Brand & Concept Development

Pull a 0.5-mile radius scan on Placer.ai or SafeGraph for your target site, then cross-reference against NYC DOH active permits and Google Maps for category density — under 4 direct competitors per 0.5-mile is a real gap, 8+ is saturated. Layer in median household income from ACS 5-year data and daytime/nighttime population split (the LION dataset on NYC Open Data is free), because a 'gap' in Tribeca that's $250 per cover means nothing if your math needs $80. Walk the block 3 separate times — Tuesday 7pm, Friday 10pm, Saturday 1pm — and count covers, not seats. If your nearest direct competitor has a 45-minute Resy wait at 8pm Wednesday, that's a real signal; if they have open tables, the gap is likely a graveyard. Budget $3-8K for a real Buxton or Buxton-equivalent site analysis if the lease is over $25K/month.

Run a 4-6 week pop-up at a host venue (Sunday or Monday slots run $500-2,500/night in NYC) and track 3 things: spend-per-cover, repeat-visit rate within 30 days, and unsolicited press coverage. Restaurant Week-style fixed-price tests at $45-65 will tell you if your concept commands a premium or lives at the floor. Eater NY, Grub Street, Resy's Hit List, and Infatuation pickup of a 4-night pop-up is the single best leading indicator that a permanent build will get press on opening. Capture every email at the door (a $99 Square subscription will do it) — anything under 40% email-capture rate means the concept isn't memorable enough. Total validation budget should run $15-40K vs. $80K-$1.5M of buildout regret.

Stack 3 sources: NYS SLA new license filings (public, weekly), NYC DOB Alt-2 filings tagged for restaurant use, and Resy/OpenTable inventory by neighborhood (you can scrape this manually in 90 minutes). When you see 6+ new SLA filings in a 3-block radius for the same cuisine in a 90-day window, the wave has already crested — you're now late by 12-18 months. The 2026 actual signals are: licensed-but-spirit-free bars (NYS SLA on-premise no-spirit endorsements up ~22% YoY), Korean-corner Mexican fusion (verified 9 new Manhattan openings Q4 2025-Q1 2026), and 24-seat micro-omakase ($175-285 PP). TikTok views without a corresponding lift in Resy reservation depth is hype, not demand. Spend 8-12 hours a month reading Skift Table, Eater national, and Restaurant Business Online for real macro signals.

The working ratio in NYC is roughly 1 viable mid-tier restaurant per 1,200-1,800 daytime+nighttime population for any given cuisine (varies by ticket band). West Village can hold ~14 Italian spots in 0.6 sq mi because daytime+resident population is ~38K and tourist overlay adds 30%; LIC can hold maybe 6 because the daytime population evaporates after 6pm. Cross-check against Yelp 'Italian' filter count + 1.5x for non-Yelp listings to estimate true competitive set. If your category already has 3 spots within 4 blocks at the same price point, you need a real hook — chef pedigree, IP license, or a price-point gap (e.g., $24 pasta in a $38 pasta block). Resy waitlist depth at the top 2 competitors is the single best 'is there room' indicator.

Yes — 25-40 in-person interviews with people who match your target persona, recruited through Respondent.io ($40-80/interview) or via a 4-block hand-recruit at a neighbor coffee shop. Ask 'where did you eat the last 3 times you spent $80+?' before you ask 'would you eat at my concept?' — the first question is honest, the second is fantasy. Look for the same unprompted complaint to come up in 30%+ of interviews ('I can never get a Wednesday rez at X', 'nothing in this neighborhood opens past 11'). Anything under 25 interviews is anecdote; 40+ starts to be data. Budget $2-4K and 3 weeks for the full sweep.

For a sustainable NYC restaurant, locals (within a 1-mile walk) should drive 55-70% of weekday covers and 35-50% of weekend covers — anything more tourist-dependent and you're a 4-year exit, not a 12-year operation. Times Square / Hudson Yards / South Street Seaport flip those ratios and the concept design has to follow (faster turn time, larger group bookings, English-default menu, photo-forward dishes). Check NYC & Co.'s 2025 visitor data: 64.3M visitors spent $51B, but that's distributed wildly — Times Square gets ~50M annually while West Village gets a fraction. If your concept needs Instagram-driven destination dining to hit pro forma, you're underwriting tourist revenue and need a separate playbook for the inevitable tourism dip.

The honest NYC arc is: 18-30 months at peak (Resy waitlist, press cycle, $200+ check averages holding), then a 24-36 month plateau, then concept refresh or fade — total useful lifespan 5-8 years for most non-institution concepts. Institutions like Balthazar (1997), Gramercy Tavern (1994), and Carbone (2013) are outliers, not the norm. Plan a concept refresh budget of $80-250K at year 3-4 (menu reset, room paint/textile refresh, new uniform, refreshed brand book) — this is not optional, it's underwriting. The concepts that die fastest are those that locked too hard on a 2024-era trend (e.g., 'natural wine bar with no menu') and can't pivot. Build optionality into the brand voice from day one — a name like 'House' or 'Cafe' or '[Surname]' ages better than 'Pasta Project' or 'The Kale Yard'.

One page, 8 sections: (1) one-sentence concept ('a 48-seat Lower East Side counter-service Sichuan handpulled noodle bar, $24 average ticket, 11am-11pm, no reservations'), (2) target customer in 2 sentences with age/income/walk-radius, (3) menu architecture in 6 lines, (4) check average and target covers/day, (5) competitive set (3 named restaurants with one-line gap analysis each), (6) physical footprint (sq ft, seats, kitchen %), (7) total project budget range and capital ask, (8) 90-day pre-launch timeline. If you can't fit it on one page, the concept isn't tight enough. Investors and architects both read this same page — they need it to align before money or drawings move. Iterate 4-6 times before you treat it as locked.

Start with 80-120 candidate names across 5 buckets: surnames (Carbone, Sadelle's), places (Llama Inn, Estela), descriptors (Atomix, Le Bernardin), invented words (Daddo, Atoboy), and one-word gut-punches (Don, Lilia). Cut to 25 by killing anything you can't pronounce on first read or that has a tortured spelling — your hostess will repeat it 600 times a week. Check USPTO TESS (free), NY DOS Corporation Name database, and Instagram + .com availability for the surviving 25. Anything that requires owning '[name]nyc.com' because '[name].com' is taken is a real signal you should keep going — second-position names cost 3x to rank and 5x to brand-recall. Final 6 go to a 50-person friend test for pronunciation, recall after 24 hours, and gut association.

NYC speaks 200+ languages and the top 10 (Spanish, Mandarin, Cantonese, Russian, Korean, Yiddish, Polish, French, Italian, Bengali) cover ~90% of the relevant audience for vetting. Run the name through native-speaker review for those 10 — Upwork can source vetters at $35-75 each, total under $700 for the full sweep. Specifically test for: profanity, sexual innuendo, religious offense, and political loading (Russian and Mandarin are the highest-risk for the last). Brand name 'Iso' tested clean in Italian and Japanese but means 'frozen' in Finnish, which doesn't matter for a NYC pizza spot — the risk filter is 'does it make sense to a NYC immigrant taxi driver hearing it for the first time'. Document the clearance in writing in case a competitor later claims your name is offensive.

Get the .com if it's under $5K — it's still the default address bar guess, especially for the 35+ demographic that drives $80+ tickets. Below 35, social handle (especially Instagram and TikTok) matters more than domain, so secure exact-match @handle on IG, TikTok, X, and Threads day one. .nyc works for hyperlocal concepts but expect 15-25% of customers to land at the .com first and bounce. For premium domains, the brokers to call are HugeDomains, Sedo, and Saw.com (the last for $25K+ deals); a $5-15K domain is normal for a 4-letter restaurant name and you'll pay it back in zero ad-spend over 3 years. Don't launch with a hyphenated or '-nyc' suffix domain if you can possibly avoid it.

Check the geographic scope of their trademark first — if they hold a federal USPTO mark in Class 43 (restaurant services) for the exact name, you cannot legally use it nationally regardless of where they operate. If they only hold a state registration (e.g., California Secretary of State only), you can probably use it in NY but expect a cease-and-desist within 18 months if you grow to press visibility. Email them once at the corporate address and ask: 30% of small operators will sell you a coexistence agreement for $5-25K (geographic carve-out, both parties stay in their region). The other 70% will say no and you should pick a different name — litigation runs $40-150K and a year, and you'll lose if their mark is older. Don't get clever with adding 'NYC' or 'New York' as a differentiator; USPTO and courts treat geographic descriptors as non-distinctive.

Run a 50-person SMS test: text the name as plain text (no audio) to 50 contacts and ask 'how do you say this'. If less than 75% pronounce it the way you intended, the name will cost you 5-12% of word-of-mouth referrals because people don't recommend what they can't say out loud confidently. Atomix, Atoboy, Llama Inn — all clear. Cosme (KOZ-may) lost recall for 18 months until press normalized it. If your name needs a phonetic explainer in your IG bio, that's not a brand voice choice, it's a tax. Counter-cases exist (Le Coucou, Frenchette) but they're carried by chef pedigree and a $400M Restoration Hardware-anchor neighborhood — most operators don't have that cushion.

Use a generic LLC name (e.g., '142 Orchard Holdings LLC') and operate the restaurant as a DBA — this protects the brand name as separate IP from the operating entity and makes it easier to license, sell, or open a second location under different ownership. NY DOS DBA filing is $25 county-level; the LLC formation runs $200 plus the publication requirement (~$500-2,000 depending on county; NY County publication is the most expensive). If you ever want to do a brand deal, expand to a hotel partnership, or sell the IP separate from the operating lease, the holding-company-plus-DBA structure is what your acquirer will demand. Setting it up day one costs ~$1,500 total; restructuring later costs $8-25K in legal fees.

Address numbers (e.g., 'ABC Kitchen' was originally 35 East 18th Street's neighborhood spot) are fine if you own the lease long-term and don't plan to expand — but if you ever move, the brand goes with the address and you've trained customers to associate you with a building you no longer occupy. Year-based names ('2024', 'Estd. 2024') age into liability the moment they read as dated, usually by year 3. Trend-tied names ('Brutalist Cocktails', 'Bone Broth Bar') compound the same risk. The names that travel best are surnames, single common words ('Don', 'Estela'), or invented short words ('Atomix') because they don't anchor to a moment. If you want to honor a year or address, make it a secondary tagline, not the primary mark.

Start free at USPTO TESS (now Trademark Search), search exact match plus 2-3 phonetic variations in International Class 43 (restaurant, bar, hotel services) and Class 30/29 if you'll sell packaged goods. A clean TESS search is necessary but not sufficient — common-law unregistered marks in NY can still block you under Lanham Act §43(a). For the real clearance, hire a trademark attorney for a $750-1,800 'knockout + full search' that includes Thomson CompuMark or Corsearch reports covering common-law, state, and federal databases. Anything under $750 is the attorney just running TESS for you. Budget 4-6 weeks before filing; most fatal conflicts surface in week 2-3.

USPTO filing fee is $350 per class for the standard TEAS application (effective Jan 18 2025 fee restructure replaced the old TEAS Plus / Standard split), plus a $200 surcharge if your goods/services description doesn't match the USPTO ID Manual exactly. Restaurant operators typically file Class 43 (restaurant services) — that's $350. Add Class 25 (apparel/merch) for another $350 if you'll sell hats and tees. Attorney fees run $1,200-2,500 to draft and prosecute the application; total all-in for one class with attorney is roughly $1,600-2,900. From filing to registration runs 12-18 months in 2026 — USPTO is currently 8-10 months just to first examination. File 'intent-to-use' (1(b) basis) before you sign a lease so you lock the priority date; convert to use-based after you open and serve the first customer.

Yes — NY Department of State accepts state trademark filings under NY GBL Article 24, $50 filing fee, ~6-8 week turnaround. State registration only protects you within New York and is junior to any federal mark; it's most useful as a stopgap if you can't yet file federal (e.g., your concept isn't operating yet and you don't want to pay USPTO ITU fees). For 95% of NYC restaurants, skip state and go straight to USPTO federal — you get nationwide priority, the right to use the registered ® symbol, and access to federal court. The exception is if you're using a clearly geographically descriptive name that USPTO will refuse — state registration is your only path. Filing both costs an extra $50 and gives you minor evidentiary value in a NY-only dispute.

Class 43 is the table stakes — it covers restaurant services, bar services, catering, and hotel services. Add Class 25 if you'll sell branded apparel (hats, tees, hoodies — easily $40-180K/yr revenue at a hot concept). Add Class 30 (coffee, sauces, dry goods) if you'll do CPG or wholesale, Class 29 (prepared meals, frozen) if you'll do meal kits, and Class 33 (alcoholic beverages excluding beer) or Class 32 (beer) if you'll private-label. Each class is $350 USPTO fee. Most operators file 43 only at launch and add 25 + 30 in year 2 once merch and packaged-goods revenue is real. Filing classes you don't actually use exposes the registration to cancellation for non-use after 5 years, so don't over-file.

A descriptiveness refusal under Lanham Act §2(e)(1) is the most common first office action for restaurant names, and you have 3 months to respond (extendable once for 90 days at $125 fee). Three paths: (1) argue distinctiveness with consumer recognition evidence, (2) amend to the Supplemental Register (preserves filing date, no presumption of validity, can convert to Principal after 5 years of continuous use), or (3) add a secondary distinctive element to the mark. Costs $400-1,500 in attorney time depending on complexity. If your name is something like 'New York Pizza Bar', you'll lose path (1) and should accept Supplemental Register. If it's something like 'Carbone' (a surname), file 2(f) acquired distinctiveness with sales/press evidence. Don't ignore — abandonment is automatic at 6 months and your filing fee is gone.

USPTO requires a Section 8 declaration of continued use between years 5-6 ($325 per class) and a combined Section 8 + Section 9 renewal between years 9-10 ($650 per class), then every 10 years after. Miss either window and the mark is canceled — no grace period beyond the 6-month late filing window with $125 surcharge. Set 3 calendar reminders (year 5, year 9.5, and 6 months before each renewal) and assign a single owner; this is the most common way mature operators lose their brand. Send a specimen of current use (a menu, a website screenshot showing the mark in commerce) — the specimen is what trips up half of all maintenance filings. Total 10-year maintenance cost per class is roughly $1,000.

File the standard-character wordmark first — it protects the name in any font, color, or stylization, which is the most flexible coverage. The logo (called a design mark or stylized mark) is a separate $350 filing per class and only protects that specific design; it's worth filing if your logo has a distinctive non-text element (a custom illustration, an iconic symbol). For most restaurants the right sequence is: wordmark Class 43 at launch, then add the logo Class 43 at year 2 once the visual identity has stabilized. If you redesign the logo in year 3, the original logo registration becomes near-worthless — another reason to lead with wordmark. Total to file both wordmark + logo in one class is ~$700 USPTO + $1,800-3,500 attorney.

Mid-tier hospitality brand studios in NYC charge $25K-$75K for a full identity (Mucca, Apartamento, Order, Triboro, Hyperakt sit in this range); top-tier (Pentagram, Gretel, Collins, Roanne Adams RoAndCo) start at $85K and run to $250K+. The $25-75K package should include: 3-4 logo directions narrowed to 1, primary + secondary palette, primary + display typography licensed for digital + print, 6-12 brand pattern/illustration assets, 1-2 packaging touches (matchbook, takeout box, coaster), and a 30-60 page brand book. Avoid anyone who quotes under $15K for hospitality — you'll get template work that 4 other restaurants will also have. Expect 10-14 weeks from kickoff to delivered files. Demand source files (AI, Figma, font licenses) in writing — many cheaper studios deliver only PDFs.

It has to read at 4 sizes: a 14pt menu footer, a 1.5-inch matchbook, a 36-inch awning, and a 1-inch Instagram avatar. Test it printed in pure black on white before you fall in love with the color version — if it dies in 1-bit, the construction is weak. It needs a primary lockup (full mark + name), a secondary mark (icon-only or monogram for socials), and a horizontal version for headers. Avoid more than 2 typefaces in the lockup, avoid more than 4 ink colors (each ink is real money on awnings, neon, and packaging — a 4th color can add $1,200-3,500 to a single signage run), and stay away from gradients (they don't survive embroidery, foil, or single-color print). Custom-drawn type beats a stock font 9 times out of 10 and costs $4-12K extra.

Anchor to 1 hero color, 1 ink (usually black or warm-black like #1A1A18), and 2-3 secondary support colors — total 4-5 colors max. Avoid the 2024-2026 trend palette (the dusty terracotta + sage + cream combo is everywhere and will read as 'mid-2020s' by 2028). Pull palette references from 2 sources outside hospitality (architecture, fashion, mid-century print) — restaurants that source colors from other restaurants all start to look like each other. Pick colors that survive the printing realities you'll actually use: Pantone-matchable for brand stationery, achievable in vinyl awning, and renderable in ceramic-tile mosaic if your storefront uses tile. Test each color under your actual restaurant lighting (warm 2700K dimmer LED reads warmer; daylight reads cooler) — a palette that's beautiful on a designer's calibrated display can be muddy at 7pm dinner service.

Budget $400-2,500 per typeface family for hospitality use. You need 4 license types: desktop (for in-house design work), web (for your website by traffic tier — usually under 250K page views/month is base tier), app (if you build a native ordering app), and ePub (for digital menus / press kits). Foundries like Klim, Commercial Type, Grilli Type, and Dinamo charge $500-1,500 per family for desktop + web combined; Adobe Fonts and Google Fonts are free but you trade exclusivity (Google Fonts are on millions of sites). For a hospitality brand, paying for one custom-feeling display face (Klim's Söhne, GT America, Forma DJR) plus a free Google body face (Inter, Söhne Mono substitute) is the right balance — total around $1,200-2,000. Save the licenses in your shared drive; renewing later costs 30-50% more than buying right.

Write a 1-page voice doc with 4 sections: (1) 3 adjectives that describe the voice ('warm, dry, never precious'), (2) 3 things the voice never does ('never uses exclamation points, never explains itself, never apologizes for being busy'), (3) 8-12 phrases the brand uses verbatim ('the kitchen is glad to / we're holding the table / we'd love to see you back'), (4) 8-12 phrases the brand never uses ('absolutely!', 'amazing', 'please be advised'). Read it to a new hire on day 1 and put it in the staff handbook. Test against real outputs monthly — pull 5 IG captions, 5 customer email replies, and 5 menu descriptions and ask 'does this sound like the same voice'. Customers can't articulate brand voice but they feel the inconsistency immediately and it cuts repeat-visit rates 8-15%.

The Day-1 minimum kit: logo (3 lockups), color palette with hex/RGB/CMYK/Pantone, 2 typefaces with full license docs, 6-10 brand patterns or textures, photographic style reference (15-20 example images that define lighting, framing, food styling), and packaging templates for the 6 things every NYC restaurant ships (matchbook, takeout box, coaster, branded cocktail napkin, to-go bag, business card). Add: menu template (food + cocktail + wine), email signature template, IG story templates (3-4), and PowerPoint or Keynote pitch template for press and partnerships. A complete launch kit runs $35-95K all-in including studio time and asset production; trying to do it later in pieces always costs 60-100% more and the system fragments. Source files in Figma + AI + InDesign, fonts and patterns in a shared Dropbox or Google Drive, full brand book PDF for distribution.

Leave it alone for the first 5 years unless there's a real reason — premature refreshes are the single fastest way to undo brand equity and confuse repeat customers. Real reasons: ownership change, expansion to a second concept that needs to relate visually, legal forced change (trademark loss), or the original was so off-strategy that revenue is being measurably suppressed (rare, and you'd know in year 1). When you do refresh, evolve don't replace — keep the silhouette, color anchor, or type personality so customers recognize continuity. Big-bang rebrands lose 4-12% of repeat customers in the first 6 months. Budget for a real refresh is $40-150K depending on scope, plus $15-60K to re-do all signage, menus, packaging, and uniforms (signage alone runs $8-25K for a single-storefront facade swap).

Use AI for ideation (mood boards, palette exploration, naming brainstorms, copy drafts) but not for final deliverables — a logo or pattern that came directly from Midjourney is unprotectable as IP because the US Copyright Office (2023 guidance, reaffirmed 2024-2025) won't register AI-generated works without substantial human authorship. That means competitors can copy it freely and you can't sue. Worse, several Midjourney-generated logos have surfaced as near-duplicates of other brands' marks because the model trains on existing logos. Use AI to compress designer time (saving maybe 30-40% on the initial exploration phase) and pay a designer $15-40K to take the strongest direction to a defensible, registrable, original final mark. For a serious NYC concept with $500K+ buildout, the $20K design saving is not worth the IP exposure.

No — that's a demographic, not a persona, and it leads to generic concept design. A real persona is a 1-page profile of one named imaginary person: 'Maya, 34, marketing director at a fintech in Flatiron, lives in Carroll Gardens with partner, household income $245K, dines out 4-6 nights/week at $60-180 PP, books on Resy 5 days ahead, follows @theinfatuation and @platt'. Build 2-3 personas covering ~80% of your covers — typically a primary (60% of revenue), secondary (25%), and edge (15% — the press / influencer / industry diner who drives word-of-mouth). Pin them in the staff handbook so the floor team knows who they're serving. The goal is that any decision (music volume, menu price, reservation policy, IG cadence) can be checked against 'does this work for Maya?'

Use the Geoffrey Moore / Crossing the Chasm template, hospitality version: 'For [Maya, the 34-yo Flatiron marketing director], who [wants a Tuesday-night spot she can walk to from the office and not feel like she's eating in a hotel lobby], [Concept Name] is a [category, e.g., 38-seat Mediterranean wine bar] that [single key benefit, e.g., serves a $58 prix fixe with a 6-glass wine flight]. Unlike [the closest 2 competitors], we [single point of differentiation, e.g., source from 4 NYS-only producers and the wine list is rebuilt weekly].' One paragraph, 50-80 words, fits on a 3x5 card. If you can't fit it, the positioning is muddy. Re-test it against staff every 6 months — drift is the enemy of brand.

Pull the Resy / OpenTable price tier ($, $$, $$$, $$$$) for the 12 closest restaurants in your radius and find where there's a real gap — most NYC neighborhoods have 5-7 spots at $$ ($30-45 PP), 4-6 at $$$ ($45-95 PP), and 1-3 at $$$$ ($95+). The smartest concept-positioning move is the 'one tier above the cluster' play — if your block is dense at $35 average, a $58-72 concept with intentional design and one signature dish stands out without entering institutional-tier risk. Anchor your check average to your target persona's stated dine-out spend (Maya's $80-180 band), not to your food cost (food cost is a constraint, not a positioning input). Run the math backwards: covers/day × check avg × 7 days × 4.3 weeks = monthly revenue; if that doesn't cover rent at 8-10% of revenue, the price tier is wrong for the rent.

Pick one hero day-part and let the others follow — most successful NYC restaurants get 55-75% of revenue from one band (dinner for most, brunch for a smaller set, late-night for very few). Brunch margins are often higher than dinner (egg dishes are 18-22% food cost vs 28-34% for dinner protein) but the cover average is 40-55% lower, so the math only works if you can do 2 turns 11am-3pm Saturday and Sunday. Late-night (post-11pm) is a small but real positioning opportunity post-COVID — the FiDi/Lower East Side gap for kitchens open until 2am is genuine, but staffing is brutal (kitchen labor is +25-40% after midnight) and SLA hours matter (most NYC SLA on-premise licenses go to 4am but neighborhood opposition can cap you earlier). Don't try to be 3 things on day 1 — open with dinner, layer brunch in month 4, late-night in month 8 if at all.

Five low-cost methods stacked: (1) read 30 Google + Yelp reviews of the 4 closest direct competitors and tag the recurring complaints/praise — your persona's pain points surface in their own words; (2) follow 50 IG accounts of likely customers (people who tag the 4 competitors) and screenshot what else they post; (3) sit in 2-3 nearby coffee shops on a Tuesday morning and a Saturday and observe; (4) interview 6 hostesses or bartenders at competitors over a beer ($40 + an hour of their time gets you 90% of customer-behavior data); (5) buy a $300 Placer.ai or AlphaQuery report on the radius for visit cadence and home-zip distribution. Total cost under $1,500, time about 30 hours, and you'll have more usable persona data than 80% of new operators.

First 90 days, expect persona drift — your real customer is rarely exactly the imagined one. Track it: pull 200 covers' worth of OpenTable/Resy data (party size, day, time, repeat rate, zip) and survey 50 walk-ins on age band, neighborhood, and how they heard. If the gap is small (10-25%), tune the concept toward the actual customer (music, menu pacing, pricing). If the gap is large (40%+) — say you positioned for couples doing date night and you're getting groups of 6 doing birthdays — you have a positioning failure that you can either embrace (rebrand internally as a group spot, raise minimums, change the seating mix) or fight (hostess script, pricing, marketing all push back to the original persona). Fighting takes 6-12 months and 15-25% of revenue. Most operators should embrace the truth.

Hotel F&B has 3 personas to balance: house guests (typically 25-45% of covers, captive but not loyal), neighborhood locals (the hardest to win — they assume hotel = mediocre), and destination diners (the hero persona for press). The tension is that house guests want predictable, locals want a reason to come back, and destination diners want a story. Solve it by positioning the restaurant first as a stand-alone (its own name, separate IG, separate Resy listing, dedicated entrance if possible) and the hotel second — Eleven Madison's old positioning under the MetLife building, Le Coucou inside The Ludlow, and The Modern at MoMA all do this. Drive the local capture by giving locals a price advantage on a specific day-part (industry night, chef's counter Mondays); house guests will pay the rack rate. Budget 18-30% of room nights as restaurant covers if positioning works.

Three concentric rings: (1) direct competitors — same cuisine, same check average, within 0.5 mile (usually 3-6 places); (2) substitute competitors — different cuisine, same check average and same occasion (date night, group dinner) within 1 mile (usually 8-15 places); (3) reference competitors — the 3-5 NYC institutions in your category citywide that customers and press will compare you to regardless of geography. Direct competitors define your pricing and menu architecture, substitutes define your reservation pacing and design ambition, and references define your press strategy and chef hire. A common mistake is to over-index on direct competitors and ignore that your real Tuesday-night fight is with the Italian spot 3 doors down, not the Greek place across town. Build the matrix on a single sheet — 12-25 rows, 8-12 columns (price, cuisine, capacity, hours, Resy availability, IG followers, last press hit, signature dish, design vibe).

Operational systems are fair game — reservation policy (Carbone's call-for-rez phone tree, Don Angie's no-walk-ins), service flow, hours, and even specific roles (chef de cuisine + sommelier + GM trio). Menu format ideas (the prix fixe + supplements model, the bar-snack-into-mains shape) are borrowable. Design language (a banquette layout, an open kitchen pass, a brass bar rail) is derivative not infringing. What crosses the line: signature dish replication (a literal copy of the Carbone spicy rigatoni with the same garnish and same plating), trade dress (a near-identical room layout + color palette + uniform combination — Two Pesos v Taco Cabana 1992 SCOTUS holds room trade dress is protectable), and chef-poaching from a single competitor as a coordinated raid (NY non-compete law allows narrow restrictions; intentional interference with contracts is litigable). Studying competitors is mandatory; copying their distinctive expression is a $50-200K legal exposure.

Triangulate 4 inputs: (1) seat count × turn assumption (1.5-2.5 turns weeknight, 2.5-3.5 weekend) × check average × 6.5 effective days/week × 52 weeks; (2) Resy/OpenTable booking depth (you can see 30-day-out availability — full = high turns; wide-open = struggling); (3) hours of operation and day-parts; (4) public review counts and IG follower growth as a momentum proxy. A typical NYC 60-seat dinner-only concept at $90 average doing 1.8 turns weeknight / 2.5 weekend nets roughly $4.5-5.5M annual revenue. Cross-check against staff count (a hostess on a slow Wednesday tells you nothing; a 6-person FOH on Thursday at 7pm tells you everything). Eater closing reports and the NY Post real estate section sometimes leak actual revenue figures during sales — those are the cleanest data points you'll get.

Top 6 patterns from the past 5 years of Eater NY closings: (1) opening with a $5M+ buildout and 100+ seats in a $25K/mo rent — leverage kills you in any 90-day soft patch; (2) launching with a chef whose press value depends on one prior restaurant's reputation that won't transfer; (3) overstaffing FOH (1.0+ FOH per 8 seats; the right ratio for a 60-seat dinner is closer to 1.0 per 12 seats); (4) menu of 28+ items at launch — every dish over 18 dilutes execution and inventory; (5) Resy + OpenTable + Yelp Reservations all live simultaneously (causes overbooking and double-counts); (6) opening in August or January (the two slowest NYC dining months — open in March, May, September, or October instead). Each of these has killed multiple concepts in the $2-8M revenue band.

Map the 'comparable opening' press cycle for 5 restaurants similar to yours that opened 12-36 months ago: which outlets covered them, in what order, and on what timeline. The standard NYC press cycle for a well-positioned new restaurant is: Eater NY first-look (week 1-3), Grub Street roundup mention (week 2-6), Infatuation review (week 6-12), New York Magazine mention (week 8-16), New York Times review by Pete Wells / now Brian Velasco / Melissa Clark (month 4-9 if you get it). Concepts that get 0-1 of these in 90 days are usually flat to declining at month 6. Concepts that get 4-5 are full through month 18. Hire a PR firm ($5-15K/month for 3 months around opening) only if you have a real press hook (named chef, unusual cuisine, notable backers) — otherwise the spend doesn't move the needle.

You don't compete on volume, marketing budget, or buildout polish — Major Food Group spends $4-8M on a single restaurant buildout (Carbone Miami was reportedly $14M+) and Tao Group has 60+ venues and a celebrity rolodex. Compete on what they structurally can't do: hyper-specific cuisine (a 32-seat Sichuan concept beats Carbone in its 4-block radius for that craving), genuine chef ownership (groups rotate chefs every 18-30 months), neighborhood loyalty (free regular's perks, walk-in priority, names remembered), and price-tier escape (be either 30% cheaper or 30% more expensive — never within 10% of their check average). Their weakness is institutional pacing — a group will take 14 months to pivot a menu, you can do it in 2 weeks. Watch their lease expirations on PropertyShark and CoStar; expansion-driven groups consolidate every 4-6 years and there's a 12-18 month window of distraction during each consolidation.

Stand on the corner with a clicker on 3 different days and 3 different times — ideally Tuesday 6-7pm, Thursday 8-9pm, and Saturday 1-2pm — and count pedestrians passing your storefront for 15 minutes each. Multiply by 4 to get hourly. NYC primary retail corridors run 600-2,000 pph at peak, secondary 200-600 pph, residential side streets 50-200 pph. Cross-reference with NYC DOT pedestrian count data (free, published 2-3 times a year for ~100 locations citywide) and Placer.ai mall/storefront data ($500-2K for a single block report). For a destination restaurant where customers come on purpose, foot traffic matters less than for a casual or counter-service concept where impulse walks-in are 20-40% of covers. A $35 PP counter spot on a 200-pph block will starve; the same concept on an 800-pph block can hit 180 covers/day.

Pull median household income (MHI) by census tract — NYC tracts vary from $35K (parts of the Bronx, East New York) to $325K+ (parts of Tribeca, West Village, UES). The working ratio is: a casual dinner check average of about 0.04-0.08% of MHI is sustainable. So a $200K MHI tract supports $80-160 dinner; a $90K tract supports $36-72. Tribeca (~$280K MHI) supports $112-224 PP comfortably; East Williamsburg (~$95K MHI) supports $38-76 PP. You can index higher than the local MHI if your trade area pulls from a wider radius (Williamsburg pulls Manhattan diners on weekends, adding 30-50% effective income to the calc). But fighting the math — a $145 PP concept in an $85K MHI tract — burns 12-18 months of capital before you accept the price needs to come down.

FiDi has ~470K daytime workers and ~63K nighttime residents — that's a 7.5x flip and it's the single biggest reason FiDi restaurants struggle on weekends and at dinner. Hudson Yards is similar (huge daytime, thin nighttime). Conversely, Park Slope or Astoria has nearly inverted ratios — strong nights and weekends, dead at lunch. Match your concept's day-part to the population pattern: a $24 lunch counter wins in FiDi and dies in Park Slope; a $90 dinner spot wins in Park Slope and starves in FiDi unless you do happy hour and corporate group dinners. NYC LION dataset (free on NYC Open Data) gives you population by tract by hour. The mistake is opening 7-day dinner-only in a daytime-heavy zone — you've signed up to be empty 4 nights a week.

The destination-diner threshold is roughly 4 blocks (~0.2 mile, ~5 min walk) from a major subway stop — beyond that and you'll see a measurable 15-25% drop in weeknight covers vs an equivalent concept on top of the line. Within 2 blocks of a multi-line interchange (Union Sq, Times Sq, Atlantic-Barclays, Court Sq, Jay St-MetroTech) is a real premium worth $8-25K/mo more in rent. The exception is a true destination concept (3+ Michelin stars, $300+ tasting menu, IG-driven hype) where customers will walk 8 blocks and Uber from outside the city — that's maybe 1% of restaurants and you don't get to claim it on day 1. For a hotel restaurant, the rule changes — house guests don't care about subway proximity but local capture does, so the same 4-block math applies for the local-capture portion of the business plan.

Walk the block at 4 times: 9am (commuter flow, who's opening for the day), 1pm (lunch crowd, what's open and what's closed), 7pm (dinner energy, what kind of customer this block attracts after dark), and 11pm-1am (whether the block dies, has noise complaints, or transitions to nightlife). Look at neighbors — if the next door is a methadone clinic, a 24-hour bodega, or a noise-complaint nightclub, that affects your lunch crowd and your liquor license review. Count restaurants vs retail vs residential ground-floor — a block with 5+ restaurants in a row has a built-in dining audience but also competition; a block with 1 restaurant and 6 retail has lower dining traffic but you own the category. Check NYC 311 complaint data by address (free) for the past 24 months — neighbors who complained about a previous tenant will complain about you.

Adjacent retail anchors drive incremental traffic — a flagship next door (Apple, Glossier, Aritzia, RH) brings 800-3,000 daily visitors, and 8-15% will eat within a 2-block radius of the store. A grocery anchor (Whole Foods, Trader Joe's, Wegmans) drives different traffic — locals during the week, families on weekends. A hotel anchor (especially a 200+ key 4-star) gives you 40-100 captive house-guest covers/day if you position right. Conversely, an anchor closing or going dark (the Bed Bath & Beyond closures of 2023, the Whole Foods Bowery closure 2024) drops adjacent restaurant revenue 12-22% within 6 months. Read the Real Deal and CoStar weekly for anchor-tenant moves in your target zone before lease signing — the data is 3-6 months ahead of foot traffic impact.

NYC's Dining Out NYC program (made permanent under Local Law 121 of 2023, expanded under Mamdani-Flynn 2026 to restore year-round roadway dining for qualifying streets) lets you apply for sidewalk and roadway cafe permits — but eligibility is site-specific. Sidewalk cafes need 8+ feet of clear pedestrian path remaining after your seating; roadway cafes need a non-bus, non-bike-lane, non-truck-route street with a 25 mph or lower speed limit. Apply through DOT's online portal; permits run $1,050-2,650/year base plus per-sq-ft fees and require approval from DOB, DOT, and Community Board. A 12-seat sidewalk cafe can add $180-450K annual revenue at NYC dinner check averages — material for a 60-seat operation. Check the address against the DOT's eligible-streets map before signing the lease; ineligible sites are a hidden 15-25% revenue loss.

Opening menu: 14-22 dishes total is the sweet spot for most NYC concepts. Fewer than 12 reads as too tight unless you're explicitly tasting-menu only; more than 24 dilutes execution, increases prep waste 8-15%, and slows ticket times. Architecture: 4-6 starters / small plates, 6-9 mains, 2-3 sides if mains don't include them, 3-5 desserts. For a wine-bar or bar-forward concept, flip the ratio (8-12 snacks/small plates, 4-6 mains). Lock the opening menu 4 weeks before doors open so the line cooks have rep time; the first menu change should be week 6-10 and should swap 3-5 items based on actual sell-through, not chef boredom. Use a sales-mix report from your POS weekly — anything under 5% sell-through after 30 days gets cut.

Pick a clear position on the spectrum and commit — the death zone is the middle. Hyper-authentic concepts (Atomix, Atoboy, Casa Enrique, Wu's Wonton King, Atoboy) thrive because they signal expertise and let press anchor on 'finally, real X'. Adapted concepts (Carbone, Don Angie, Llama Inn, Cosme) thrive because they reframe a cuisine for a NYC palate at a NYC price point and own that combination. The middle — 'authentic-ish but with American garnishes and bigger portions' — gets compared unfavorably to both poles. Decide based on the chef's actual training and your operator persona — Maya the Flatiron marketing director will go to Atomix once a quarter for the experience and Carbone monthly for the night out, but she won't go to a 'modern Italian-American with Asian touches' concept twice. Document the position in the concept brief and audit every menu addition against it.

Yes — a single signature dish drives 20-40% of mention rate in reviews and IG tags, and gives press a hook to anchor the story. Carbone's spicy rigatoni, Don Angie's pinwheel lasagna, Lilia's mafaldine, Atomix's signature 10-course progression, Cosme's husk meringue with corn mousse, Lucali's pizza, Joe's Shanghai's xiao long bao — every NYC institution has one. Design criteria: visually distinct (you can identify it from a photo), technically reproducible by a $24/hr cook (not just the chef), priced 15-30% above the menu average so it carries margin, and shareable in the way the room is set up (a tableside finish, a presentation board, a steam release). Engineer it to be Instagrammable without being gimmicky — a real dish that happens to photograph well, not a photo prop. Test with 30 friends before opening; if 80%+ remember it 24 hours later, you have it.

NYC working ranges by category (food cost as % of menu price): Italian pasta-driven 24-30%, steakhouse 32-40%, sushi/omakase 30-38%, modern American 28-34%, vegetarian/vegan 22-28%, pizza 18-26%, Mexican casual 22-30%, fine dining tasting menu 30-38%. Bar / cocktail food cost should run 18-22% to subsidize labor-intensive food. The blended target for most NYC restaurants is 28-32% food cost, 30-36% labor — anything pushing food past 34% means menu engineering or pricing is broken. Use a weekly menu engineering report (popularity vs profitability quadrant) and prune the bottom-left items (low sales, low margin) ruthlessly. Negotiate produce with Baldor and protein with Pat LaFrieda or DeBragga — bulk relationships save 4-9% on food cost vs spot-buying.

Beverage should be 28-42% of total revenue for a sit-down NYC restaurant — anything below 25% means under-monetized, anything above 45% means food underperforming. Lead with one anchor: wine-led (Estela, Wildair, Frenchette — wine drives 40%+ of rev), cocktail-led (Death & Co, Attaboy, Dante — spirits 50%+), or beer-led (Spuyten Duyvil, Tørst — beer 35-45%). Anchor choice drives glass program, license endorsements, and bartender hire profile. Pour cost: spirits 18-22%, wine 28-35% (lower if BTG-led), beer 22-28%. Wine BTG (by-the-glass) program is the highest-leverage menu engineering tool — a 6-glass program at $14-22 per glass with rotating selections drives both check average and sommelier hire credibility. Don't build a 600-bottle list at opening; start at 80-150 bottles and grow with sell-through.

NYC menu pricing is anchored on the highest-priced item the customer compares to — your most expensive entree sets the perceived ceiling for the entire menu. If your top entree is $42, the rest of the menu reads as 'reasonable'; if it's $58, the same items at $32 read as 'cheap'. Use no dollar signs and no decimals (research from Cornell Hospitality Quarterly: removing $ and decimals lifts spending 8-12%). Price avoidance points are $19, $29, $39, $49 — moving a dish from $39 to $42 typically doesn't change order rate but adds direct margin. Anchor a $24 starter and a $48 entree on the menu even if you don't expect them to sell heavily — they shift the sweet-spot dishes ($16-19 starters, $32-38 entrees) into the value zone. Audit the menu pricing every 90 days against food cost movement; ingredient inflation in 2024-2026 has been 6-12% YoY in several categories.

NYC tracks ahead of national norms here — at minimum, 25-30% of your dishes should be naturally or easily-modifiable to vegan, vegetarian, gluten-free, and dairy-free. The market reality: ~10% of NYC diners self-identify as vegetarian or vegan in 2026, ~6% report celiac or gluten sensitivity, and ~25% are an 'accommodator' (eating out with someone in one of those groups). A menu that excludes those 25% loses the entire party booking. Designate dishes clearly with symbols (V, VG, GF, DF) — the FDA's voluntary major-allergen disclosure framework is widely adopted in NYC and avoiding misrepresentation reduces 311/DOH complaints. Avoid the trap of one tortured 'vegan option' bolted onto the menu — design 3-5 dishes that happen to be naturally vegan or GF and are good in their own right. Train the floor on cross-contamination protocols (separate fryer for GF, no flour-dusted surfaces) — cross-contact lawsuits run $25-150K in settlement.

Two principles: name the hero ingredient and one preparation, then stop. 'Sungold tomato, smoked ricotta, basil oil' beats 'A vibrant celebration of summer's bounty featuring locally-sourced heirloom Sungold tomatoes from Norwich Meadows Farm, hand-pulled smoked ricotta, and house-made basil oil drizzle'. Cornell research shows descriptive labels lift sales 27%, but only if they describe the food, not the farm-bio. Cite a farm or producer only for 1-3 hero items per menu, not every dish — overdoing it reads as performative. Cap descriptions at 12-15 words. Use proper noun ingredients sparingly (one 'Calabrian chili' or one 'Castelvetrano olive' per menu, not five). Avoid these dead words in 2026: 'curated', 'elevated', 'chef-driven', 'seasonally-inspired', 'thoughtfully sourced', 'craft' — all read as filler.

Quarterly menu refreshes are the NYC norm (4 changeovers/year), with 30-50% of dishes rotating each quarter and the signature/anchor items staying. Ingredient-driven concepts (vegetable-led, seafood) can do 6-8 changeovers/year tracking actual harvest windows; protein-led or pasta-led concepts can stretch to 2-3 per year. The mistake is changing too much — swapping more than 60% of the menu in one revision confuses repeat customers and erases your sell-through data. Print menus on a Wednesday for a Thursday rollout (slowest to slowest day). Budget $400-1,200 per menu reprint depending on stock and run, plus 6-12 hours of chef R&D time per new dish — total quarterly menu refresh cost typically $3-8K including staff training. Don't print 'seasonal menu' on the menu itself — it dates the document and forces premature reprints.

Hire 9-14 weeks before opening so the chef can build the menu, hire the line, and run 4-6 weeks of testing — anything tighter and you'll open with an unexecuted concept. Three credibility signals beyond resume: (1) 3+ years at a restaurant in your direct cuisine category at the price tier you're targeting (a chef who ran sauté at a $200 PP Italian spot can run a $90 PP Italian; not a $90 PP Sichuan); (2) demonstrated menu authorship (ask for menus they wrote, not just cooked from); (3) a tasting interview where they cook 4-6 dishes from your concept brief on a real budget and timeline. Compensation in 2026 NYC: chef de cuisine $95-145K, executive chef $140-220K (small concept) to $250-450K+ (large or hotel), plus 1-3% revenue or profit share for executive level. Lock IP terms in writing — recipes developed during employment are work-for-hire only if the contract says so explicitly under NY law.

Two-page document, 9 sections: (1) concept brief copy (so the designer reads it fresh), (2) target customer persona, (3) operational must-haves (seat count, kitchen size, bar size, ADA bathroom count, BOH back-of-house dimensions), (4) program adjacencies (entry-host stand-bar-dining-kitchen flow), (5) reference imagery — 25-40 images on a single mood board PDF, captioned with what specifically you like ('this banquette pitch', not just 'this room'), (6) brand identity package (logo, colors, type), (7) budget range and not-to-exceed, (8) timeline including DOB filing, MEP, and target opening date, (9) explicit exclusions ('no exposed Edison bulbs, no subway tile, no shou-sugi-ban'). Hand it to 3 architects for fee proposal — fees range $35-180/sq ft for full architectural + interior design on a NYC restaurant, or fixed $80-450K for a typical 2,000-4,500 sq ft buildout.

Working NYC density: 16-22 sq ft per dining seat (including aisles and circulation), 4-6 sq ft per standing-bar guest. A 2,500 sq ft total restaurant typically yields 60-90 dining seats + 8-14 bar seats after kitchen (25-32% of total area), bathrooms (NYC requires 2 single-occupant or gender-separated for >35 seats per 2014 NYC Plumbing Code Table 403.1), back-of-house, and circulation. The seat count drives the entire pro forma — too few and you can't cover rent; too many and the room feels empty at 6:45pm. Sweet spot for a NYC dinner concept is 55-85 dining seats: small enough to feel curated, big enough to hit $4-7M revenue at 1.8-2.4 turns. Run the math 3 ways before locking layout: at design capacity, at 75% capacity (your real average), and at 50% capacity (your slow-Tuesday floor) — the concept must work at 75%.

NYC working ratios: 25-30% of total restaurant area for the back-of-house including kitchen, prep, dishpit, dry storage, walk-ins, office, and staff bathroom. A 2,800 sq ft restaurant typically dedicates 700-840 sq ft to BOH; cutting below 22% saves rent but creates an unworkable kitchen that caps your covers. Within BOH, the kitchen line itself is roughly 60-70% (450-600 sq ft for a 70-seat dinner concept). Tasting-menu and fine-dining concepts need more (35-40% BOH because of plating space and prep complexity); fast-casual and pizza concepts need less (18-22%). The single most expensive design mistake is undersizing the kitchen or dishpit — you cannot retroactively add 100 sq ft of kitchen, and it will limit your throughput forever. Brief the architect with a target covers/hour and let the kitchen designer (a separate consultant, often Whaley or Strategic) size accordingly.

Target 68-76 dBA at the table during peak service — Eater and Infatuation now publish dB readings in reviews and 80+ dBA gets called out as 'unbearably loud'. NYC restaurants with hard surfaces (concrete floor, tin ceiling, exposed brick, no curtains) routinely measure 82-92 dBA at full service, which kills repeat-visit rate among 35+ diners. Design treatments that work: acoustic ceiling clouds or panels (Soft Sound, Snowsound, Baux — $35-110/sq ft installed), upholstered banquettes against hard walls (banquette material absorbs 40% of mid-range frequencies), drapery on at least one wall, and floor area rugs in soft seating zones. Budget 3-6% of total buildout for acoustic treatment; $25-60K is typical for a 2,500 sq ft room. Hire an acoustic consultant ($4-12K) for any room over 80 seats — the math is cheap insurance against a 'loud' reputation that takes 18 months to undo.

Layer 4 light types: ambient (room-fill, dimmer-controlled), task (above the bar, kitchen pass, host stand), accent (artwork, architectural features), and decorative (pendants, sconces, candles). Color temperature 2400-2700K for dinner (warmer reads as romantic; cooler reads as clinical), 3000K acceptable for brunch/lunch concepts. Use 0-10V or DALI dimming, not on/off — programmable scenes for lunch / dinner / late-night that the GM can set with a tap. Hire a lighting designer ($8-35K for a typical NYC restaurant) — the architect can brief lighting but specialist designers like Lighting Workshop, Tillotson Design, or Focus Lighting deliver dramatically better results for the spend. Expect $40-120K in fixtures and $25-80K in installation labor; total lighting line item is typically 5-9% of buildout. Test the dim curves at 5pm, 7:30pm, and 10pm before sign-off.

Six material categories that fail in 12-24 months at NYC volume: (1) untreated wood tabletops — spec sealed wood with marine-grade urethane or move to stone/laminate; (2) light fabric upholstery — spec performance fabrics (Crypton, Sunbrella, Maharam Energy) rated 50K+ double-rubs; (3) brass that's not lacquered — patina is fine if intentional, but expect tarnish complaints; (4) terrazzo or limestone in high-traffic zones without sealing — re-seal annually $4-12K; (5) banquette piping in white or cream — spec dark contrast piping or wipeable vinyl; (6) painted walls below 4 ft — spec wainscot, tile, or wipeable epoxy paint in service zones. Budget 8-15% of FF&E line for replacement reserve in years 2-3. NYC restaurant traffic is 3-5x residential, and design choices that work in a magazine shoot fail in operation.

NYC restaurant buildout in 2026 runs $400-900/sq ft for a casual to mid-tier concept and $900-1,800+/sq ft for fine-dining or hotel-grade work — total project costs for a typical 2,500 sq ft restaurant range $1.0M to $3.5M+. The split is roughly 35-45% construction (walls, MEP, HVAC, plumbing), 18-25% kitchen equipment, 12-18% FF&E (furniture, fixtures, equipment for dining room), 8-12% design fees (architect + interior designer + MEP engineer + expediter), 5-8% permits and DOB/FDNY filings, 8-15% contingency. NYC-specific cost drivers: union vs non-union labor (union adds 25-45%), 421-a / Local Law 92 / Local Law 94 retrofit requirements for older buildings, and ConEd electrical service upgrade ($25-180K if you need >200 amp). Get 3 GC bids, demand line-item detail, and add 15% contingency on top of the highest bidder's number — every NYC restaurant project goes over.

ADA Title III + 2010 ADA Standards plus NYC Local Law 58 of 1987 require: at least one wheelchair-accessible entrance (32-inch clear width minimum), 5% of dining seats accessible (minimum 1 table) with 30x48-inch clear floor space and 27-inch knee clearance, accessible route through the dining room with 36-inch minimum aisle, and at least one accessible single-user bathroom OR accessible stalls in each multi-user bathroom (60-inch turning radius). NYC enforces via Department of Buildings on plan review and accepts ADA complaints via 311 and the AG's office. Common operator mistakes: bar-only seating with no accessible bar (need a 34-inch high section with 27-inch knee clearance for at least one position), entry steps without ramp, bathrooms below grade without lift. Retrofit costs $15-80K; designing it in from day one adds <$5K. Drive-by ADA lawsuits in NY are real — settlements run $4-25K plus your legal fees.

30-60 pages, 8 sections: (1) brand essence — concept brief, mission, voice in 4 pages; (2) logo system — primary, secondary, monogram, clear-space rules, do/don't examples; (3) color — full palette with hex/RGB/CMYK/Pantone and usage proportions; (4) typography — typefaces, weights, hierarchy examples, body/display ratios; (5) photography — style references, do/don't gallery; (6) brand patterns and graphic system; (7) applications — menu, signage, packaging, uniforms, social templates; (8) tone of voice — written examples of headlines, captions, email replies, error messages. Deliver as a clickable PDF + a Figma library + a Google Drive of source assets. Budget $15-45K to produce as part of brand identity package, or $8-20K standalone if you have the components. Update annually — version it (v1.0, v1.1) so the team knows what's current.

Three-tier system: (1) source files (AI, Figma, RAW photos, font installers) in a private Google Drive or Dropbox restricted to brand owner + designer — these are master files and should never be sent to vendors directly; (2) production-ready files (PDF logos, packaged photo crops, exported PNG/JPG, Word/PPT templates) in a shared 'Brand Assets' Google Drive accessible to internal staff and approved vendors; (3) public-facing press kit (web-resolution JPG/PNG of logo + 8-15 photos + 1-page bio) on a /press page of your website behind a simple email gate or via a Dropbox public link. Tag every asset with usage right (web only, print only, all use), version date, and approval status. Lost or unmanaged assets cost you in vendor confusion, off-brand outputs, and recreated work — a $300-800/yr Brandfolder or Frontify subscription is worth it once you have 5+ vendors touching the brand.

Hire a hospitality-specialty photographer (NYC rates: $2,800-7,500/day for food + interior shoot) and shoot 3 categories on opening day: (1) hero food — 12-18 dishes plated correctly under your service lighting; (2) interiors — 15-25 frames including the empty dining room, the bar, the entry, the kitchen pass; (3) people — 8-15 frames of staff in motion, but no posed forced-smile shots. Shoot with natural light supplemented by tungsten/LED that matches your service color temp (2400-2700K), not stock studio strobes that flatten food. Avoid the 2024-2026 trend of overhead-flat-lay food photography — it dates fast and obscures texture. Get all images in 3 crops (square 1:1, vertical 4:5, horizontal 16:9) for IG, Resy, and press. Plan a refresh shoot at month 6 (capture menu evolution) and annually after — total photography spend year 1 is $8-25K including pre-opening, opening, and 6-month refresh. NYC photographers to know: Eric Helgas, Eric Wolfinger (when in NYC), Liz Clayman, Andrew Bui, Daniel Krieger.

Build 12 production templates into the brand system at launch: (1) menu (food + cocktail + wine), (2) reservation confirmation email, (3) post-visit thank-you email, (4) press kit one-sheet, (5) IG square post (4 variants), (6) IG story (4 variants), (7) IG carousel covers, (8) gift card design (digital + physical), (9) business card, (10) staff name tag, (11) PowerPoint pitch deck (for partnerships and press), (12) event proposal PDF. Build them in Figma or Canva for Teams ($150/yr per seat) so non-designers can update without breaking the brand. Lock the template file with text-only edit permissions and put the master in the brand library. The hours saved over year 1 typically pay back the $8-18K it cost to build templates 4-5x over.

Day -90: announce the project on a single IG post (chef, concept, neighborhood, target month — no exact date) and capture an email signup at a /coming-soon landing page. Day -75 to -45: 6-10 IG posts of buildout, recipe testing, named hires; pitch Eater NY first-look at Day -45 (they need 3 weeks lead time). Day -45 to -21: confirmed press hits (Eater first-look, Grub Street roundup), Resy listing live with future-date booking, soft-open invites to 80-150 industry friends + media + neighborhood VIPs. Day -21 to -7: 3-5 friends-and-family seatings (35-60 covers each, comped), refine service flow. Day -7 to 0: paid soft-open at 50-70% capacity for 4-7 days at $25-50 cover or comped wine pairing. Day 0: full opening, Resy at 100%. Total pre-launch marketing spend $8-35K plus PR firm if used. The mistake is launching cold with no Day-90 runway — you need 6-8 weeks of buzz before doors open or the opening week is dead.

PR firm is worth it if you have a real press hook (named chef with prior coverage, unusual cuisine, notable backers, design by a published architect) and your buildout is over $1.5M. NYC hospitality PR firms charge $5-15K/month with 3-6 month minimum — total $30-90K for the launch window. The good firms (BoCo Communications, Becca PR, Wagstaff, Bullfrog & Baum, Magrino) have direct relationships with Eater, NYT Food, Grub Street, Infatuation, NY Mag, and trade press, and they can secure 4-8 placements in the first 90 days that would take you 200+ unanswered cold emails to attempt. If you don't have a press hook, the same money is better spent on opening party + influencer dinners + paid media. Self-PR works for concepts under $500K buildout where the chef-owner has personal media skill — outside that, the PR firm is the leverage.

Run 2 soft-launch phases: (1) Friends & Family — 3-5 nights, 35-60 covers each, 100% comped, no press, no Resy, the goal is to find systems failures (POS bugs, ticket-time gaps, wrong glass at table, broken bathroom lock); (2) Soft Open — 5-10 nights at 50-70% capacity, $0-50 nominal cover OR menu at 50% off, Resy live but capped, no press reviewers (you'll know them by name — keep a list). Walk the floor every service. Debrief the team for 30-45 min after each service with a one-page punch list. Don't invite the NYT Critic, Pete Wells's successor (Brian Velasco), Steve Cuozzo, Adam Rapoport, or Hannah Goldfield during soft open — formal reviewers traditionally wait 30+ days post-opening but operators occasionally trip them in early. Total soft-launch comps cost $15-50K and prevents $200K+ of opening-month revenue loss from broken systems.

Yes — but pick 8-15 carefully, not 60. The NYC accounts that genuinely move Resy bookings in 2026: @theinfatuation (1.4M+), @grubstreet (~1M), @eater (~3M), @nytfood (~1M), Pete Wells (now critic emeritus, residual influence), @platt (Adam Platt, ~80K), and a tier of operator-respected micro-accounts (@thefoodb1tch, @newforkcity, @manhattan_dines, @brunchboys, @nycfoodhunter — each 100K-1M). Influencer threshold for measurable cover lift is roughly 80K+ engaged NYC followers. Invite to a dedicated press dinner (15-25 seats, 4-course menu, 90 minutes, no other guests) — this is a working dinner, not a party. Comp the meal but never pay for posts (FTC §16 CFR Part 255 requires #ad disclosure for paid posts which kills the editorial credibility). Track Resy bookings in the 7 days after each post to measure actual lift; budget $4-12K for the press-dinner program.

Limited yes for paid social (Meta + TikTok), almost never for paid search. Allocate $3-8K in pre-launch and $2-5K/month for the first 90 days into Meta/Instagram boosted posts targeted to 1-mile radius around your address, 25-55 age, 'foodies' / dining-out interest. Expect $1.20-3.50 per IG profile click and $0.40-1.20 per landing-page email signup. TikTok organic outperforms paid for restaurant content — invest in 2-3 in-house short-form videos a week instead of TikTok ads. Google Search ads almost never pencil for restaurants because direct-name searches are already finding you and category searches ('best Italian Williamsburg') click into Yelp, Resy, and Google Maps before paid links. The exception is a hotel restaurant where Google Search ads on hotel-guest searches can deliver $2-5 per qualified click — worth $1-3K/month if you have hotel partnership.

Three deal structures dominate NYC hotel F&B: (1) full lease — operator pays rent ($75-200/sq ft for a quality 4-star NYC hotel ground floor), keeps all revenue, full P&L control, but no comp/IRD/breakfast obligation unless negotiated; (2) management agreement — hotel owns the F&B P&L, operator earns 3-7% of net revenue + 8-15% of GOP (gross operating profit), takes brand exposure but no upside; (3) brand license + JV — operator licenses the name and concept (royalty 4-8% of revenue) plus equity participation 20-40%, hotel runs operations under operator brand standards. Negotiate hard for: separate restaurant entrance and signage, control over hours, control over menu/pricing, exclusion from required IRD (in-room dining) or breakfast service unless paid separately, exit rights if hotel sells. Le Bernardin's deal at the Equinox Hotel Hudson Yards, The Modern at MoMA, and Le Coucou at the Ludlow are the comparable structures to study. Get a hospitality lawyer ($800-1,800/hr in NYC) — these contracts run 80-150 pages.

Standard hospitality brand license terms in 2026: 4-8% of gross revenue royalty + $150-400K upfront license fee per location + a development territory commitment (e.g., minimum 3 locations in 5 years in defined market). Carbone licensed Miami (2021), Las Vegas (2022), Hong Kong, Riyadh, Dallas — Major Food Group's deal structure runs ~6% royalty plus chef-development support. For your license to be worth that, you need a registered Class 43 trademark (ideally also Class 25, 30), a documented brand book + operations manual, and a track record of 3+ profitable years at the original. License agreements should require: brand standards manual compliance, mystery-shopper rights, menu approval, marketing approval, indemnification for trademark infringement caused by licensee, and a termination right with 90-day cure period. Quote no less than 5% royalty even on a friend deal — anything less and you can't fund the QA team that protects the brand.

Three collab types and the rules for each: (1) Chef-guest residencies — short (1-4 nights), one ticketed seating, 50/50 revenue split or fixed honorarium $3-15K depending on chef stature, exclusive press push, no brand-name conflict. Risk: low. Lilia x Don Angie, Frenchette x Eleven Madison Park alums all use this format. (2) Product collabs — co-branded packaged goods (sauces, granola, coffee beans, glassware, candles) with a CPG partner; structure as a license with 6-12% royalty to the brand owner, term 12-24 months. Risk: medium — protects via approval gate on each SKU. (3) Fashion / lifestyle crossovers — limited-edition merch with apparel brand, restaurant edition of a wine, exclusive cocktail at a hotel bar. Cap collabs at 4-6 per year per brand to preserve scarcity. Every collab needs a one-page agreement covering: trademark use approval, term, territory, revenue split, indemnification, and termination. Charging $0 for early collabs is fine; charging $0 for the 5th collab undervalues the brand and trains partners.