Somebody asks what you would charge to print four small parts. You look at the spool, do the sum everyone does — material times three — and say twenty dollars.
On the job below, twenty dollars is nine dollars short of what it costs you to make. Not short of profit. Short of cost.
That gap is not exotic. It is the ordinary result of pricing the object instead of the process, and it is why a lot of people quietly conclude that printing for other people is not worth it. They are right, at their prices.
The job
Four identical PLA parts on one plate. 340 g of filament between them, 7.5 hours on a machine that meters about 100 W, a spool that cost $20 a kilogram, and 45 minutes of your hands: slicing, plate prep, watching the first layer, popping them off, deburring, bagging. One print in ten on this machine goes in the bin.
Run that through the cost calculator at $25 an hour for your time and 30 cents an hour for machine wear:
| Filament, 340 g at $20/kg | $6.80 |
| Electricity, 0.75 kWh | $0.13 |
| Machine time, 7.5 h | $2.25 |
| Consumables | $0.20 |
| Your 45 minutes | $18.75 |
| Failure allowance at 10% | $1.04 |
| Job cost | $29.17 |
Three times the filament is $20.40. The rule of thumb does not reach cost, let alone price.
The line that dwarfs everything else
$18.75 of a $29.17 job is your own hands. The plastic — the thing the customer thinks they are buying, the thing you quoted on — is $6.80.
Take your time out of the calculation and the same job costs $10.42, which is $2.61 a part. That is the number most hobby quotes are secretly built on, and it is the number that makes "three times filament" feel generous. It feels generous because you have already decided, without noticing, to work for nothing.
There is nothing wrong with working for nothing if that is a choice you are making. It stops being a choice the moment somebody else is scheduling your evening.
Margin is not markup, and the confusion always costs you
Second trap, and it is pure arithmetic.
Add half again to your cost and you have not taken half the sale. You have taken a third of it: the profit is half the cost, and the cost is two thirds of what the customer pays.
If what you meant was "half of what they pay is mine", the price is the cost divided by 0.5. On a per-part basis that is $14.59 against a $7.29 cost — a 100% markup on cost, which is the same thing as a 50% margin on the sale, described from the other end.
People who say they work on 50% and calculate it the first way are running at a third, and discovering it at the end of a quarter. The resale pricing guide has the mechanics; the markup question has the shorter version.
What a minimum charge is actually for
A twenty-minute keychain and a nine-hour bracket need almost exactly the same 45 minutes of you. Slicing, plate, first layer, removal, cleanup, packaging, message thread. The machine time scales with the job; you do not.
That is what a minimum charge encodes. It is not a surcharge for being annoying, and it is not padding — it is the fixed cost of touching a job at all, and the reason the second identical part of anything is always disproportionately cheap to make. A price list built purely per gram gets this exactly backwards: it charges the most for the jobs that cost you the least attention, and the least for the ones that eat an afternoon.
If you want one structural change to your pricing, make it this: a floor, plus a per-part figure, and quote per plate rather than per part where the plate is the real unit of your time.
The three costs you cannot bill and still have to recover
Quoting. Every job you price and do not win was real time. If you win one in three, the two you lost are an overhead on the one you won.
The design conversation. "Can you make it a bit thicker" is twenty minutes in CAD plus a re-slice plus a re-print. If it is not written down as chargeable, it is a gift, and customers cannot tell the difference between a gift and a norm.
Reprints you chose to do. The failure allowance in the table covers prints that failed. It does not cover the ones that succeeded and that you reprinted because a layer line annoyed you. That is a quality standard, it is a legitimate business decision, and it is a cost you are absorbing on purpose — so absorb it on purpose, in the price, not by accident, out of your margin.
The failure rate is yours to fix and theirs to fund
One in ten on the job above adds $1.04. That is small because the parts are small. On a long print it is not small, and the multiplier is worse than the rate — the failure allowance is 1/(1−f), not f, because you need more than one attempt per delivered part.
Two things follow. A customer should never see a line item called "failures", because it reads as an admission that you are unreliable; it belongs inside your part price, quietly. And a failure rate you have not measured is a failure rate you are funding out of pocket, which is why what a failed print really costs is worth more attention than most people give it.
Quoting so it survives contact
A few habits that hold up.
- Quote a total, not a breakdown. A customer given a breakdown will argue with the cheapest line, which is always the plastic, and never with the expensive one, which is always you.
- Price the material as a band, not a figure. PLA at retail spans roughly a factor of two, and the cheap end has its own failure rate. Quoting against your actual last purchase, not against the cheapest listing you have ever seen, is the difference between a margin and a hope.
- Put a validity period on it. Filament prices move, and a quote that is six months old is a promise you did not mean to make.
- Charge for the file if you made the file. Modelling is a separate service that happens to be delivered as a print.
- Say no to the ones that are cheaper to buy. Sometimes the honest answer is that it is cheaper to buy the thing, and saying so buys more trust than the job was worth.
Where to go from here
If you want a starting figure rather than a method, what to charge for a 3D print gives one, with the assumptions attached. But the place to start is not a price list. It is knowing your own cost per job, on your own machine, with your own failure rate — what a print actually costs you works one all the way through, input by input.
Then decide, explicitly and in writing to yourself, what your hour is worth. Everything above is downstream of that one number, and it is the only input on the whole list that nobody else can supply. If the answer turns out to be that the work does not pay at any rate a customer would accept, that is a real answer, and it is much better arrived at now than after you have committed to a delivery date. Anyone thinking about doing this at volume should look hard at what a small print farm costs to run before buying a second machine, because the second machine does not halve the 45 minutes.