Start With the Process Window, Not the Price List
Almost every assembly quotation looks the same on the surface: bare boards, components, solder, test. The real differences sit inside the process window — the range of designs a shop can build repeatedly without bending its own rules or outsourcing the difficult parts. Before comparing numbers, compare that window against your actual artwork and stackup.
Begin with board construction. A supplier that only handles single-sided and double-sided rigid boards will struggle the moment your design calls for rigid-flex or flex constructions, and a shop that treats those as occasional orders may not have stable yields. Ask directly about layer count (the range can run from 2L up to 58L), laser-drilled microvias, blind and buried vias, cavity boards, via-in-pad, and heavy copper builds up to 30 oz. If your product depends on any-layer HDI, confirm whether staggered, stacked, and skipped microvia structures are routine rather than experimental.
Fine-pitch work deserves its own conversation. Designs with BGA pitch down to 0.15 mm and trace width/spacing at 0.002″ leave very little margin for a printer or a placement program that is only “close enough.” The same logic applies to controlled impedance: a stated tolerance of ±5% is only meaningful if the manufacturer verifies it on test coupons rather than assuming the laminate behaves as modeled.
Material familiarity matters just as much as machine capability. High-speed and RF designs often depend on specific laminate families — Isola grades such as I-Speed, ASTRA MT77, I-TERA MT40, and IS680; Rogers materials including RO4003C and RO4350B; plus Arlon, Taconic, and Panasonic MEGTRON laminates. A partner who orders these rarely will quote longer and may improvise substitutions. Finishes follow the same rule: ENEPIG, hard gold, gold fingers, and selective gold each solve a different problem, and mechanical details like countersink holes, press-fit pins, castellated holes, and semi-flex cutouts should be confirmed in advance rather than discovered at first article.
The practical test is simple. Send your Gerbers, stackup, and BOM, then read the DFM response. A manufacturer that flags potential assembly conflicts, panelization issues, or component clearance problems before quoting is telling you something useful about how the build will go.
Quality Assurance Is a Paper Trail, Not a Promise
Every supplier claims good quality. What separates them is whether quality is documented, traceable, and repeatable — and whether you can inspect the evidence before placing an order.
Certifications are the entry point, not the finish line. Industry credentials such as ISO 9001, ISO 14001, and IPC certifications appear on the shortlist of nearly every serious operation, so ask for current certificates and check that the scope actually covers the work you need. A certificate for bare board fabrication does not automatically extend to assembly and test, and a scope statement written for commercial electronics may not address the expectations of an aerospace or medical buyer.
Then move to the records behind the certificate. Ask how material lots are tracked from incoming inspection through finished goods, how first article inspections are performed and archived, and how impedance and microvia integrity are verified. For assembly specifically, ask what inspection coverage exists at each stage and who signs off before shipment. You do not need a lecture on equipment brands; you need to know that a defined procedure exists and that someone is accountable for following it.
The most revealing questions concern failure. What happens when a batch is rejected — by your incoming inspection or by the end customer? Who absorbs rework and scrap costs? Is there a formal corrective action process that traces a defect back to its root cause, or does the response stop at “we will remake it”? A supplier with a mature nonconformance system will answer these questions quickly and without defensiveness.
Engineering communication belongs in the same evaluation. Measure how long DFM questions take to come back, whether feedback arrives as specific recommendations or vague warnings, and whether you deal with one accountable engineer or a rotating queue. On complex builds, a single responsive technical contact is often worth more than a marginally lower unit price, because it prevents the small misunderstandings that turn into scrapped panels.
Location, Volume Fit, and the Real Cost of a Cheap Quote
Geography is a legitimate selection criterion, provided it is weighed honestly. Buyers in the United States often prefer a domestic PCB assembly manufacturer for time-zone overlap, faster communication, simpler shipping, and reduced customs friction. Buyers with cost-sensitive, high-volume programs frequently look to Asian suppliers for scale and supply chain depth. Neither choice is automatically correct — what matters is whether the supplier’s location supports your logistics, your responsiveness expectations, and your volume profile without adding hidden overhead.
Volume fit is the criterion most often overlooked. A manufacturer optimized for high-volume, highly automated runs may treat a fifty-piece prototype as an inconvenience, while a quick-turn specialist set up for prototypes and small batches may not be the right home for a mature product shipping in thousands. Ask where your program sits relative to their typical order, and whether the same process controls apply to low-volume work.
Customization and flexibility follow close behind. If your design needs tailored tooling, non-standard test fixtures, mixed-technology boards, or a staged ramp from prototype to production, confirm that those services are offered as a matter of course. Benchmarking against published comparisons helps here — reading an overview of leading pcb assembly manufacturer profiles, including their regional strengths and service models, gives you a realistic sense of what a credible operation offers before you send out RFQs.
Finally, treat cost with discipline. A low quotation saves money once; a build that fails at the customer saves nothing at all. Add up the true cost of a poor partner — rework, expedited freight, engineering time, delayed launch, and reputational damage — and the gap between the cheapest bid and a well-matched one usually narrows sharply. The right selection process compares capability, documentation, communication, and logistics side by side, then lets price settle into its proper place among them.
