For Founders · 2026

How to Build Your First Supply Chain

By Josh SCM · Updated 2026-10-27 · 12 min read

You have a design, a landing page, and maybe a Kickstarter countdown. What you don't have is a supply chain — a repeatable way to turn money into products that arrive on time, at the quality you promised, at a cost that leaves you margin. This is the roadmap we walk founders through: five phases, what each one actually costs and how long it takes, and the five decisions that are painful to reverse.

The part most founders get backwards

The instinct is to start with the factory: search a platform, message ten suppliers, compare quotes, pick the cheapest. That sequence fails for a reason that has nothing to do with supplier quality — before you have a specification, you don't have a quote, you have a guess.

Two factories quoting "the same" product without a shared spec can differ by 40% and both be honest, because they priced two different objects. And every quote you receive before the spec is frozen locks in assumptions you'll spend the next three months unwinding.

So the order is: spec → shortlist → sample → pilot → mass → repeat. Skipping ahead doesn't save time. It moves the work to the phase where changes are most expensive.

Phase 0 — Freeze the spec before you message anyone

Your spec pack (technical pack) is the single document your entire supply chain runs on. For apparel it means measurements, fabric composition and weight, trims, labelling, and construction notes. For hardware: drawings with tolerances, materials, finishes, component brands, and functional test criteria.

• Materials with grade and spec — not "cotton fabric" but composition, weight (GSM), and finish.
• Dimensions with tolerances — what range is acceptable, stated in numbers.
• Components named explicitly — zipper brand and type, screw standard, chipset revision. Named components cannot be swapped quietly later.
• Finish and colour standard — a Pantone reference or physical reference beats "matte black."
• Functional test criteria — what must be true for a unit to pass (wash cycles, drop test, load test, run time).
• Packaging spec — polybag, carton dimensions, inserts, labelling, barcodes. It affects cost twice: packaging and freight.
• Target cost and target MOQ — so the factory can engineer toward your constraints instead of guessing.

Spec freeze is a decision, not a state of mind. Write "Revision 1.0 — approved for quotation" on the file, date it, and send the same file to every supplier. From that moment, every change goes through a version bump and a re-quote. This one habit prevents more disputes than any contract clause.

Founder version: if you can't fill all seven fields, you're not ready to quote yet — and that's fine. Missing fields are exactly what a sourcing partner helps you close.

Phase 1 — Build a shortlist you can actually compare

Aim for three to five candidates per core process. Fewer than three gives you no leverage and no fallback; more creates a comparison you can't manage in parallel.

Screen on four things before you talk price:

• Capability fit — they make your category and your order size is meaningful to them. A factory running 200,000 units a month will not prioritise your 500.
• Licences and exportability — business licence, and whether they can export under their own name or need an agent.
• Quality system in practice — in-process inspection, final inspection records, calibrated equipment. Ask for evidence, not adjectives.
• Communication — response speed and technical depth in the first three emails predict the next twelve weeks unusually well.

Then send one identical enquiry to all of them: same spec file, same quantity, same Incoterm, same request for a line-by-line breakdown. Identical input is what makes the output comparable.

Use the verification point-by-point with shortlisted suppliers — factory audit checklist →
And structure the quote request so the numbers actually line up — cost breakdown request template →

Phase 2 — Sampling and the golden sample

Sampling is where the design meets reality, and it almost always takes two or three rounds. Budget for that instead of being surprised by it. Each round should be met with written, numbered feedback ("collar 3mm narrower; stitch density 10 per inch"), never "looks slightly off, please improve."

• Prototype (looks-like) — validates form, fit, and construction. Usually the fastest round.
• Pre-production sample (works-like) — uses production-intent materials and processes. This is the round that reveals real problems.
• Golden sample — the approved unit, signed by both sides, that defines mass production.

Two operational rules. Keep at least two physical golden samples — one at the factory, one with you. Photos compress defects. And pay for samples and express shipping rather than asking for free ones: paying buys you responsiveness, and it's the cheapest goodwill you'll ever purchase.

Phase 3 — Pilot run and your first QC gate

A pilot — typically 5–15% of your intended order — is the dress rehearsal. It proves the factory can make 200 of something, not just one, using production tooling and line settings. Skipping straight to mass production is the most common and most expensive shortcut first-time founders take.

From the pilot, three things get locked:

1. Process parameters — machine settings, cycle times, line layout, operator count. These become the standard for mass production.
2. Defect categories — separate critical (unsafe, non-functional), major (visible, affects sale), and minor (cosmetic, tolerable). Write the agreed acceptable quality level into the contract.
3. Yield and takt — real scrap rate and real units per day, which let you sanity-check both the quoted lead time and the quoted price.

Inspect the pilot before you pay the balance. Inspection after payment and before shipment is the only point where you still hold leverage.

Phase 4 — Mass production, freight, and cash

Mass production is mostly project management. Weekly or bi-weekly status from the line — photos, completion percentage, materials received date — costs nothing and catches problems while they're still fixable. Silence for two weeks in the middle of a run is a signal, not a coincidence.

Three things to settle before the run starts:

• Payment structure — the common pattern is a 30% deposit with the balance before shipment or against bill of lading. Whatever you agree, note that your cash is tied up for the full production plus transit cycle, and plan your runway around the second number.
• Duties and HS classification — your landed cost depends on the HS code and destination, not just the unit price. Classify before you ship, not when the broker calls.
• Freight mode and timing — ocean is slow and cheap, air is fast and expensive, and both move with seasonality. Chinese New Year and the Q3 peak window change price and capacity materially.

Landed cost is the only number that matches your bank statement — built line by line, never as a flat multiplier. Reference: the landed cost teardown →

Phase 5 — Repeat production and cost-down

The first order makes you a customer. The second order makes you a partner — and that's where unit cost actually falls. Once you're repeating, the levers open up in this order:

• Volume tiering — ask for the price curve at 1×, 3×, and 5× quantity and understand where setup costs stop dominating.
• BOM re-engineering — the largest lever, and it belongs here, after you have real performance data on what can be safely changed.
• Tooling ownership — if you paid for the mould or pattern, it's yours and it should be transferable. That single clause keeps your factory honest for the life of the product.
• Packaging and consolidation — improving carton cube or combining shipments reduces cost on two lines at once.
• Second source on standby — keep one additional qualified factory that has been vetted and sampled, even at zero order volume. Negotiating with a credible alternative is fundamentally different from negotiating with none.

Timeline and budget at a glance

Illustrative ranges for a first-time founder's project. Tooled categories (injection moulding, custom hardware) sit at the long end. Verify against your own project.

Phase	Typical duration	What it costs you	Exit criterion
0. Spec freeze	1–3 weeks	Your time, possibly design/engineering support	Dated spec pack sent to all suppliers
1. Shortlist & quotes	2–4 weeks	Minimal; possibly audit travel	3–5 comparable line-by-line quotes
2. Sampling	3–6 weeks	Sample fees + express shipping per round	Signed golden sample, two physical units held
3. Pilot run	2–4 weeks	Pilot units at or above unit price + inspection	Pilot passes agreed AQL; process parameters locked
4. Mass production	4–8 weeks	Deposit (commonly 30%), then balance before shipment	Pre-shipment inspection passed
Freight & clearance	Sea 3–6 weeks / Air 3–10 days	Freight, insurance, duty, brokerage, inland delivery	Goods landed at your warehouse
5. Repeat & cost-down	Ongoing	Engineering time, second-source sampling	Unit cost falls on documented changes, not pressure
Total to first landed goods	12–20 weeks	Tooling, testing, samples, and freight sit outside unit price — budget them separately

The most common budgeting failure: funding the unit price and discovering tooling, testing, and freight afterwards. Those lines are real and they arrive early in the cash cycle.

Five decisions that are expensive to reverse

Decision	Why it's hard to undo	Do this instead
Changing spec mid-run	Repriced when you have zero leverage; materials already bought	Freeze in writing; version every change; re-quote each revision
Not owning your tooling	You can't move factories without rebuilding it	Pay for it, get ownership and transferability in writing
Single-source with no qualified backup	Every future negotiation happens from a position of dependency	Vet and sample a second factory even at zero volume
Over-ordering against unvalidated demand	Inventory is locked cash; discounts later destroy margin	Pilot small, then buy into validated sell-through
Accepting DDP without seeing the breakdown	You lose visibility into every cost lever at once	Buy FOB with your own freight and duty lines modelled

Your first 90 days

1. Weeks 1–3: write and freeze the spec pack. Date it "Rev 1.0." Set your target cost, target MOQ, and target delivery date.
2. Weeks 3–6: build a shortlist of 3–5 factories. Send one identical enquiry with the request for a line-by-line cost breakdown.
3. Weeks 6–8: compare quotes on equal Incoterms and equal spec. Interview the top two — ask who else they make for in your category and volume.
4. Weeks 8–12: start sampling with your front-runner and keep a second supplier warm with a duplicate sample order.
5. Throughout: open a landed cost sheet with every line separated — unit price, tooling, samples, testing, freight, duty, brokerage, defect reserve.
6. Before any deposit: confirm tooling ownership, payment milestones, defect AQL, inspection timing, and lead time in writing.

The mistakes that cost the most

• Quoting before specifying. You collect confident numbers about an undefined product, then spend weeks reconciling them.
• Picking on unit price alone. The cheapest quote usually wins on a specification you didn't read closely enough.
• Skipping the pilot. The defects that only appear at 200 units appear after you've paid for 2,000.
• Inspecting after paying in full. Then defects become a negotiation instead of a remedy.
• Treating freight and duty as an afterthought. Landed cost, not unit price, decides whether you have a business.
• Never documenting anything. In a dispute, the written spec and the signed golden sample are the only things that carry weight.

If MOQ is currently what's blocking you, start there — how to negotiate MOQ →
And if you're at the very beginning, follow the full sequence — the complete sourcing playbook →

FAQ