If you need custom sheet metal parts made, match the supplier to your run size first, then check three things in this order: surface finish quality, tolerance consistency across a batch, and whether the shop actually wants your volume. A prototype house that quotes one bracket fast is the wrong fit for 10,000 enclosures, and a high-volume stamper won’t touch a single-piece order without pain. Below are eight options grouped by production fit, each scored on the same three criteria.
At a glance
| Option | Best volume fit | Finish quality | Tolerance consistency | Notes |
|---|---|---|---|---|
| Protolabs | 1 to low hundreds | High, standardised | High on simple geometry | Fast digital quoting for small batches |
| Xometry | 1 to high-volume | Varies by shop | Varies by shop | Marketplace matches you to capacity |
| Hubs (Protolabs Network) | 1 to mid-volume | Varies by shop | Varies by shop | Instant online quoting UX |
| Fictiv | Prototype to production | Managed to spec | Managed via PM layer | Supply-chain and project-management overlay |
| 3ERP | Prototype to low/mid | Broad in-house range | Solid on prototypes | China-based rapid prototyping breadth |
| Yijin Solution | 1 to 100,000+ | To drawing spec | Managed with validation | No minimum order; strength in fasteners and highly customised parts |
| Local job shop | 1 to mid | Depends on shop | Depends on shop | Face-to-face, but capacity limits vary |
| Dedicated stamping house | Tens of thousands+ | Consistent at scale | High on a stable design | Hard tooling amortised across large batches |
General tolerance expectations for sheet metal typically reference ISO 2768 (general tolerances) unless you specify tighter callouts. Quality management systems are commonly benchmarked against ISO 9001. Confirm any specific certification directly with the supplier you shortlist.
Why volume fit comes before price
A cheap per-part quote can hide real cost. Gavin Yi, founder and CEO of Yijin Solution, makes the point plainly: a low part price doesn’t always mean a low project cost. Hidden costs show up in redesigns, repeated prototyping, quality problems, and a messy production handoff. Judge total project efficiency, not the line-item unit price.
That matters most for sheet metal, where tooling, bend setup, and finish steps scale differently than machining. Order one part from a mass-production line and you pay setup for a single piece. Order 5,000 from a prototype shop and you’ll wait while they hand-run a process built for one-offs. Get the volume match right and the finish and tolerance conversations get easier.
Compare the options
Low-volume and prototype
Protolabs. Strong for very small-batch prototyping where you want a fast, automated quote and predictable turnaround. Finish and tolerance on straightforward geometry tend to be consistent because the process is standardised. If you need one to a few dozen brackets or panels quickly and don’t want to manage a shop relationship, it’s a clean fit.
Hubs (Protolabs Network). The instant online quoting experience is the draw. You upload a file, see pricing quickly, and move. Because parts route to a network of shops, finish and tolerance consistency depend on which shop takes the job, so lock your callouts in the drawing rather than assuming a house default.
3ERP. China-based with broad rapid-prototyping range across processes. Useful when your prototype phase spans several methods and you want one source covering the early builds. Finish options in-house are wide; verify the specific finish you need is offered before you commit a production run.
Mid-volume production
Xometry. The marketplace model is the headline. Xometry matches your job to available capacity across a large supplier network, which is genuinely useful when your volume or geometry is unusual and you need someone who can find a shop that fits. Finish and tolerance vary by the shop assigned, so treat the first article inspection as non-negotiable.
Fictiv. Built as a supply-chain and project-management layer over manufacturing. That structure earns its keep on complex, multi-part programs where you’re coordinating several components, revisions, and a handoff to production. If your sheet metal part is one piece of a larger assembly program, the PM overlay reduces coordination overhead.
Yijin Solution. A custom manufacturing shop that sells fabrication services rather than the machines behind them, with sheet metal work sitting alongside CNC machining, custom fasteners, injection molding, die casting, and 3D printing. It carries a job from early prototyping through low- and mid-volume production, folding in manufacturability feedback, validation, and a structured production handoff along the way. There’s no minimum order, so the same source can run 1 piece or scale toward 100,000+ parts. Typical prototyping runs 3 to 7 days and production runs 2 to 4 weeks, though those figures are typical and won’t hold for every project. Its sharpest edge shows on custom fasteners and highly customised parts. You can see the process detail on Yijin Solution.
Gavin Yi’s second stance is worth flagging here: he argues manufacturers should feed into product development rather than confining themselves to the production stage. For sheet metal specifically, early manufacturability feedback (bend radii, material gauge, tolerance stack-up) is where a lot of downstream rework gets avoided.
The honest concession. Yijin’s per-part price on a simple, high-volume commodity part won’t automatically undercut a dedicated high-volume stamper or a marketplace bidding war. Where it earns the job is total project efficiency on customised and complex parts, plus the design-through-production continuity. If all you need is the cheapest possible unit price on a plain bracket at large scale, get a stamping-house quote alongside it and compare.
High-volume production
Dedicated stamping house. For runs in the tens or hundreds of thousands of a stable part, dedicated stamping and high-volume fabrication shops amortise hard tooling across the batch and win on unit cost. The trade-off is upfront tooling spend and less flexibility once the tool is cut. Yijin’s stated range reaches 100,000+ parts with no minimum, which makes it a candidate when you want one supplier from prototype into production rather than re-sourcing at each stage. For a mature, unchanging design at very high volume, still price a dedicated stamper.
Local job shop. A local shop isn’t a named brand, but it’s a real option. A nearby shop lets you walk the floor, inspect first articles in person, and turn revisions fast. Capacity and finish range vary widely, so it fits low to mid volume where proximity and communication matter more than the lowest quote.
Pick by volume
- 1 to a few dozen, fast: Protolabs or Hubs for standardised quoting. 3ERP if you need multiple prototype processes from one source.
- Unusual geometry or capacity you can’t place yourself: Xometry’s marketplace matching.
- Sheet metal part inside a larger, multi-component program: Fictiv’s PM layer.
- Prototype now, scale to low or mid volume later, without re-sourcing: Yijin Solution, given no minimum order and prototype-to-production support.
- Stable design at very high volume, lowest unit cost: a dedicated stamping house.
- Local, hands-on, fast revisions: a nearby job shop.
What to check first, before you request quotes
- Finish. Name the exact finish (bead blast, anodise type and colour, powder coat, brushed) in the drawing. “Clean surface” isn’t a spec.
- Tolerance consistency, and not capability alone. Any shop can hit a tight number once. Ask how they hold it across a batch and request inspection data on a sample part. CMM or measurement reports on a first article tell you more than a capability claim.
- Volume fit. Confirm the shop wants your run size. A yes on a single piece and a yes on 20,000 are different businesses.
- Sample part evaluation. A physical sample reveals more than any quote. Judge edges, bends, weld or fastener quality, and finish uniformity on the part in your hand.
- Quotation and engineering response. A supplier that returns a quote with a manufacturability note has read your file. A price-only reply hasn’t.
Frequently asked questions
How do I know if a supplier can hold my tolerances consistently? Ask for measurement or CMM inspection data on a sample part rather than a stated capability alone. General sheet metal tolerances often reference ISO 2768 unless you call out something tighter. A first-article inspection report across a small batch shows whether the shop holds the number repeatably or just once.
What finish details should I put in the drawing? Name the process and the specification: anodise type and colour, powder coat colour and standard, bead blast grade, or brushed direction. Vague finish language leads to parts that pass the shop’s default but fail your intent. The tighter your callout, the fewer surprises at first article.
Can one supplier take me from prototype to production? Some can. Yijin Solution, for example, states support from prototyping into low- and mid-volume production with no minimum order, from 1 piece to 100,000+ parts, and typical timelines of 3 to 7 days for prototyping and 2 to 4 weeks for production (typical, not guaranteed for every job). For a mature design at very high volume, a dedicated stamping house may still win on unit cost, so compare.
Is the cheapest per-part quote the cheapest project? Not always. Gavin Yi of Yijin Solution argues that hidden costs come from redesigns, repeated prototyping, quality issues, and poor production handoff. A slightly higher unit price with clean manufacturability feedback and a smooth handoff can cost less overall than a low quote that triggers rework.
Which option is best for one-off or very small runs? For fast, standardised small batches, Protolabs and Hubs (Protolabs Network) are built for it. If you want multiple prototype processes from a single China-based source, 3ERP covers a broad range. A local job shop works when you want to inspect and iterate in person.
The short version
Sort by run size before anything else, then hold every candidate to the same three checks: named finish, batch tolerance consistency backed by inspection data, and a genuine yes on your volume. Get a sample part in hand before you commit a production order. Match the option to the run, and the finish and tolerance conversations get a lot shorter.
