MOQ vs. Annual Demand: Which Classic Parts Are Actually Worth Commissioning

A unit price is a comparison. Years of cover is a consequence. The money leaves on day one — deposit, tooling, first article, balance, freight, duty — and comes back one part at a time over however long the stock lasts. Everything else is downstream of how long that is.

The grid below is arithmetic, not data: each cell is the minimum divided by annual sell-through. We publish it because the arithmetic is what people skip, not because we have measured anything. There is no public, measured sell-through or inventory-turn figure for this segment — we looked, found only distributor-software marketing, and will not invent one. Put your own back-order numbers in the left column.

The left column is the whole decision, and it is the number people are most willing to lie to themselves about. Evidence divides cleanly into things somebody has already paid for and things somebody has expressed a feeling about. The diagnostic question, asked out loud: how many of these people have already given me money for something? A want and a purchase order are separated by price, so ask the price question in the same breath as the interest question — "would you take three at this number" collapses an enthusiastic forum thread to its real size fast, and it is much cheaper to be disappointed now.

Buyers routinely conclude that "Taiwan has high minimums" when what they actually met was a hard-tooling process. The same 250-piece request is unremarkable in one process and close to absurd in another, and knowing which is which changes both the shortlist and the part you choose to start with.

Because value preserved is not the same as cash available. A brace for a 1966 car has no model-year cliff, no supersession event and no moment where the platform disappears and strands your stock — genuinely different from the modern aftermarket, and the reason many small reproduction brands survive purchase orders that would kill a collision-parts seller. But storage, insurance, handling and damage keep running, and the real cost never appears on an invoice: every dollar on the shelf is a dollar not tooling the next part, so your growth rate is set by how much capital you can keep off that shelf.

It also matters who you compete with for shelf space in the customer's mind. The segment is consolidating into well-capitalised multi-brand platforms that can carry slow inventory in a way you cannot: on 28 July 2026 Holley Performance Brands announced the sale of the Scott Drake and Brothers Trucks brands to CJ Pony Parts, whose leadership said the intent was to expand inventory depth and support wholesale partners. For scale context only — this is the whole US specialty-equipment market, not the classic segment, which is not separately published — SEMA's 2026 Market Report puts US specialty-equipment sales at $52.92 billion for 2025. That is not your addressable market; it is a reminder that the buyers you pitch have alternatives with balance sheets.

So sequence. Tool one part, let it sell through far enough to test your demand estimate, then tool the next. The characteristic failure is not one bad part — it is five tooled in one enthusiastic year, all five moving at the same modest rate, and nothing left to fund the sixth, which was the good one.

A classic-parts business almost never has one part worth a purchase order. It has fifteen small ones spread across several small Taiwanese makers who each have their own minimum, and the economics only work if you treat them as one buy rather than fifteen decisions. Plan that buy around your cash cycle, not any single supplier's minimum: pick one or two buying windows a year, size each to what you can fund and store, and make the candidate parts compete for space inside it. A part that cannot earn its place this year is not rejected, it is queued. This is a purchasing problem, not a freight problem — the container-versus-LCL break-even has its own article, and a freight optimisation should never drag a part into the buy that did not earn its way in. What breaks in practice is schedule, every time.

Below roughly a few dozen units a year of demand you can actually evidence, no offshore tooling program pencils out, and we will say so before taking a sourcing fee. The alternative is not defeat. Buy at dealer terms from an established US reproduction house: Classic Industries' B2B program, for one, publishes that initial investment is not necessary for legitimate automotive businesses to apply, that qualifying businesses receive discounts regardless of the dollar value of the order, and that there are no minimum monthly order volumes to maintain. Wholesale margin on somebody else's inventory beats gross margin on nine years of your own. The other route is domestic small-batch manufacture — higher per piece, and you never own the stock.

One Taiwan-specific limit deserves naming plainly, and it comes from our own project experience rather than any published figure: small Taiwanese makers will generally not warehouse finished goods and release them against monthly calls the way a US distributor will. Blanket orders with scheduled releases exist, but they are the exception and are usually earned over several completed programs. Arriving with a cash plan that assumes one is the commonest way a first-time buyer's numbers fall apart in month two.

And if your demand is real but lumpy — a marque club buys forty in a month, then nothing for two years — that is a group-buy, not a stocking program, and group-buys fail in predictable ways: deposits collected before tooling is proven, participants who vanish between deposit and balance, a first article that fails while forty people expect delivery. Run one only if you can fund the whole order without the deposits.

If the part does clear the bar, describe it on our request page in plain English — vehicle, years, body style, position, material, finish, and the quantity you actually intend to buy. Taiwan Supplier Hub returns a shortlist of at least three verified manufacturers that have confirmed they want the project within 14 days, US$99, refunded if we cannot deliver. If the honest answer is that the part should not be commissioned, that is the answer you get.