All notes
HARDWARE NOTE No. 03
What a supplier RFQ should ask that yours probably doesn't

Hardware Notes · No. 03

What a supplier RFQ should ask that yours probably doesn't

A clean RFQ gets you a clean quote — often on the wrong thing. The questions that actually price your project are the ones founders tend to leave out.

A clean RFQ gets you a clean quote — and a clean quote on the wrong scope is exactly how a budget that looked healthy in month one detonates in month six. The supplier answers precisely what you asked, prices the happy path, and stays silent on tooling ownership, the volume curve, change orders, and the variance behind the lead time — because you did not ask. Below are the questions that price the real deal, not the demo, plus a copy-pasteable checklist to send every supplier so you are comparing the same thing instead of five differently-shaped fictions.

1. Tooling ownership & control — in writing

If the factory owns the mold, your second source is a hostage negotiation. The single most expensive omission in a hardware RFQ is the tooling clause, because the mold is the one asset that converts a supplier from a vendor you can replace into a single point of failure you cannot. Get all of this on paper before the deposit clears:

  • Title. State plainly that you own the tool on final payment, with the tool number, cavity count, and a photo recorded against your name. “We made the tool” is not ownership.
  • The right to move it. Your contract should let you pull the mold to another factory on notice, with no “tool debt” or hostage fees. A tool you legally own but cannot physically retrieve is worth nothing.
  • Storage & maintenance. Who stores it, in what humidity, and who re-polishes or re-shims it as it wears? A neglected mold quietly degrades your part for a year before anyone admits it.
  • Shot-count warranty. A class-101 tool is built for roughly a million-plus shots; a soft “bridge” tool may give you 50,000–100,000. Make the supplier state the rated life and who pays to rebuild it when it is reached.
  • T1/T2 sample sign-off. Define that production cannot start until you approve first-article samples (T1, then T2 after corrections) against a dimensioned drawing. The sign-off is your acceptance gate — skip it and you have agreed to whatever falls out of the press.

None of this is adversarial. A good factory expects these terms; a factory that flinches at them is telling you something for free.

Supplier rfq review

2. The MOQ ladder, not the MOQ

Do not ask “what is the MOQ.” A single MOQ number tells you almost nothing — it hides where the cost actually lives. Ask for unit price at 500 / 1,000 / 5,000 / 10,000, ex-works, with tooling/NRE quoted as a separate line. The shape of that curve is the real information.

Here is a representative ladder for a small molded-and-assembled product, with $8,000 of tooling (NRE) amortized two different ways:

Order qtyPiece price (EXW)Tooling per unitAll-in per unitTotal cash outlay
500$6.80$16.00$22.80$11,400
1,000$5.40$8.00$13.40$13,400
5,000$4.10$1.60$5.70$28,500
10,000$3.70$0.80$4.50$45,000

Two lessons fall straight out of the arithmetic. First, amortization dominates at low volume: the same $8,000 tool adds $8.00 to every unit at 1,000 pieces ($8,000 ÷ 1,000) but only $0.80 at 10,000 pieces ($8,000 ÷ 10,000). Your all-in cost collapses from $13.40 to $4.50 — a 3× swing — and most of that move is the tool spreading out, not the factory getting more efficient. Second, the piece price itself drops from $6.80 to $3.70 as setups, purchasing, and labor get shared across a bigger run.

Now the “low-MOQ trap.” A supplier who proudly offers a 500-unit minimum is quoting you $6.80 a piece — about 66% more than the $4.10 you would pay at 5,000 ($6.80 ÷ $4.10 ≈ 1.66). A low MOQ is not a discount; it is the privilege of buying small at a premium. That can be exactly right for a first production run where you would rather eat $2.70 a unit than gamble $28,500 on demand you have not proven. But you can only make that call if you can see the ladder. Ask for it.

3. Change-order / ECO costs after tooling

Factories quote the happy path; the money is in the revisions. A product that never changes does not exist — you will move a boss, change a connector, swap a resin, or chase a failed drop test — and every change after the tool is cut has a price the original quote conveniently omitted. Ask, in writing, what each of these costs and how long it takes:

  • A mold revision. Welding shut and re-cutting a feature, or adding a slide, can run from a few hundred dollars for a minor steel-safe tweak to several thousand for a major change — and if the change removes steel you cannot add back, you may be buying a new insert or a new tool.
  • A BOM line change. A new component means re-qualifying a sub-supplier, updating work instructions, and often a minimum buy of the old part you are now scrapping.
  • Re-inspection & re-validation. A changed part may need a fresh first-article inspection, new test fixtures, or re-certification — line items that dwarf the part cost.

The phrase to insist on is “steel-safe”: design and review the tool so early changes add material rather than remove it, because adding steel (a smaller cavity, more plastic) is cheap and removing it (a bigger cavity, less plastic) is not. A supplier who can quote you a clear ECO schedule has done this before. One who waves it off will discover the price the day you need a change — and so will you.

4. The cost breakdown & sub-tier visibility

Ask for the quote broken into material, labor, overhead, and margin, and ask who their sub-suppliers are for the critical parts. A factory that will not break the number down is quoting you a price, not offering a partnership — and you cannot value-engineer a number you cannot see inside. The breakdown tells you where to push: if material is 70% of the cost, you negotiate resin grade and wall thickness, not labor; if labor dominates, you redesign the assembly to shave seconds off the line, not the BOM.

Sub-tier visibility matters just as much. The factory you signed with may not make your most important part — the connector, the battery cell, the display, the custom IC. If a Tier-2 supplier two steps down the chain goes dark, allocates capacity to a bigger customer, or quietly substitutes a cheaper part, it is your line that stops, and you will not see it coming if you never knew the part existed. Ask which components are single-sourced, who makes them, and whether there is a qualified alternate. The honest answer is sometimes “there is no second source for this cell” — which is precisely the risk you needed surfaced at RFQ, not at month six.

Supplier fai tooling review

5. Lead time — and its variance

“Four weeks” is half an answer. The average lead time does not blow up your launch — the variance does. Ask “four weeks ± what, and what drives the spread?” The drivers are predictable, and a supplier who names them is one who actually understands their own operation:

  • Raw-material lead time. The factory may build in four weeks, but if the resin, the PCB laminate, or a long-lead IC is on a 12–16 week clock, that is your real lead time. Ask what is on order and what is on the shelf.
  • Shared lines. Your run shares a press, an SMT line, or a paint booth with other customers. When a bigger order jumps the queue, your “four weeks” becomes seven, and nobody calls you.
  • Chinese New Year. For roughly the last two weeks of January or first three of February, plus a ragged ramp-down before and ramp-back-up after, much of China and Taiwan’s supply base stops. Anything you need in Q1 must be planned in Q4 — this is the single most predictable schedule killer in Asian manufacturing, and it surprises a new founder every single year.
  • The bullwhip. Small swings in your forecast amplify as they travel up the supply chain: a 10% bump in your order can become a 40% scramble at the cell or connector maker. Smooth, honest forecasts buy you reliability; lumpy ones buy you stockouts.

Pin down lead time as a range with named causes, and you can build a launch schedule with real buffer instead of a single optimistic number you will miss.

6. Quality & acceptance terms

“Good quality” is not a spec; it is a feeling. Acceptance has to be defined numerically before the first carton ships, or every defect becomes an argument you are positioned to lose. Nail down:

  • AQL levels. Agree the Acceptance Quality Limit per defect class — a common starting point is AQL 0.65 for critical defects, 2.5 for major, 4.0 for minor, sampled per ISO 2859. This defines how many defects in a sampled lot trigger rejection of the whole lot.
  • FAI / first-article inspection. A full dimensional and functional check of initial parts against the drawing, signed off before mass production. No FAI, no green light.
  • IQC. Incoming-quality control on their raw materials and sub-components — you want the factory catching a bad cell or off-spec resin at their dock, not in your returns.
  • Scrap & rework liability. State plainly who pays when a lot fails: the factory eats the scrap and rework on their defects, and a failed AQL inspection means they re-screen or re-make at their cost, not yours. Without this clause, “rework” lands on your invoice.

7. Payment terms & IP/NDA realities

The standard structure is a deposit against tooling and a balance against shipment — commonly 30% down, 70% before the goods leave (or against a copy of the bill of lading). Resist paying 100% up front: your remaining balance is the only leverage you hold once the order is in motion. For larger first orders with a new supplier, an escrow or a letter of credit puts a neutral party between the cash and the cartons.

On IP, set expectations honestly. An NDA is worth signing, but its practical enforceability across borders is limited and slow — treat it as a filter for serious counterparties, not as armor. Your real leverage is structural: own the tooling, hold a balance payment, split a sensitive sub-assembly across two suppliers so no single factory holds the whole design, and keep firmware and the highest-value know-how out of the BOM you hand over. The tool you own is both your supply-chain insurance and your strongest IP lever — whoever controls the mold controls the negotiation.

8. Logistics & Incoterms

The quote’s Incoterm decides who owns the goods, the cost, and the customs paperwork at each leg — and the same “$4.50 a unit” can land at wildly different real costs depending on the three letters next to it:

  • EXW (Ex-Works). Cheapest sticker, most work: you take ownership at the factory door and arrange — and pay for — everything after, including export clearance. Easy to under-budget.
  • FOB (Free On Board). The supplier delivers the goods loaded at the named port and clears export; you own ocean freight, insurance, and import. This is the usual sane default for comparing Asian quotes.
  • DDP (Delivered Duty Paid). The supplier delivers to your door, duties and import clearance included. Simplest for you, highest unit price, and you are trusting them to handle your country’s customs — verify they actually can.

The trap is comparing an EXW quote against a DDP quote as if they were the same number. They are not: freight, duty, insurance, and brokerage can add a meaningful fraction to the landed cost. Normalize every supplier to the same Incoterm before you rank them, and know exactly which side clears customs on each end.

Red flags & common mistakes

  • No written tooling clause. If ownership, storage, shot-count, and the right to move the mold are not on paper, assume you do not own it — and that your second source is a negotiation from weakness.
  • Accepting “the MOQ” without the ladder. A single minimum hides the volume curve and the low-MOQ premium. No ladder, no real comparison.
  • Ignoring ECO costs. Quoting only the happy path guarantees a surprise the first time you need a change — which is always.
  • A quote with no breakdown. A lump-sum price you cannot see inside cannot be value-engineered or trusted. Walk it back to material / labor / overhead / margin.
  • No variance on lead time. An average with no ± and no named drivers is an optimistic guess you will be held to and will miss.
  • Paying 100% up front. You have just thrown away your only leverage. Keep a balance payment tied to acceptance and shipment.

The RFQ checklist to send every supplier

Paste this into your RFQ verbatim. The goal is not to interrogate — it is to make every supplier answer the same questions so you are comparing real deals, not sales pitches.

#Question to send
1Do we own the tool on final payment? State tool number, cavity count, rated shot-count, storage location, and our right to move it.
2Quote unit price (EXW) at 500 / 1,000 / 5,000 / 10,000, with tooling/NRE as a separate line.
3What does an engineering change cost after tooling — mold revision, BOM-line change, re-inspection — and is the tool designed steel-safe?
4Break the unit price into material, labor, overhead, and margin.
5Which components are single-sourced, and who are the Tier-2 suppliers for the critical parts?
6What is the lead time ± range, and what drives the variance (raw-material lead time, shared lines, holidays)?
7What AQL levels apply per defect class, and is FAI and IQC included? Who pays for scrap and rework on your defects?
8What are the payment terms (deposit / balance), and do you accept escrow or L/C for the first order?
9Which Incoterm is this quote (EXW / FOB / DDP), and who clears customs on each end?
10Will you sign our NDA, and which sub-assemblies can be split or kept in-house to protect the design?

The supplier who answers tooling, the MOQ ladder, change orders, the cost breakdown, lead-time variance, quality terms, payment, and Incoterms straight at RFQ is the one who will still be straight with you at month six — everyone else is quoting you a price, not a partnership.