As a 450 m² ground-floor venue (~150 seats), the base case is genuinely attractive — ~$184k EBITDA (24% margin) and a ~2.5-year payback on ~$450k of fit-out. But it rests on pulling ~190 covers/day in an off-promenade location. The conservative case (~135/day) only covers cash costs, not the fit-out. Before signing, validate the one number everything hinges on: achievable covers/day.
What 450 m² becomes
A single-floor venue of this size supports roughly 150 seats once you net off the kitchen and back-of-house. The layout that fits the strategy: a street-facing café/bakery counter for daytime and winter trade, behind it a main dining hall for lunch and dinner.
| Zone | Area | Seats |
|---|---|---|
| Main dining hall | ~190 m² | ~115 |
| Café / bakery counter + grab-and-go | ~50 m² | ~25 |
| Entry / waiting / host | ~20 m² | — |
| Front of house | ~260 m² | ~140–150 |
| Kitchen, prep, cold/dry store | ~110 m² | — |
| Staff, WC, circulation | ~70 m² | — |
| Back of house | ~180 m² | — |
| Total | ~450 m² | ~150 |
Planning ratios: ~58% front-of-house, ~40% back-of-house, ~1.8 m² per seat. A real architect's plan will shift these ±10%.
The numbers — built bottom-up
Everything below is an indicative model, not a forecast or audited figure. The structure is industry-standard; the inputs are calibrated to Samarkand (low wages, low food cost, ~12,500 UZS/$). Replace the four driver assumptions with your real quotes and the rest recomputes.
Driver assumptions (base case)
| Driver | Value | Basis |
|---|---|---|
| Seats | 150 | ~260 m² FOH ÷ 1.8 m²/seat |
| Operating days | 360 | Closed ~5 days/yr |
| Avg check (blended) | $11 | Coffee/brunch ~$6 · lunch ~$9 · dinner ~$15 |
| Covers/day (annual avg) | ~193 | ~1.3 seat-turns/day, seasonally weighted |
Revenue — three scenarios
| Scenario | Covers/day | Covers/yr | Avg check | Revenue/yr |
|---|---|---|---|---|
| Conservative | ~135 | ~48,600 | $10 | ~$490k |
| Base | ~193 | ~69,400 | $11 | ~$760k |
| Optimistic | ~240 | ~86,400 | $12 | ~$1.04m |
Covers/day is the single biggest swing — and the hardest to guarantee off the promenade. Treat the conservative case as the downside you must survive.
Build cost (CapEx)
| Item | Estimate |
|---|---|
| Fit-out & finishes (450 m² × ~$450) | $200,000 |
| Commercial kitchen equipment | $90,000 |
| Furniture & fixtures (~150 seats) | $60,000 |
| Bakery & coffee equipment | $25,000 |
| POS / IT / signage / branding | $25,000 |
| Pre-opening (licences, hiring, training, opening stock, launch) | $30,000 |
| Contingency (~5%) | $20,000 |
| Total CapEx | ~$450,000 |
≈ $1,000/m² all-in (≈ 5.6 bn UZS). Range $360k–$560k depending on finish level. Excludes the building shell (assumed provided by the developer).
Annual P&L — base case
| Line | % of rev | USD / yr |
|---|---|---|
| Revenue | 100% | $760,000 |
| Food & beverage cost (COGS) | 32% | −$243,000 |
| Gross profit | 68% | $517,000 |
| Labour (~30–35 staff) | 22% | −$167,000 |
| Rent (450 m² @ ~$8/m²/mo) | 6% | −$45,000 |
| Utilities (power, gas, water) | 4% | −$30,000 |
| Marketing (high — you pull demand) | 5% | −$38,000 |
| Other opex (R&M, supplies, delivery, insurance, bank) | 7% | −$53,000 |
| EBITDA | 24% | $184,000 |
| Depreciation / amortisation (CapEx over ~6 yr) | 10% | −$75,000 |
| EBIT (pre-tax profit) | 14% | $109,000 |
| Tax (placeholder ~15% — confirm regime) | 2% | −$16,000 |
| Net profit (est.) | ~12% | ~$93,000 |
Unit economics & key ratios
| Metric | Value | Read |
|---|---|---|
| Prime cost (COGS + labour) | 54% | Healthy (target < 60–65%) |
| EBITDA margin | ~24% | Strong for full-service |
| Revenue / seat / yr | ~$5,070 | Low check market — normal here |
| Revenue / m² / yr | ~$1,690 | — |
| Simple payback | ~2.5 yr | CapEx ÷ EBITDA |
| Pre-tax ROI | ~24% / yr | EBIT ÷ CapEx |
| Cash break-even | ~110 covers/day | ≈ $432k revenue — covers cash costs |
| Full break-even | ~145 covers/day | ≈ $566k revenue — incl. depreciation |
The seasonality problem in cash
Base-case monthly revenue. The business earns its year in Apr–May and Sep–Oct; Dec–Feb run at or below cash break-even.
A booming market with a brutal winter
Samarkand drew ~7.2M visitors in 2024 (2.5M foreign), double 2019, and average stay rose to 3 days — a genuine tailwind. But tourist demand is sharply seasonal: prime in spring (Apr–May) and autumn (Sep–Oct), hot but busy in summer, and a cold, quiet Dec–Feb. That seasonality is exactly what the cash chart above reflects, and why a local-regular base is not optional.
Relative tourist-demand index by month. Shape is evidence-based (best-months guides, climate, "1M+ arrivals/month from April"); exact values are a model, not measured Samarkand data.
~15 venues within 1 km — and a service-quality gap
Rivals cluster along University Boulevard and around Gur-e-Amir, north-east of and uphill from the site. The highest-volume tourist venues sit at only 3.8–4.0 stars — the city's chronic weakness is service consistency, which is your cheapest edge.
| Venue | Type | Rating | Reviews | ≈ Dist |
|---|---|---|---|---|
| Boulevard Restaurant & Bakery | European, bakery-café | 4.7 | 115 | 0.8 km |
| Platan | Eastern European | 4.0 | 967 | 1.4 km |
| Emirhan | BBQ / European | 4.2 | 405 | 1.2 km |
| Old City | European / Central Asian | 3.9 | 274 | 0.9 km |
| Oasis Garden | Steakhouse | 3.8 | 242 | 0.9 km |
Boulevard Restaurant & Bakery (4.7) is the quality benchmark and closest direct rival; the rest compete on volume, not consistency.
Where the city's menu has holes
Opportunity score (0–10) = demand × scarcity × fit. Research-based estimate, not hard data.
An all-day café-bakery + dining hall
With no rooftop, the whole concept lives on one floor — which is fine, because the daypart stack is what makes the economics work, not the view.
Morning–afternoon
specialty coffeebakery counterall-day brunchhealthy / vegLunch–dinner
signature plovpremium-casual Uzbekfamily dininghalalWinter engine
café + bakerylocal regularsprivate eventsdeliveryWhat can break the model
| Risk | Severity | Mitigation |
|---|---|---|
| Covers/day below ~145 (off-promenade pull) | High | Maps/SEO day one, hotel referrals, delivery, boulevard signage; test footfall before full build-out |
| Winter cash trough | Med-High | $40–60k buffer; café/bakery + local loyalty; events |
| Fit-out overrun | Medium | Fixed-price contractor; ~5% contingency; phase non-essentials |
| Ramp-up (slow first 6–12 months) | Medium | Pre-marketing; soft open; the payback assumes a mature year, not year 1 |
| Service inconsistency (industry-wide) | Opportunity | Train hard, hire English speakers — beats the 3.8–4.0 incumbents |
Note: the payback (~2.5 yr) is on a mature base year. Realistically year 1 ramps from a lower base, so cash payback from opening is more like 3–4 years even in the base case.
Validate the drivers, in this order
Achievable covers/day
The whole model hinges here. Count footfall on the lane at peak and 7pm; pull competitors' busy hours on 2GIS/Yandex; run a weekend pop-up or bakery counter to test real pull before building.
Real CapEx
Fixed-price fit-out + kitchen quotes from 2–3 local contractors. This sets the payback.
Rent / ownership terms
Are you renting the ground floor or do you own it? Negotiate a rent-free fit-out period. Ownership lifts EBITDA ~$45k.
Tax regime & licences
Confirm turnover-tax vs VAT+profit-tax at ~$760k revenue, plus catering/alcohol licensing — it moves net profit several points.
Pre-market & soft-open
Open the Instagram/Google/Yandex/2GIS presence now; lock hotel-concierge referrals; soft-open to de-risk the ramp.
Market data cross-referenced: TripAdvisor, eva-darling.com, Advantour/Wanderlog/Eurasia.travel, Uzbekistan stats & press (xs.uz, Euronews/UNWTO), Numbeo/Hikersbay prices, 2GIS. Financial model built on standard restaurant cost ratios calibrated to Samarkand wage/price levels.
Honesty notes: (1) All figures are an indicative model for decision support, not a forecast, audited accounts or investment advice — every driver must be replaced with local quotes. (2) Covers/day is assumed, and is the dominant uncertainty given the off-promenade location. (3) The two scene images are AI visualizations faithful to the developer's renders, not architectural drawings. (4) The seasonality shape is evidence-based; exact monthly numbers are modelled. (5) Tax is a placeholder pending regime confirmation. (6) Payback uses a mature base year; year-1 ramp is slower.