Companies do not buy machined parts; they buy working programs that deliver usable parts at predictable cost. The unit price is the number everyone quotes and almost no one should compare first, because the costs that actually shape a machining program live around it - in the tooling, the setup minimums, the freight terms, the quality escapes, and the engineering time spent managing the supplier. This guide builds the cost model a company should run before committing a part family to a source.
The model is simple to state and demanding to fill in: total program cost equals the delivered cost of good parts, plus the cost of the failures and delays that a supplier's performance creates, plus the overhead your team spends on the relationship. Buyers who run this model discover that the cheapest quote is rarely the cheapest program, and that the suppliers worth building around are the ones whose total cost is predictable, not just low.
1. Why is the unit price the wrong number to compare first?
The unit price on a quote is built on assumptions that differ between suppliers: the material grade and section, the number of setups, the finishing process, the inspection depth, and the freight term. Two quotes can differ by twenty percent on the same drawing while both are accurate, because one assumed a stocked bar and a single setup while the other priced a bought-in section and a second operation.
The unit price also hides the costs that arrive after the quote. A supplier whose parts fail incoming inspection transfers rework and expediting into your cost whether or not the invoice changes. A supplier whose dates slip forces emergency freight and line delays that never appear on its quote but appear in your profit. The quote is the visible surface of a cost structure that extends far beyond it.
That is why the first discipline is normalization. Put the identical drawing, material, finish, quantity, and trade terms in front of every supplier, and ask for itemized quotes. Only when the quotes describe the same scope does the unit price become a number worth comparing, and even then it is one line in the program model rather than the whole story.
Companies should also price the decision lag. Locking a part family to one supplier for a year on the strength of a single quote leaves the model stale as material markets and freight rates move. Building an annual re-quote into the process keeps the cost lines honest, and the review pays for itself on the first part whose price has drifted from the market.
2. What does the program cost model look like?
The program model has three layers. The first is the delivered cost of good parts: the itemized quote times the quantity, plus the tooling and setup that the first order carries, plus freight and duty under the agreed terms. This layer is what most buyers compute, and it is the smallest layer of the real total.
The second layer is the cost of performance: the inspection your team runs on arrival, the rework and scrap when parts miss the drawing, the expedited shipments when dates slip, and the line or launch delays when a supplier fails. This layer is invisible on the invoice but real in the accounts, and it is where a low-priced supplier quietly becomes an expensive one.
The third layer is the overhead of the relationship: the engineering hours spent reviewing quotes, chasing status, resolving non-conformances, and qualifying alternatives. None of this appears in the unit price either, yet it is a genuine cost that scales with the number of suppliers and their reliability. A program model that includes all three layers gives a company a number it can actually decide against.
3. Where do tooling and setup land in the first order?
Every machining program starts with one-time engineering: the program development, the fixture, and the prove-out that turn a drawing into a repeatable process under the quality system. On the first order these costs sit on top of the unit price, which is why the first article of a new part is dramatically more expensive than the tenth release of the same part. The fixed costs are the reason quantity curves exist.
The commercial question is who pays for the fixed work and who owns it afterward. Some suppliers fold the tooling into the first order; others quote it separately and credit it back after a volume threshold. Some hold the program and fixture for your part so the repeat order is cheap; others rebuild the engineering each time and charge for it again. The terms around fixed costs shape the economics of every subsequent order.
Ask the question directly: what does the first order carry that the repeats will not, and what is the repeat price on the same drawing? A supplier that can name the fixed costs and the repeat price is showing you a program view. One that quotes a flat unit price with no repeat logic is either averaging the fixed costs into every part or planning to re-engineer on each order.
4. How does quality appear as a number in the cost model?
Quality arrives in the cost model as the defect rate times its consequences. A part rejected at incoming inspection costs the unit, the return, the expedited replacement, and the line time you spent finding it. A part that escapes inspection and fails in service costs more again. The per-part price of a supplier's quality is its defect rate multiplied by those consequences, and that product is real money.
The mitigation is the inspection you run yourself. Companies that inspect incoming machined parts pay for that inspection, and the cost belongs in the model - either as a line, if the supplier is unproven, or as a saving, if the supplier's history lets you sample instead of inspect. A proven supplier removes an inspection cost that a new supplier requires.
This is why the reliable supplier with the higher unit price often wins the model. Its parts arrive correct, so the inspection line shrinks, the rework line disappears, and the engineering overhead drops. When those lines are added to its unit price, the total is frequently below the cheap supplier whose failures generate costs that never appear on its quote.
5. How should freight and trade terms enter the comparison?
Freight can decide whether sourcing a part from China is economical at all, and the mode is the lever. Air freight moves a first article or an emergency in days but adds heavily to the unit cost. Sea freight costs a fraction per part and arrives in weeks, which is why repeat volume travels by sea with inventory covering the transit window. The mix of the two modes is a line in the cost model.
The trade terms decide where the freight cost and the risk sit. Ex-works and FOB leave the freight, the duty, and the transit risk with the buyer; delivered terms put them with the supplier and fold them into the price. Comparing quotes across different terms is meaningless until the freight lines are normalized to the same boundary.
The professional habit is to compute the delivered cost per part under each freight mode and each term, and to let the inventory cost of the slower mode enter the model. A part that moves by sea needs weeks of stock ahead of it, and that working capital is a cost. The mode that looks cheapest on the freight line is not always the mode that wins the model.
6. What is the real cost of tolerances and finishes on the drawing?
The drawing sets the process, and the process sets the price. A tolerance that requires measurement on every critical feature adds inspection time and documentation to every order. A finish that needs a coating batch adds a process and its minimums. Each specification is a recurring cost, and on a part ordered repeatedly the recurring costs dwarf the one-time engineering.
The highest-value review is the critical-callout audit. Walk the drawing feature by feature and ask whether each tight tolerance or fine finish serves a function - a seal, a bearing seat, a wear surface. Where the answer is no, relax the callout and remove its cost from every future order of that part. Where the answer is yes, keep it and price it knowingly.
The finish deserves the same audit. A machined finish or a light bead blast is cheap; a named anodize, a plating, or a fine Ra carries a process and a schedule. Confirm the finish matches what the application needs, because on many drawings the finish line is where the quote and the reality diverge.
7. How do you value reliability when building the supplier base?
Reliability has a price you can estimate: the cost of the failure it prevents. If a supplier's delay stops a line for a day, the value of on-time delivery is that day's cost, and a supplier whose dates hold is worth a premium up to that value. Companies that price reliability this way make sourcing decisions on economics; companies that ignore it make them on the invoice and pay the difference in crises.
The split between suppliers is a risk decision best made with the model in hand. Consolidating volume earns better pricing and scheduling, while a qualified second source protects the parts whose failure would hurt most. The size of the protection should reflect the cost of the failure, not a general preference for redundancy.
The scorecard that feeds the model should measure the three lines that create cost: on-time performance against confirmed dates, defect rate at your incoming inspection, and the responsiveness of the commercial process. Track them per supplier per quarter, and let the trend lines move the volume split. Suppliers who see their performance priced into their volume respond faster than any negotiation.
8. How do you build a machining cost review that pays for itself?
Run the review on the parts that spend the most, and run it with the model rather than the invoice. For the top part families, compute the delivered cost of good parts, add the performance layer from the last twelve months, and add the overhead of managing the source. The parts that surface as overpriced against their peers are the ones worth re-quoting or re-designing.
Re-quoting is only half the value; the design review is the other half. A part whose critical callouts and finishes were set years ago carries costs from decisions nobody remembers. Re-auditing the drawing against the current function removes those costs permanently, and a company that does this on its high-spend parts sees the machining budget move without a single negotiation.
Finally, make the review a rhythm rather than a reaction. Review the supplier scorecards quarterly, re-quote the high-spend families annually, and re-audit the drawings on the parts that never changed. Machining cost is not a number a supplier gives you once; it is a total your company assembles from many lines, and the company that assembles it honestly is the company that controls it.
Frequently Asked Questions
Should a company negotiate unit price or total program cost?
Program cost. The unit price is one line inside a larger total that includes tooling, setup charges, freight, expediting, quality escapes, and the engineering time your team spends managing the supplier. A slightly higher unit price from a reliable supplier can beat a lower one from a source that fails, and the program total is what actually hits the business.
What is total landed cost for a machined part?
It is the delivered cost of usable parts at your door: the unit price times the quantity, plus tooling and setup, plus freight and duty, plus the cost of inspection and any rework or scrap, plus the overhead of managing the supply. Buyers who compare only unit prices ignore the lines that separate a cheap supplier from a cheap program.
How does supplier quality show up in the cost per part?
Quality failures arrive as cost: parts rejected at incoming inspection, expedited replacements, line stoppages while you wait, and engineering time spent on the return process. A supplier with a low defect rate removes those costs even when its unit price is a few percent higher. Quality is not a soft value; it is a line in the cost model.
Why do repeat orders of the same part often cost less?
Because the one-time costs do not repeat. The program, the fixture, and the proven process were paid for on the first order, so the repeat order carries only material, machining, and inspection. A supplier that does not lower the repeat price is charging you for engineering it no longer performs.
How much does freight add to a machined part from China?
Enough to decide the order of magnitude of your savings. Air freight can add a significant fraction to the unit cost on small orders, while sea freight spreads over weeks of inventory and costs a fraction per part. The mode should follow the purpose - air for first articles and emergencies, sea for repeat volume - and the terms decide who pays for the machining services.
What is the cost of a tolerance you do not need?
A precision callout on a feature the assembly never checks buys inspection time, documentation, and process risk on every order of that part. Relaxing it to what the function requires removes those costs permanently, which is why reviewing the critical-callout list is one of the highest-value cost exercises a design team can run.
How should a company split its volume between suppliers?
Consolidate the volume that a supplier serves well to earn better pricing and scheduling, and keep a qualified second source for the parts whose interruption would hurt most. The split is a risk decision priced in real terms: single-sourcing saves money until it fails, and the cost of the failure sets the value of the backup.
What are the 2026 price drivers a company should track?
The same levers that always mattered - material markets, energy, freight rates, and factory capacity - plus the cost of compliance and certification on regulated programs. The useful habit is to track the material index for your main grades and review freight quotes periodically, because those two lines move more than the machining rates.
Want the program view, not just a unit price?
Send your part list and annual volumes to [email protected] and ask for the itemized program quote - tooling, repeat price, quality history, and freight options - to feed your cost model.