-mode catalog. Fast-casual chains (Sweetgreen, Cava) have been repeatedly dinged. 7Shifts/Harri auto-flag premium-pay triggers; Excel won't.

Close every day by 11am — Z-report, declared cash, declared tips, comp/void log, prior-day labor hours, vendor invoices logged into MarginEdge or QBO. Operators who run weekly-only close find the 200-bps food-cost miss 6-9 days late, by which time the leak compounds to $10-25K on a $1.5M-revenue restaurant. The closed-restaurant pattern is consistent: no daily flash, no weekly variance review, period close 14-21 days late, owner sees the P&L 5 weeks after the month — and the runway is gone before action is possible. Modern stack: POS daily sales journal auto-syncs to QBO/Xero/R365; MarginEdge ($330/mo) ingests vendor invoices via OCR (95%+ line-item accuracy) and produces real-time food/labor cost the morning after. Manual entry is obsolete for any operator over $1M revenue per the accounting bible. Period close target: 4-7 days on R365, 7-10 on MarginEdge+QBO, 14-21 on pure-manual = sell the business.

Three patterns leak more cash than every other source combined: (1) **No-sale + voided-check skim** — server rings the order, prints check, takes cash, voids the check, pockets the difference; fix = manager-PIN required for voids in Toast/Aloha/Square + daily void-log review against shift sales pattern. (2) **Comp abuse / sweetheart comps** — server comps a friend's drink/dessert and the cap drifts; fix = post comps to a separate expense account (not netted against revenue per accounting bible), audit comp % weekly, terminate at >2% of gross sales repeated. (3) **Tip-pool / cash-tip misallocation** — drawer comes up short and cash tips are reclassified to cover; fix = blind drop, per-shift drawer assignment, FLSA-compliant tip-pool config in POS (2018 CAA amendments — BOH can join only if employer pays full min wage), and daily declared-tips reconciliation with payroll export. PCI DSS 4.0.1 enforcement (March 31 2025) made cloud POS auto-handle compliance; legacy on-prem requires SAQ filing + acquirer fines $25-100/mo.

Calendar every FDNY and DOHMH date the day you sign the lease — annual FDNY Type-I hood inspection, FDNY Place of Assembly CofA renewal (required ≥75 occupant load, 4-8 weeks to issue), DOHMH letter-grade cycle (initial 2-3 weeks post-open; A=0-13, B=14-27, C=28+ violation points; re-inspection 7+ days if >13). Miss the FDNY PA renewal and FDNY can suspend occupancy on inspection — a one-night closure during a Saturday service is $40-80K lost revenue plus social-media damage. Miss the DOHMH cycle and a B/C posted in the window is a marketing disaster — Eater NY covers it within 48 hrs per the consultant bible failure-mode catalog. Pre-stage a mock DOHMH inspection 1 week before the expected window — Tier-2 NYC consultants (JK Consulting, JEM, Bidwell) bill this at $2-5K and clients hit A on first inspection 85%+ of the time. Mamdani-administration DOHMH cadence in early 2026 is reportedly more outreach-focused, but track A-grade rates by borough quarterly — the policy is not enforcement-relief.