The 6-Step Custom Parts Ordering Checklist That Would Have Saved Me $11,000

The $11,000 Mistake That Turned Into This Checklist

I've been handling custom manufacturing orders for six years. In that time, I've personally made (and documented) 17 significant mistakes, totaling roughly $11,000 in wasted budget—not counting the engineering hours I'll never get back. I now maintain our team's ordering checklist. It's caught 47 potential errors in the past 18 months.

This article is for you if you source parts—metal 3D prints, CNC machined components, or custom one-offs—and you've ever been surprised by a cost increase or a rejected batch. Six steps. Every step has a checkpoint. Do them in order.

Step 1: Confirm the Machine Can Actually Produce Your Part

The first mistake I made in this role—back in 2017—was assuming a metal 3D printer could handle any geometry I threw at it. I designed a bracket that was 15 mm taller than the build volume of the vendor's machine. The file was ready to submit. So glad I checked the published specifications on the 3D Systems official website before sending it. I was one click away from a two-day "cannot build" response and a week of redesign.

What exactly do you verify? The bounding box of your part, including supports. The minimum feature size. The minimum wall thickness. And the maximum unsupported overhang angle, which matters more for metal powder bed fusion than most engineers assume. As of February 2025, the material datasheets on the 3D Systems official website include all four values for each metal alloy. Read them.

Checkpoint: confirm all four values in writing before you submit a file.

Step 2: Ask What Post-Processing Is Actually Included

The quote said "metal 3D printed bracket, $85 per piece." The final invoice said $142 per piece. The gap was entirely post-processing: stress relief, support removal, bead blasting, and a couple of CNC finishing passes. I didn't ask. I assumed "printed" meant "finished." It doesn't.

Why does this matter? Because post-processing affects your lead time as much as your budget. A print can be done in three days and still sit in queue for two weeks waiting for the vendor's wire EDM. On a tight schedule, that's the difference between hitting a milestone and explaining to your manager why the prototype is late.

Checkpoint: ask for a step-by-step process list, from print to packed box, and confirm which steps are inside the quoted price.

Step 3: Put Numbers on Your Tolerances

Here's the thing: "tight tolerance" means nothing to a machine shop. Neither does "nice and snug." I once ordered CNC multi-axis Swiss turning parts with a note saying "keep it tight." The vendor held ISO 2768-m and called it a day. The parts came back—all 60 pieces—and they didn't fit the thermal assembly they were designed for. $900 of scrap and a red face at the kickoff meeting.

Look, I'm not saying you need to become a GD&T expert. But you do need numbers. If you don't know what tolerance to specify, ask the vendor what their standard is, then decide if that works for your application. One sentence in the RFQ would have saved that $900: "Critical bore tolerance: ±0.02 mm. Confirm capability before quoting."

Checkpoint: every dimension that matters has a number on it, and the vendor has confirmed they can hold it.

Step 4: Ask About Tooling Before You Commit

This is the one most people skip, and the one that's cost me the most schedule time. In September 2024, I designed a machined slot that required an indexable side milling cutter the vendor didn't have in-house. They ordered the tool. One week delay. $300 tooling charge on a $2,100 order. The part was fine. My timeline wasn't.

Most vendors will tell you if they have the tooling—but only if you ask. It's not a trap; it's a capacity question. Add it to your RFQ as a standard line:

"Is this feature within your standard tooling envelope? If not, what is the special tooling cost and lead time?"

I'm not 100% sure people realize how much this question saves. I've seen the same oversight add more than $2,500 in special tooling charges on larger production runs. One email upfront would have prevented all of it.

Checkpoint: the vendor has confirmed tooling availability in writing before you approve a quote.

Step 5: Spell Out the Process You Mean

I'm still a little embarrassed by this one. In 2023, I sent an RFQ that said "laser finishing required." The vendor came back with options that were clearly aimed at cosmetic dermatology. They thought I was asking about skin treatment. That's when I learned how wide the gap is between industry terms. If you search "coolpeel vs co2 laser," you get clinical results. If you search "industrial CO2 laser cutting," you get machining results. Same acronym, completely different worlds.

In a B2B RFQ, vague process names are dangerous. "Laser finishing" could mean laser marking, laser cutting, laser engraving, or laser polishing. Each one needs a different machine and produces a different result. Spell it out:

"CO2 laser cutting, 3 mm 6061 aluminum, deburred, no heat tint on cut edges."

It feels overkill. It's not. Ambiguity in sourcing is expensive in both money and weeks.

Checkpoint: your RFQ names the process, the material, and the expected outcome.

Step 6: Compare Total Cost, Not Unit Price

The data said Vendor B was 15% cheaper. Same material, similar specs, better payment terms. My gut said something was off. I went with the spreadsheet. What I didn't know was that their "similar specs" didn't include the bore tolerance my design required. The rework, the rush freight, and the extra engineering time put us $1,400 over Vendor A's original quote. The savings vanished. It became a loss.

Part of me wants to say "always trust your gut." Another part knows my gut was only shouting because I had incomplete data. The fix isn't intuition. It's a total cost calculation that includes setup fees, shipping, post-processing, probable scrap, and the cost of your own time to fix problems.

I use a simple formula:

(unit price × quantity) + setup + shipping + (expected rework rate × unit cost) + engineering hours.

Most buyers stop at the first three. That's what I did. That's how a 15% "saving" turned into a $1,400 loss.

Mistakes I Still See (and Made)

A few wrap-up notes, drawn from repeat incidents—mine and the people I train:

• Not reading the material datasheet before designing. The published specs for 3D Systems metal 3D printers tell you minimum wall thickness, surface finish as printed, and maximum unsupported angle. Those numbers should influence your design, not surprise you after the first order.
• Not asking what's in stock. A cheap price on a material with a six-week lead time is not a bargain. Ask about bar stock and powder availability in the RFQ.
• Assuming "standard" means the same thing to every vendor. One shop's standard tolerance is another shop's precision work. Get them to say it explicitly.
• Approving a file from a screen without rotating the model in space. I did this once. The feature was flipped. The shop followed my file, not what I meant. My mistake. (Mental note: I really should stop approving files in a hurry.)

That's the checklist. Six steps, each with a checkpoint. If you take only one thing from this, make it Step 4—the tooling question. It's the cheapest question to ask and the most expensive to skip.