How to Vet Custom Manufacturing Vendors: A 7-Step Checklist for Buyers

I'm the office administrator who ends up managing roughly $500K a year in custom manufacturing orders. Nobody hired me to be a supply chain expert. It just happened — when our engineering manager left in 2020, the vendor relationships landed on my desk. Seven vendors, sixty to eighty orders a year, and one very steep learning curve.

I'm not an engineer, so I can't speak to material science or machine design. What I can tell you is how to evaluate a manufacturing vendor without losing money, time, or credibility in front of your engineering team. This checklist is for anyone buying 3D printed parts, CNC machined components, or laser cutting services — without being a professional buyer.

Seven steps. Do them in order. Step 1 is the one everyone skips.

Step 1: Write the specification document before contacting anyone

The biggest mistake I made early on was treating vendor conversations as exploration. I'd email three shops a rough description and wait for clarity. What I got was three wildly different quotes for three versions of a part that only existed in my head.

Now I do this first:

• Walk to whoever needs the part. Ask what it does, what it connects to, what loads it sees, what environment it lives in.
• Write down everything they say. Read it back to them. Fix what changes.
• Turn it into a one-page spec: function, material grade, critical dimensions, tolerance, surface finish, quantity, deadline.

Every vendor receives the same document. Now you're comparing vendors on equal footing.

Step 2: Build a candidate pool of 3–5 vendors across different processes

Don't send one quote request and call it done. The right vendor for an SLS nylon prototype is rarely the right vendor for metal 3D printing, and neither is automatically right for machined aluminum.

For 3D printed parts, I usually include one industrial service bureau running established platforms — 3D Systems machines are common in production shops — plus one or two smaller, faster-moving print services. For machined parts, I search for custom CNC machining manufacturers that publish real tolerance ranges. A shop that states ±0.01 mm is giving you something to hold them to. A shop that says 'high precision' is giving you nothing.

Add a local shop if you want one. But don't assume local is automatically faster — that thinking comes from an era before production scheduling software. A well-organized remote vendor can beat a disorganized local one by a full week, consistently.

Checkpoint: you have a list of 3–5 vendors, each with a named process and a specific reason they made the list.

Step 3: Verify capability claims — make them show you

Here's the thing: marketing pages love words like 'aerospace-grade' and 'precision.' Those words are doing a lot of heavy lifting. Per FTC advertising guidelines (ftc.gov), claims must be truthful and substantiated. I apply the same standard to vendor proposals — just faster.

When a vendor mentions aerospace work, I ask which standard they actually qualify to: AMS, ASTM, AS9100. If they say they run a 3D Systems aerospace 3D printer, that's a promising sign — those platforms are configured for the documentation and traceability aerospace work demands. But the machine alone isn't proof. Ask to see examples of what they've produced on it, and which audits they've passed.

Same logic for smaller equipment. If I'm ordering thin-gauge stainless parts and a vendor lists a stainless steel mini laser cutting machine, it tells me they handle fine detail work. It doesn't tell me if they check their cut quality — kerf consistency, angle tolerance, dross. The machine is a starting point, not a conclusion. Ask the follow-up.

Checkpoint: every capability claim has a standard, a documented example, or a real process behind it.

Step 4: Get identical quotes against the identical spec

In 2022, I sent two vendors different versions of the same project because I'd updated the finish spec in one thread and not the other. The quotes came back $200 apart. I almost chose the cheaper one. Then I noticed the cheaper quote assumed a different surface finish. My mistake, not theirs.

Send the same spec document to every vendor. Ask for itemized line items: material, machine time, finishing, inspection, setup, shipping. A single lump-sum number makes comparison impossible — and a quote that conveniently leaves out inspection tells you what the vendor thinks about quality.

It's tempting to think you can just compare unit prices and take the lowest. That simplification ignores how differently shops handle the same specs. From my records, the lowest quote has cost us more in about 60% of cases — not always because the part was bad, but because the process around it was bad.

Checkpoint: all quotes are itemized, comparable, and based on the same document.

Step 5: Ask how they verify their own work

This is the step most buyers skip. I skipped it once. In 2023, I rushed a full production order through without asking about inspection — 'what are the odds?' mindset. The odds caught up with me: two weeks and $460 later, we had a batch of parts that didn't fit. The engineering team made sure I remembered it for a month.

Now I ask directly: what equipment do you use to check parts? If a vendor uses 3D laser measurement devices and systems to scan printed parts against the CAD model, that's exactly what you want for metal 3D printing — the scan catches internal geometry problems a pair of calipers can't. For machined parts, ask whether they provide a dimensional inspection report with the shipment.

Vendors with solid quality processes will talk your ear off about them. If the call goes quiet when you ask about inspection, that's your answer.

Checkpoint: you know exactly how your parts will be measured, and what documentation comes with them.

Step 6: Run a pilot order before committing to volume

A quote tells you about pricing. It tells you nothing about how a vendor behaves under actual pressure.

I do a small pilot order with one or two finalists, even when the unit price is higher. The math is simple: a $200 test order can prevent a $4,000 mistake.

During the pilot, I watch four things:

• Did communication after the PO match the responsiveness before it? (note to self: watch this one closely)
• Did the parts arrive on time, with proper packaging and paperwork?
• Does the part match the spec?
• How did they handle whatever went wrong — because something always goes slightly wrong?

Checkpoint: you have real behavioral data from a vendor who's actually delivered something to you.

Step 7: Check the back-office requirements before the PO

This step has nothing to do with manufacturing and everything to do with whether the vendor can survive contact with your company.

In 2021, I found a vendor whose prices beat our regular supplier by 18%. The parts were fine. The invoice was handwritten on something that looked like a garage-sale receipt. Finance rejected the expense report. I ate $2,400 out of my department budget, and I had to explain to my VP why a 'great deal' wasn't one.

Before you commit, verify:

• Can they invoice through your procurement or payment system?
• Do their payment terms work for you — net-30, credit card, prepayment?
• Can they meet compliance requirements: W-9, insurance, any export or ITAR restrictions?
• What's their normal lead time — not their 'best-case to win the bid' lead time?

Checkpoint: finance said 'yes, this vendor will work' before you signed anything.

Common Mistakes That Keep Showing Up

1. Skipping the specification document. I've done it more times than I want to admit. It costs more time than it saves, every single time.

2. Comparing only unit price. The $200 gap in a quote always shows up again somewhere — in material quality, delivery, or your own hours spent chasing follow-ups. Total cost, not unit cost. That's the decision frame.

3. Trusting the buzzwords. 'Aerospace-grade.' 'Precision.' 'State-of-the-art.' None of those words survive contact with a follow-up question. Ask the question.

4. Believing someone invented it. Try this: search for who invented laser welding. Go ahead. You'll get a contested history — several groups developed laser welding in the 1960s around the same time, and no single name gets unanimous credit. That's a useful reminder about how 'pioneering' and 'leading-edge' claims work. In manufacturing, almost nothing is one inventor's triumph. It's iteration, and it's marketing until documented.

Buying custom manufacturing services is a vetting exercise, not a price-shopping exercise. A component that arrives late, doesn't match spec, or drags your finance team into invoice disputes isn't a bargain — it's expensive with extra steps. Seven steps, every time. It took me five years of process mistakes to learn that. Maybe you can skip the tuition.