3D Printer FYIBuy a printer when a named bottleneck repeats. Do not buy one because a vendor said it will “revolutionize” the shop. Many businesses should stay on a service bureau.
Price any in-house hour with the cost worksheet. Pick hardware with how to choose an FDM printer.
Start with the workflow bottleneck, not the printer
Write one sentence: “We wait ___ days for ___ parts, and that delay costs us ___.” If you cannot fill the blanks, you are shopping.
3D printing is the wrong tool when you need injection-molded cosmetics, certified metals, or fifty thousand identical widgets. It is a candidate when you need shape now, jig tomorrow, or twenty unique SKUs.
Prototype, tooling/jigs, replacement parts, customization, low-volume production
| Job | In-house tends to win when | Outsource tends to win when |
|---|---|---|
| Prototype | daily iteration, designer on site | exotic material, fine cosmetics |
| Jigs/fixtures | tribal knowledge, same workbench | one-off precision you cannot measure |
| Replacement | obsolete plastic widget, known fit | structural/safety-critical, unknown resin |
| Customization | customer waits in the shop | you would reprint the same fail twice |
| Low-volume production | you already hold QC and packing | labor/failure would dominate price |
In-house advantages
Same-day iterate. No NDA upload. You learn which filament class actually survives the job. Fixtures stay next to the machine they serve.
Hidden costs and operational burden
Someone maintains the machine. Someone owns failed prints at 11 p.m. Emissions and fire policy still apply in a shop (NIOSH 2024-103). CAD time is usually larger than print time. Spare nozzles and sheets are a line item.
Outsourcing advantages
You buy capability (SLS nylon, MJF, industrial FDM, finishing) without owning it. A service bureau or local shop quotes lead time. You are not on the hook for calibration theater.
Utilization and break-even
Hours of paid/internal work per month you can actually feed the printer: ___
Fully loaded cost per hour (worksheet): ___
Monthly bureau spend you would replace: ___If the printer would sit idle three weeks out of four, keep outsourcing and rent time at a makerspace for spikes. Do not invent a utilization number to justify a purchase.
Quality control and repeatability
Can you measure the feature that matters (calipers, go/no-go, first-article photo)? If not, a pretty print is not a part. Production means a frozen profile and a coupon every material change — see calibration.
Skills and CAD requirements
Someone must own CAD that matches the part. Downloading random STLs is not a product pipeline. If the only CAD skill is Tinkercad, your in-house ceiling is classroom-grade widgets.
A staged pilot plan
- Outsource three real jobs. Keep files and quotes.
- Print the same jobs at a library/makerspace. Time labor honestly.
- Only then buy the smallest machine that clears those jobs.
- Revisit after 90 days of the worksheet, not after unboxing.
Decision matrix
Score 0–2 each. Buy in-house only if the total is high and a person is named.
| Question | 0 | 2 |
|---|---|---|
| Bottleneck is physical parts, not sales | ||
| Same geometry repeats monthly | ||
| We can inspect the critical dimension | ||
| We have CAD + a maintenance owner | ||
| Bureau lead time is the actual pain | ||
| We will run the cost worksheet |
If the painful part is finishing, certification, or metal, keep the bureau. If the painful part is “waiting for a $12 bracket,” a modest FDM cell can be enough.
Sources
- NIOSH: Approaches to Safe 3D Printing
- Service-bureau capability lists (e.g. Hubs, local ISO shops) — compare your quote, not a blog price
- How to choose an FDM printer
- Your last six purchase orders for printed parts (the real dataset)