Chapter 12

Roadmap & capital

Five phases, each with its own capital requirement and its own exit test. No phase begins until the previous one has passed.

Sequence

Capital required by phase
USD. Cumulative line shows total deployed.

Phase 0 in detail — the one that matters most

Almost every failed import business skipped this phase. It exists to answer two questions before any capital is committed to inventory: what does it cost to acquire a sample order, and how many of those become real orders? Everything else in the model is arithmetic once you know those two numbers.

What you build

  • Arabic-first storefront, 25–35 curated designs
  • Product sourced per order from existing local distributors at trade price
  • Sample pack as a live, purchasable product
  • Postcode delivery calculator with real freight quotes
  • Paid social and search in one city, not one country

What you do not build

  • No container, no importing, no inventory
  • No warehouse, no 3PL contract
  • No conformity certification yet
  • No trade tier, no account managers
  • No second market
Phase 0 exit test

Proceed only if cost per sample order is under USD 22 and sample-to-order conversion exceeds 20% across at least 150 sample orders. Miss either and the correct decision is to iterate the range and the creative, not to import. Missing both twice is a no.

Capital breakdown, Phase 1

Where the launch capital goes
One market, first containers, full compliance.

Note how small the inventory line is relative to a typical retail launch. A 40HQ of porcelain lands for roughly . Six to eight containers cover the opening range. The risk is not the size of the cheque — it is buying the wrong designs, which is exactly what Phase 0 is for.

Market sequence and why

The sequence follows the weighted entry score, with one deliberate override: Bahrain is served from a hub rather than entered as a market. Every structural quality in Bahrain is favourable — no domestic manufacturing, lightest regulation, good payments — but at 5.5% of the regional market it cannot repay a standalone warehouse and entity. Served by truck from Dammam across the causeway, it is incremental revenue at near-zero marginal fixed cost.

Milestones and decision gates

GateWhenPass conditionIf it fails
Demand provenEnd of Phase 0 Sample CAC < $22; sample→order conversion > 20% Iterate range and creative; re-test once. Twice failed is a no.
Unit economics provenMonth 9 Positive contribution per order; inventory turning within 60 days Hold at one market, cut fixed cost, fix the range before expanding
Channel provenMonth 18 Trade > 30% of revenue; repeat rate > 20% Do not open market three. Consumer-only economics rarely carry this model
Scale provenMonth 30 Group EBITDA positive across four markets Consolidate to the two profitable markets; exit the others cleanly

The first ninety days, concretely

  1. Weeks 1–2. Choose the lead market. Register the entity and the domain. Identify 40 candidate designs from Foshan and Castellón catalogues. Verify every Chinese factory against the GCC cooperating register before anything else.
  2. Weeks 3–4. Order sample sets from 8–10 factories. Build the Arabic storefront. Source the initial range from local distributors for the per-order model.
  3. Weeks 5–6. Photograph properly — the single highest-leverage spend in this business. Room scenes and detail shots, not factory catalogue images. Launch the sample pack.
  4. Weeks 7–10. Run paid acquisition in one city. Measure sample cost and conversion relentlessly. Talk to every sample buyer personally.
  5. Weeks 11–12. Review against the Phase 0 gate. If passed, place the first purchase orders on the 8–12 designs that actually sold, from both routes, and start the conformity certification in parallel.
The one instruction that matters most

Do not buy a container before Phase 0 passes. Every failure mode in this business traces back to inventory bought on conviction rather than evidence. The sequence above costs USD 22,000 and two months to eliminate the largest risk in a USD 300,000 plan. There is no version of this where skipping it is the faster route.