Layer
My daughter's elementary school PTA found out I owned "one of those 3D printer things" about four years ago, and every fall since, that fact resurfaces right before the harvest festival fundraiser. This year the ask was bigger than usual: 300 custom keychains with the school mascot, to sell at the festival booth for $5 each, with three weeks' notice. I said yes before doing any of the math, which is how I ended up learning more about production printing in three weeks than I had in the previous four years combined.
The math that should have scared me
Three weeks sounds like plenty of time until you actually schedule it. I have one printer, and the keychain design — a 45mm mascot silhouette, about 4mm thick — took roughly 11 minutes per piece once I'd optimized the plate layout to fit six at a time on the bed. Six keychains per 11 minutes doesn't sound bad until you multiply it out: 300 keychains at six per plate is 50 plates, and 50 plates at 11 minutes each is just over 9 hours of pure print time, which should have been the easy part.
The actual time sink was everything around the printing: loading a new plate every 11 minutes for 9 hours doesn't happen in one sitting when you also have a job and a kid who needs picked up from the actual harvest festival planning meetings. Spread realistically across evenings and weekends, the print queue alone ate most of two weeks.
What broke first
Not the printer — my color-swap habit broke first. The design looked best in two colors, mascot body and a contrasting accent for the school initials, which meant either a manual filament swap mid-print (slow, and risky on a design this small) or batching by color, printing all 300 in the base color first and painting or swapping the accent separately. I switched to batching after the first eight attempts at a mid-print manual swap left visible color-bleed on five of them.
The failures that actually cost time
Across 50 plates, I had:
- 3 plates with one keychain each popping loose mid-print from a corner lifting, salvageable by reprinting just that one piece in the next batch
- 1 full plate failure from a filament tangle on the spool, caught about 40 minutes in and restarted
- 6 individual keychains with a small layer shift from me bumping the printer enclosure while swapping spools, which I didn't catch until final inspection
Total loss: about 10 keychains' worth of time and filament out of 300, roughly 3%, which in hindsight was lower than I expected going in blind on a design I'd never mass-produced before.
The part nobody warns you about
Printing 300 of anything is the easy half of a project like this. Nobody warned me about the keyring assembly — 300 split rings, each one opened slightly and threaded through a printed loop, is genuinely more physically tiring on your hands than anything the printer did. I recruited my daughter and two of her friends for an hour of "keychain assembly line" with cookies as payment, which turned out to be the most efficient hour of the entire project.
What sold, and what didn't
The mascot design outsold a second, more generic "school pride" design I'd also printed as a backup, by roughly four to one. Kids and parents wanted the specific, recognizable character, not a generic shape with the school colors — a lesson that seems obvious written down but wasn't obvious to me while I was choosing what to print weeks earlier. We sold 274 of the 300 at the festival and the PTA sold the remaining 26 at a second smaller event the following week.
Would I do it again
Yes, but differently. Next time I'd ask for a four-week window instead of three, batch the color swap decision before printing a single piece instead of discovering the problem eight attempts in, and recruit the assembly-line help before the final week instead of as a last-minute scramble the night before the festival. The actual printing turned out to be the most predictable part of the entire project — everything that went sideways was logistics, not plastic.
One thing I'd tell anyone asked to do this
If your own kid's school, or anyone's, ever asks you to do something like this with "just one printer," multiply your honest time estimate by at least 1.5 before you say yes. The print time was never the bottleneck. The bottleneck was always going to be a human sitting at a desk, swapping plates, inspecting parts, and threading key rings — and that part doesn't get faster just because the printer is fast.
The booth itself
Watching people actually buy the keychains at the festival booth taught me almost as much as building them did. We'd laid out all 300 in neat rows sorted by color, which looked tidy but turned out to be the wrong display choice — kids wanted to dig through a bin and "find" their favorite one themselves, the way they would with a claw machine or a bargain bin. Switching to a shallow cardboard box they could rummage through, instead of the neat rows, visibly increased how long kids lingered at the table, and the PTA volunteer running the booth said the same thing happened at the dunk tank and the cupcake walk next to us: a little bit of searching and choosing made the purchase feel more like an activity than a transaction.
Pricing also mattered more than I expected going in. We'd debated $4 versus $5 per keychain for most of the planning meeting, and settled on $5 mostly because it's an easier number to make change for with cash. In hindsight, bundling — three for $12 instead of $5 each — would probably have moved more volume, since several parents bought exactly one keychain per kid in their group rather than letting siblings pick different designs, which a bundle price would have nudged toward buying more.
What I'd print differently next year
If the PTA asks again next fall — and after this year's turnout, I'd bet they will — I'd design the mascot keychain with a slightly thicker base and a reinforced loop from the start. A handful of the 274 sold keychains came back to me over the following weeks with a cracked keyring loop, almost always from being looped onto a backpack zipper rather than an actual key ring, which puts more side-load stress on that thin connection point than it was designed for. A small design change now, informed by how people actually used the first batch, would save a round of repair prints later.