Pricing 3D prints: filament, machine hours, and failed plates

There is a number your slicer gives you that feels like a cost, and it is not one. It says grams and it says hours, and most sellers price off the grams. That is how a print that occupied a machine for the better part of a day gets sold for the price of its plastic.

Plastic is the cheapest thing in a 3D-printing business. The expensive things are the hours the plate was busy, the prints that failed on their way to becoming something, and the finishing work that happens after the machine is done and cannot be sped up by buying another printer. This chapter costs those honestly.

The method underneath is the same one in pricing handmade products: materials, plus labor at a real rate, plus a share of overhead, into a true cost per unit. What follows is what those three mean when a machine does the making. For the method on its own, argued on a single hand-made object with no printer in the picture, read pricing handmade products without underpricing yourself first.

Material is per gram, but you never buy it per gram

Your material cost per print starts from a cost per gram, and the honest cost per gram is not the spool price divided by its labeled weight. It is the spool price divided by the grams you actually get product out of.

The gap between those two is real and it is all waste that never becomes anything sellable:

• Purge and prime lines at the start of every plate, and every filament change on a multi-material print.
• Brims, rafts, and skirts.
• Support material, which the slicer counts in its gram figure and which goes straight into the bin.
• The tail end of a spool that is too short to trust on a long job.
• Filament ruined by moisture, tangles, or a jam.

Resin has the same shape with different names: the bottle price divided by the millilitres that reached a finished part, after supports, failed layers, the film of resin that drains back cloudy, and what is lost in washing.

Cost per gram = spool price ÷ grams that ended up in a sellable part.

The measurement that gets you there costs nothing. Weigh a spool at the start of a week and again at the end, count the parts you can actually sell from that week, and divide. That single number quietly captures purge, supports, failures, and the unusable tail without you tracking any of them individually. Do it across a few weeks and you have a material cost per part that a slicer estimate will never match.

Machine hours are the real material

A printer is not free while it runs, and more importantly the plate can only do one thing at a time. Work out an hourly rate the same way a shop rates any machine, from four parts you can measure:

• Power. Nameplate draw is peak, not average, so it will overstate you badly. Put a plug meter on the printer for a full job and read the kilowatt-hours, then multiply by the rate on your own bill.
• The printer, amortized. Purchase price divided by a realistic lifetime in printing hours. Be honest about “realistic” rather than optimistic.
• Wear parts. Nozzles, belts, bed sheets and build plates, hotend components, and for resin machines the vat film and the screen, which has a rated life measured in exposure hours. Each is a replacement cost divided by the hours it lasts.
• Maintenance time. Tramming, cleaning, unclogging, and the tinkering that keeps a machine printing. It is labor, it is recurring, and it belongs in the rate rather than in your evenings.

Add those and you have a cost per machine hour. A print then spends the hours it takes at that rate, exactly like a material. This is the same machine-time convention the 3D-print costing page uses when it walks a print through filament plus amortized machine time, and it is the one habit that makes long prints price themselves correctly.

One consequence is worth sitting with. If your plate is the constraint on how much you can sell, then every hour it spends on a low-margin item is an hour it cannot spend on a good one. Adding a printer buys you hours; it does not make the hours free.

Failure is proportional, and that makes it different

Ceramics loses pieces at discrete stages. Printing loses them anywhere, and the cost of a failure depends on how far it got. A plate that fails in the first layer costs almost nothing. A plate that fails near the end has consumed nearly all of its material and nearly all of its machine hours and produced nothing at all.

So the number that matters is not a failure count, it is the share of your material and your hours that ended up in sellable parts. Track two things per week and you have it: total machine hours run, and hours that produced parts you could sell. The ratio between them is the uplift every good part has to carry.

The arithmetic is the same as any yield allowance. If four in five of your machine hours produce sellable work, each good part carries a quarter again on top of its own hours. Nine in ten, and it carries about a ninth. The rate is yours to measure, and it will differ wildly between a proven file you have run a hundred times and a new one you are still tuning.

Tuning deserves its own line, in fact. The prints you burn dialing in a new file are a development cost of that product. Price them into the product’s first run rather than pretending they were a hobby.

Where the per-unit cost lies to you

Six ways a 3D-print cost comes out too low:

1. Pricing the grams the slicer reported. That figure excludes purge, excludes failures, and often understates supports.
2. Leaving machine hours out entirely. The most common and the most expensive omission, and the reason a twelve-hour print and a one-hour print get sold at similar prices.
3. Not amortizing failures. Failures are a rate, not an accident, and the good parts pay for them.
4. Forgetting post-processing. Support removal, seam cleanup, sanding, gluing multi-part assemblies, priming, and painting are usually the largest labor line in the business, and they scale one for one with volume. A second printer doubles your output and doubles your finishing hours; it does not make finishing cheaper per unit. This is the line that decides whether a print farm is profitable or just busy.
5. Costing an assembly as one print. A model that comes off the plate in six parts is six chances to fail, six lots of support removal, and one product. Cost the whole assembly.
6. Ignoring the license on a file you did not design. If you sell prints of somebody else’s model, a commercial license is a cost of the product, whether it is charged per unit or as a recurring fee. Read the terms for every file you sell and count what they cost you.

Packaging and shipping deserve a mention too. Printed goods are often bulky relative to their weight, so a box that is mostly air can cost more to ship than the plastic inside it. Cost the box and the fill, not just the part.

Measuring your own numbers

Keep a plate log, in the spirit of the batch records and consistency chapter. For each plate: the file, the material, the slicer’s grams and hours, whether it succeeded, and the finishing minutes it took afterward. Alongside it, keep the weekly spool weighing and the weekly hours-run against hours-sold.

A month of that replaces every estimate in this chapter with your own figures: a real cost per gram, a real cost per machine hour, a real yield, and a real finishing time per product. Those four are the whole model. When filament prices move, only the first one changes, and you will know by how much.

Two things are worth reading beside that log. Printing produces a lot of products with respectable-looking margins that are quietly bad uses of a plate, and a thin margin is not automatically an underpriced one, which is the distinction the margin health article draws. And once a year of plates becomes a tax number, COGS for makers is the plain-English version of what that number is and what it is allowed to include.

What carries straight across

Selling is the same business the rest of this playbook describes. Where to sell applies unchanged, marketplace fees come out of the same margin, and the wholesale math in craft fairs and wholesale holds: a shop paying roughly half your retail only works if your cost, machine hours and all, still leaves room underneath. Photography matters more than most sellers expect, because a printed object’s appeal is surface finish, and that is a lighting problem before it is a printing one.

Get the cost per gram, the cost per machine hour, the yield, and the finishing time honest, and the rest of this playbook applies to prints just as it does to every other craft it covers.