Why You Can Cost a Part That Does Not Exist Yet
Founders assume manufacturing cost is something a factory tells them. It is not. A factory quote is a price: it includes the factory's margin, its current capacity situation, and its read of how sophisticated you are. A cost estimate is a build-up from physics, and you can construct one from a CAD model, a material choice, and a target volume.
You should, because the decisions it drives are irreversible. Once steel is cut you are committed to a wall thickness, a part count, and a process. A defensible cost before that point separates a product with a viable margin from one discovered to be unsellable after $40,000 in tooling. Aim for accuracy within 15 to 20 percent, which is enough to choose between processes or set a retail price. Chasing plus or minus 3 percent from CAD is wasted effort.
Step 1: Break the Product Into a Costed Parts List
Every estimate starts as a line-item list, not a lump sum. Enumerate every molded part, machined part, fastener, seal, cable, label, PCB assembly, and piece of packaging. A structured bill of materials is the correct container for this, and building it early forces you to notice the 14 screws and 3 gaskets that nobody costed.
Split the list in two. Purchased components have a price you can look up today at your quantity break; custom parts are the ones you have to estimate. Purchased parts are usually 40 to 60 percent of a mechatronic product's cost and take an afternoon to nail down, so do them first.
Step 2: Material Cost, the Easy Half
For a molded part, take the part volume from CAD in cubic inches or cubic centimeters, multiply by material density, and you have part weight. Add the runner and sprue, which for a small part on a cold runner can be 20 to 40 percent of shot weight, and less than 5 percent on a hot runner tool. Multiply by resin price per pound.
Commodity resins such as polypropylene and ABS run roughly $1.00 to $2.00 per lb in truckload quantities and $2 to $4 in bag quantities. Engineering grades like glass-filled nylon run $3 to $8, and high-performance materials such as PEEK are 20 to 40 times commodity pricing, which is why plastic selection is a cost decision as much as an engineering one.
For machined parts, cost the stock, not the part. A 2 in cube of aluminum costs the same whether the finished part weighs 8 oz or 2 oz, and the removed material is worth pennies as chips. That asymmetry is why near-net processes beat machining as volume rises.
Step 3: Machine Time, the Number That Decides
Material rarely dominates. Machine occupancy usually does.
Injection molding. Cycle time is driven by the thickest wall section, because cooling scales roughly with the square of wall thickness. A 0.080 in (2 mm) wall on a small part typically cycles in 20 to 35 seconds; 0.160 in (4 mm) can push past 60. Divide cycle time by the number of cavities to get seconds per part. A press with an operator bills around $45 to $90 per hour, so a 30-second cycle in a two-cavity tool at $60 per hour is about $0.25 of machine time per part. Uniform walls, generous draft, and no unnecessary side actions are worth more to unit cost than any material substitution, which is the argument running through the whole design for manufacturing discipline.
CNC machining. Estimate setups first, since each adds fixture time and work-holding cost, then estimate cut time from removal volume. A 3-axis mill bills roughly $60 to $110 per hour and a 5-axis machine $110 to $200; the detailed drivers are in CNC machining cost.
Sheet metal. Count laser or punch time per profile length, then count bends. Each bend is a separate handling operation, so a 12-bend bracket costs far more than a 4-bend one at the same weight.
Step 4: Labor, Finishing, and Packaging
Assembly is estimated in seconds per operation. A rough working set: 3 to 5 seconds to place and drive a screw, 2 seconds for a snap fit, 8 to 15 seconds to route and connect a cable, 20 to 45 seconds for a functional test at end of line. Multiply by a fully loaded labor rate, perhaps $4 to $8 per hour in Southeast Asia and $25 to $45 per hour in the US once burden is included.
Then add the finishing operations people forget: painting or powder coating, pad printing or laser marking, ultrasonic welding, heat-staking inserts, deburring, and cleaning. Each is a separate handling step with its own labor and yield. Packaging is real money too, often $0.60 to $3.00 for a retail box.
Step 5: Overhead, Scrap, and the Factory's Margin
Add a scrap allowance based on process maturity: 1 to 2 percent for a stable molded part, 3 to 8 percent for a painted or cosmetic part, more for a hand-finished surface. Then apply factory overhead and profit, typically 15 to 30 percent on top of loaded cost. Do not stop at the ex-works number either: freight, duty, and inbound handling add 8 to 25 percent depending on product density and tariff classification, all before you have sold anything.
Step 6: Amortize the Tooling
Tooling is capital, but it belongs in the unit economics. Divide total tool cost by the units you honestly expect to sell before a design change, not by the tool's theoretical life.
A $25,000 mold over 5,000 units adds $5.00 per part and probably kills the business case. The same mold over 100,000 units adds $0.25 and is invisible. That one calculation decides between a soft aluminum tool and a hardened steel one, a tradeoff laid out in aluminum vs steel injection molds, and whether molding is the right process at all at your volume, the question in choosing a process by volume.
Validating the Estimate
Three checks catch most errors. Compare against a physical analog: buy a competing product of similar size, material, and part count, and work backward from its retail price using the markup chain in product pricing. Send the CAD to two or three suppliers for a budgetary quote at your target volume and see whether your number sits in their range. And run a sensitivity pass: if a 10 percent resin price move changes your gross margin by more than a couple of points, material is your exposure and value engineering should start there.
If suppliers come back 40 percent above your build-up, you have missed an operation, not a rate. Usually it is a secondary process, a fixture, or a yield loss nobody mentioned.
Get a Costed Estimate Before You Commit Steel
Projects House builds pre-tooling cost models for clients deciding between processes and configurations: costed BOM, cycle-time estimates, tool amortization, and the two or three design changes with the largest effect on unit cost. Send your CAD and target annual volume through our contact form.