Tooling costs, and the fortune the first part costs
Why unit price is meaningless without the volume attached to it.
The first part off a new tool effectively costs a fortune because nearly all of the money spent making that tool exists before a single part has actually been produced, and that entire upfront cost is carried, in an accounting sense, by whichever part happens to be first, even though every part made afterwards shares the exact same tool for a small fraction of the original outlay.
A fixed bill that does not care about volume
Building a mould or a die involves a large amount of design, machining and proving work that has to happen in full before the tool can produce anything at all. That cost is set by the complexity of the shape, so a tool built to make ten parts and a tool built to make ten million parts of the same design cost roughly the same to build, at least until the higher-volume tool needs harder steel to survive its longer life.
Once the tool exists, running it for one more cycle costs comparatively little, mostly material and machine time. A part's true average cost therefore falls steeply as volume rises, from something close to the entire tooling bill on the very first part down toward the bare material and running cost once enough parts have shared that bill between them. Quoting a single unit price without stating the volume behind it tells the listener almost nothing, since the same tool, part and process can honestly be described as costing a fortune or almost nothing depending on how many units the fixed cost is divided across.
The first key and the eleven spares
Having a locksmith cut an unusual key from a blank shows the same accounting in a setting most people have paid for directly. Most of the price of that first key is the locksmith's time, setting up the cutting machine, matching the pattern precisely and checking the result turns in the lock, and that work has to happen in full whether one key or a dozen are being cut.
If a dozen identical spares are cut in the same sitting, each additional one costs only a small part of what the first did, because the expensive step, getting the machine set up for that particular key, has already been paid for by the first. Every key after it rides on that one setup cost instead of paying for a fresh one of its own, which is exactly how the millionth moulded part rides on a tool paid for long before.
Dividing one tool by two volumes
Round numbers make the gap plain. A tool costing £50,000 works out at £500 of tooling on each part if only a hundred parts are ever made, and at ten pence on each part if half a million are made, with the part, the material and the machine identical in both cases. The whole difference between a fortune and pocket change is the same fixed cost divided by two very different volumes.
That gap is why a project's first cost estimate is so often wrong in one direction or the other. A supplier quotes the tooling-heavy cost of a low volume to a customer mentally expecting the amortised cost of a high one, or the reverse, and both sides leave the conversation thinking the other has misunderstood. The single most important figure to agree before any tooling quote is discussed is the volume the quote is being priced against.
It follows that a quote which looks alarmingly expensive at low volume may be entirely fair, and one that looks suspiciously cheap at high volume may be no bargain, because both only make sense once the volume behind them is known. Two quotes for the same part priced against different assumed volumes are answering two different questions, and setting them side by side without checking tells a buyer nothing about which supplier is cheaper.
Cheaper tools for shorter runs
A low-volume project should think about tooling from the very start. Committing to an expensive, precisely engineered permanent tool for a product that will only ever need a few hundred units guarantees an unfavourable cost per part however efficient that tool is. The same product made on a cheaper, less durable tool, accepted precisely because it will never face the millions of cycles a mass-production tool must survive, can land at a far more sensible cost for the volume actually being made. Aluminium moulds, softer steels and simpler tools with fewer moving parts all exist to fill that gap.
Guessing the volume before anyone knows it
The arithmetic gets harder when future volume is genuinely uncertain. A low-cost tool chosen to suit a small first run becomes the wrong decision in hindsight if demand later grows far beyond what it was built to survive, forcing a second, more expensive tool to be commissioned partway through the product's life. Choosing a tool for an uncertain volume is a forecasting problem more than a mechanical calculation, and getting it wrong costs money either way, whether by paying for durability nobody needed or by buying a tool that wears out just as real demand arrives.