Notes for owners · Channel and customer growth
Prototype quantity to volume: how to treat a start-up hardware customer
How a board house or assembler keeps a hardware start-up from its first ten boards to its volume order: a note after every build, a log of the start-up's milestones, and terms that protect you meanwhile.
The GullySales team · Updated 17 Oct 2026 · 6 min read
On this page
Ten boards from a hardware start-up look like a nuisance job. Two parts get fitted by hand, the BOM changes twice, and the founder messages at midnight. Treat it as the first entry in an account instead. After every build, send a one-page note on what you changed, what went wrong and which parts will be hard to buy at volume. Keep a log of the start-up's next milestone and check in when it falls due. The volume RFQ often comes from an operations head the founder hired later, and your notes are how that person learns you already know the board.
A start-up buys in stages, and the buyer changes
A start-up does not go from ten boards to ten thousand in one step. First come a handful of prototypes. Then come one or two engineering builds as the design is corrected, sometimes called EVT and DVT. A pilot of a few hundred boards follows for a field trial or the first customers. Volume comes last, if it comes. Between the stages sit events you do not control: a funding round, a certification test, a large customer saying yes.
The person changes too. The founder or a design engineer buys the prototypes. By the time volume arrives, there may be an operations or supply chain head who was not there for the first build. That person sends the volume RFQ to a list of EMS firms. The list is made from what they can find on file, and from whatever the founder happens to remember.
How to quote the prototype itself, without fixing a volume price too early, is covered in our post on prototype and production enquiries. This post is about what happens between the builds.
What goes in the note after every build
Send it with the boards, as a PDF the founder can forward. Keep it to one page.
- Board name, revision, BOM revision, quantity and build date.
- Anything you changed or fitted by hand, and why: a wrong footprint, a missing polarity mark, an alternate part they approved.
- Failures found on your test, by board serial number, with the cause if you know it.
- Changes worth making before volume, such as a part that cannot be placed by machine or test points the jig will need.
- Parts that will be hard to buy at volume: long lead times, a single source, end-of-life notices.
- Their next step, as they told it to you, with the date.
The fifth item matters most. A part that was easy to buy in tens may be on allocation in thousands, and the start-up finds out only when redesign is expensive.
The start-up log
Keep one row per start-up, separate from your running customers. It is a pipeline of accounts that do not exist yet. The entries here are illustrative.
| Start-up | Board and revision | Last build | What you flagged | Next milestone and date | Others involved | Your next contact |
|---|---|---|---|---|---|---|
| Milk temperature logger | Logger main board, rev B | 10 boards, consigned parts | Footprint error on U4; sensor connector single-source | Field trial at chilling centres, ends in about eight weeks | Founder; a contract firmware engineer | Ask for trial results when it ends |
| Smart lock | Lock controller, rev A | 5 boards, turnkey | Battery holder hand-fitted | Investor demo; date not given | Founder only | Ask for the demo date, then wait |
The last column is not a reminder to chase. It is the next moment when you have a reason to call that the founder will welcome.
Protect yourself while you wait
Start-ups run out of money, change direction and sometimes vanish. Keep the terms simple and written. Take an advance that covers parts bought only for their board. Offer one free rework of your own errors and charge for re-spins of theirs. Do not stock custom parts for their volume build on the strength of a forecast; buy against a purchase order or a written commitment to pay for the parts.
A discount is also awkward to take back. If you quote ten boards cheaply to win the volume order, the founder and any operations head who joins later will read that figure as your price, and the volume quote has to explain why it is higher.
Imagine: a Bommasandra assembler and a temperature logger
Imagine an SMT assembler in Bommasandra, Bengaluru. A start-up making temperature loggers for milk chilling centres asks for ten boards with consigned parts. The details are illustrative.
The build went smoothly except for one thing. The footprint for a sensor IC was wrong, and the assembler's technician rotated two parts by hand to make the boards work. Nobody wrote it down. The boards went out with an invoice and nothing else.
The start-up raised money and hired an operations head. She sent a pilot RFQ for 300 boards, rev C, to three EMS firms she had found online. The Bommasandra house was not among them, because nothing on her file mentioned it. Rev C still carried the footprint error.
The assembler's owner heard about the pilot from the founder, almost by chance, and wrote a one-page build note for the first batch after the fact. It set out the footprint error, the hand fix and two parts likely to be hard to buy at 300. The founder was asked to forward it to the operations head. She corrected rev C before the pilot and added the house to her list. Writing the note took the owner most of a Saturday, unpaid. The pilot quote still went in against two other houses, and the start-up chose the Bommasandra house only after the owner matched the lowest price by dropping a rework allowance meant to stay. It won the build, narrowly, at a thinner margin than the owner wanted. The note now goes with every prototype batch and keeps the log above.
When to stop investing in a start-up
Keep the row but stop spending time when the start-up misses a payment on a prototype invoice, when its milestone dates slip with no reason given, or when it asks for free re-spins again and again. If your line cannot build at the volume they talk about, say so early and offer a clean handover to a larger EMS house, with your build notes. The next product may be one you can build.
Write the notes you owe
List every start-up you built prototypes for in the past year or two. Write the build note for the last batch of each, even though it is late, and send it with one question about the founder's next milestone. Keeping those accounts in view is pipeline management, and the wider picture for the trade is on the electronics and embedded systems page. If prototype customers keep vanishing before volume, tell us what you need.
