The Master Business Plan

Kolshee is an operating system, not a product. Operating systems do not have a single revenue line; they have a budget of revenue lines that share a substrate.

The read starts in the basket and moves up: a block-level view of demand, then pooled purchase orders across stores quietly buying the same SKUs, then container bookings and upstream commitments shaped by the same signal. Fleets and warehousing stay off our balance sheet — we coordinate the contracts and loads of operators already running them, instead of competing with their capital base. What we add is a decision layer, not a sixth party.

A hub is the physical form of the coordination layer: a store, a kitchen, ingredient storage, a procurement desk, dispatch coordination — co-located so the same staff can see the basket, the back room, and the inbound container. Restaurants joining a hub don't stand up their own fleets, warehouses, or procurement desks; they plug into shared lines that already serve the surrounding stores.

Before a restaurant brand is admitted, we validate the boring things: ingredient availability and substitutability, supplier consistency over a full quarter, packaging and cold-chain standards, regional distribution coverage, kitchen-side workflows, and unit economics that survive a slow week. A brand that clears the checklist can move into a new metro with materially less operational risk than opening alone.

Ethnic and halal restaurants buy a lot of the same goods that move through the grocery stores on the same block: 20kg rice bags, oil tins by the case, weekly halal protein, regional spices and canned staples. Today each kitchen negotiates alone and ends up paying 12–22% more than the store next door pays the same importer. We fold the restaurant into the grocery side's pooled PO instead of opening a parallel lane.

How it actually runs: the restaurant registers a core ingredient list and a weekly burn rate. The system matches that list against active buying windows for nearby stores, adds the restaurant's quantity to the next container, and delivers from the same regional DC — often on the store's existing daily drop. No separate procurement desk, no restaurant-only warehouse, no extra delivery contract.

The economics are direct. The kitchen takes 8–15% out of food cost and frees two-to-three weeks of working capital that used to sit in safety stock. Kolshee earns a 3–5% brokerage fee (Stream 6); the partner store sees a fuller container and a lower unit cost on its own orders.

The second-order value is brand transfer. A restaurant proven in Amman or Istanbul can open inside a Detroit or Houston hub without rebuilding its supply line — ingredients arrive on the same lane already serving the surrounding stores, with the same quality standards and the same traceability.

We don't own drivers and we don't run a fleet. Delivery is fulfilled across three partner tiers, and the partner is chosen per-order on cost, speed, and reliability — not by exclusive contract.

• ▸Tier 1 — Fleet-as-a-Service: DoorDash Drive, Uber Direct, Roadie. Open APIs, transparent per-trip pricing, no marketplace take. Kolshee pays only the direct delivery fee (≈ $4–7 for a short-radius order) and either passes it to the shopper or absorbs it inside the store's net.
• ▸Tier 2 — Local partners: regional couriers, independent taxi/van fleets, ethnic and halal last-mile operators. Direct contracts averaging 15–25% below the majors, with deeper coverage in the neighborhoods we actually serve.
• ▸Tier 3 — Store batching: orders from clusters of nearby stores are batched into a single run two or three times an hour. Effective per-order cost drops to ~$2–3.

Routing logic. At order time, the system queries live quotes from every available partner in the radius, scores each on expected delivery window and 30-day reliability, and assigns the order to the lowest-cost partner inside the acceptable quality envelope. No exclusive deal, no paid tier, no editorial preference — it's a pure algorithmic award.

Where the margin difference comes from. DoorDash and Instacart take 25–30% of the basket because they sit as intermediaries between the store and the customer and own the interface. Kolshee isn't an intermediary — the store is a network partner and the customer interface lives inside the Kolshee app. Our 6–8% take covers logistics coordination only; delivery passes through at cost. The store keeps roughly 18–22 additional margin points on the same order.

This is the layer that turns a restaurant from a customer into infrastructure. An original restaurant from Aleppo, Tehran, Lahore, or Istanbul opens a branch inside diaspora neighborhoods in the U.S. using three assets that already sit inside our network: the same ingredients (arriving on the same weekly container that serves the surrounding stores), the same supply line, and physical space inside a partner grocery store — typically 40–80 m² of underused floor in ethnic and halal stores.

Economics. Opening a Detroit branch traditionally runs $300–700K (lease, full kitchen build-out, licensing, cold-start procurement, hiring). Inside the Kolshee model the all-in falls to $40–80K because the space exists, the license is shared with the host store, ingredients ride the existing supply line, and the kitchen is partially fitted. That's not an improvement — it's a category change in entry cost.

Authenticity logic. Because the ingredients come from the same importer that serves the origin restaurant — visible in our data — we can guarantee that the kabsa served in Houston is cooked with the same rice and spice used in Riyadh. That's a verifiable claim neither UberEats nor Kitchen United can make.

Value split. The host store collects a space rent, foot traffic uplift, and incremental ingredient sales to diners who liked what they ate and grab the SKU on the way out. The origin brand takes a revenue share with quality rights. Kolshee charges an upfront license fee plus 8–12% of branch revenue for end-to-end coordination (ingredients, space, shared license, diaspora marketing).

The line we don't cross. We don't operate restaurants, don't own them, don't hire chefs. We coordinate — same posture as with the fleets. Kitchen management is either a transferred team from the origin during a training window, or a local team trained on the brand's playbook. Settlement runs through the host store's ledger.

Read the deep-dive: Bayt Halab → Dearborn (full worked example + data & ops flow diagram) →

Operational infrastructure ecosystem: how grocery, restaurant cells, procurement, and delivery run as one synchronized layer →

Kolshee gives the software away. The software is the trap, not the asset. The asset is the live event spine — every scan, every reorder, every shipment, every customer interaction — that flows through the software once it is installed. Once a store is on Kolshee, seven independent revenue lines activate against the same substrate.

The consumer-facing app is the visible top of the iceberg. Kolshee runs aggregated, intelligently batched fulfillment from clusters of nearby mesh stores, generating consumer GMV that the platform takes a clean rent on. Unlike DoorDash, the take-rate sits cleanly inside the platform's contribution to logistics + AI batching, so the merchant nets more than they would on any incumbent.

Brands, distributors, manufacturers, importers, regulators, and analysts subscribe to query-tier access to anonymized cohorts of mesh signal: SKU velocity by metro, substitution graphs, cultural calendar elasticity, regional pricing curves, importer performance leaderboards. Pricing is structured by query depth and refresh cadence. This is the Bloomberg of food.

Once Kolshee can see SKU demand at continental scale, it can commission its own private-label production of the highest-velocity heritage SKUs (rice, oils, spices, dairy, halal proteins, regional sweets) directly from origin producers. Manufacturer-grade margins (22–35%) accrue to Kolshee, while merchants gain access to a quality-controlled, culturally authentic, lower-cost in-house line.

Kolshee converts its supplier graph into an active brokerage layer: cross-border shipments routed through Kolshee earn a brokerage fee, with FX, customs, and lane optimization handled in-platform. Middle Eastern, South Asian, North African, and Latin American producers gain direct, programmatic access to U.S. mesh demand for the first time.

The AI policy engine — forecasting, restocking, pricing, staffing, route optimization — is exposed as an API consumed by non-mesh stores, chains, restaurants, distributors, and regulators. Pricing is per-call with volume tiers. This is the Stripe of food intelligence.

Restaurants — especially the dense ethnic and halal restaurant ecosystem clustered near the same diaspora communities — buy bulk SKUs through the Kolshee mesh on a B2B procurement layer with predictable pricing, intelligent reorder, and guaranteed cultural-authenticity provenance.

An origin restaurant from Aleppo, Istanbul, or Lahore opens a branch inside dedicated space in a Kolshee partner store, fed by the same ingredient lane that serves the surrounding stores. Kolshee charges an upfront license fee at branch opening, then 8–12% of ongoing branch revenue for end-to-end coordination (ingredients, space, shared license, diaspora marketing). The origin brand takes a revenue share; the host store takes space rent and traffic uplift. Market entry cost for the brand drops from hundreds of thousands to tens of thousands, which turns geographic expansion into a commercial decision rather than a capital bet.

The pilot is sized to three U.S. metros that share three traits: an Arab / South Asian diaspora density above 150K, 40+ independent ethnic and halal grocery stores, and at least one regional importer servicing those stores daily. Sequence: Dearborn, MI → Paterson, NJ → Houston, TX. Dearborn first because it has the highest Arab density in North America and the shortest distance between store, shopper, and restaurant.

Store partnership steps (4–6 weeks per store, parallelizable):

• ▸Week 1 — Diagnostic: on-site visit, shelf inventory, pull last three supplier invoices, measure price gap vs. simulated pooled-container pricing.
• ▸Week 2 — Pitch: personalized owner report showing expected annual savings (avg $22–38K for a mid-size store), two-page contract, no exclusivity.
• ▸Week 3 — Install: Kolshee POS deployment, catalog import, bank linkage, 3-hour staff training.
• ▸Week 4 — Parallel run: store keeps legacy suppliers while joining the first pooled PO. Reconcile post-delivery to prove savings empirically, not on a slide.
• ▸Week 5–6 — Switch: migrate 60–80% of purchase volume to the mesh, keep one fallback supplier. Store enters the consumer app once catalog stability hits 95%.

Migration-brand selection (starts month 5, targets 6 brands across the 3 metros):

• ▸Criteria: 10+ years operating in country of origin, 2+ existing branches, menu anchored on 8–12 ingredients already serviced by an importer in our network, owner willing to sign a soft franchise (no owned units required).
• ▸Target slate: a kabsa brand from Riyadh, a kebab brand from Aleppo, a biryani brand from Lahore, a Persian-kebab brand from Tehran, a shawarma brand from Istanbul, a pastry brand from Beirut.
• ▸Sequence: month 5 outreach, month 6 origin-site visit + recipe documentation, month 7 sign one license + fit out the in-store space, month 9 open the first pilot branch in Dearborn with a transferred crew on a 4-week handover from the origin restaurant.

Decision gates. Moving from one metro to the next is conditional on three hard targets in the prior metro: (1) 15+ active stores at 95% catalog stability, (2)measured 8%+ price advantage on a 50-SKU reference basket, (3) one migration branch running 60 days at positive contribution margin. Without those gates, we deepen the current metro instead of expanding.

• ▸Phase 1 (Year 1–2): Boston pilot. 10–15 mesh stores. Validate the engine.
• ▸Phase 2 (Year 2–4): Eastern Seaboard + Texas mesh. 220–520 stores. Activate streams 2 and 3.
• ▸Phase 3 (Year 4–7): Multi-corridor regional mesh. 1,100 → 3,400 stores. Activate streams 4 and 5.
• ▸Phase 4 (post-Y7): Conditional international expansion (MENA, South Asia, LATAM). Not part of the base case — unlocked only on demonstrated execution.

Five vertical organs at maturity: Brain (AI + data), Mesh (store onboarding + operations), Origin (supplier + importer + private label), Channel (consumer app + restaurant + B2B subs), Vault (security, compliance, regulatory). Each runs as a P&L; each contributes signal back to the central event spine.