How to Choose an Electronics Contract Manufacturer

An electronics contract manufacturer (EMS or CM) builds circuit boards and assembled electronic products for other companies — sourcing components, assembling PCBs, testing them and often handling final assembly and fulfillment. Choosing one is a decision about capability fit, test capability and communication, and the lowest quote is frequently the most expensive option.

Volume fit is the single biggest predictor of a good relationship. A large EMS will not prioritize a 2,000-unit order, and a prototype house cannot absorb 50,000 units. Match your annual volume to a manufacturer where you are a meaningful but manageable customer.

Send every shortlisted manufacturer an identical package so the quotes are comparable: Gerbers and fabrication drawing, centroid and assembly drawings, a complete BOM with manufacturer part numbers and approved alternates, test requirements, expected annual volume with order sizes, and quality expectations.

If your design is not at that level of documentation yet, the earlier stages of electronic product development cover what still needs to be produced.

An electronics contract manufacturer, also called an EMS provider, builds electronic products for other companies. Services typically include component sourcing, PCB assembly, inspection and testing, and often box build — enclosure assembly, cabling, packaging and fulfillment.

The strongest sources are referrals from your design firm or other hardware companies in your category, industry directories and trade shows such as IPC APEX, and component distributors who know which shops build similar products. Shortlist three to five and send all of them an identical quote package.

ISO 9001 is the baseline quality standard, and IPC-A-610 defines workmanship acceptance class. Medical products require ISO 13485, aerospace AS9100, and automotive IATF 16949. Ask for current certificates and the scope they cover, not just a logo on the website.

Costs split into one-time NRE — stencils, fixtures, programming and test development, commonly $2,000 to $25,000 — and per-unit assembly cost that depends on part count, board complexity and volume. Component cost is usually the largest share of unit price at low volumes.

Domestic and nearshore manufacturing offers shorter lead times, easier oversight, lower inventory risk and tariff insulation at higher labor cost. Overseas manufacturing wins on unit price at high volume. Many companies prototype and pilot domestically, then decide based on proven demand.

Fabrication files and drawing, assembly drawings, centroid file, a complete BOM with approved manufacturer part numbers and alternates, test procedures with acceptance criteria, production test firmware, packaging and labeling requirements, and a written change control process.

An electronics contract manufacturer (EMS provider) buys your components, assembles your boards, tests them, and ships finished goods. What separates one shop from another is not whether they can place parts — almost all of them can — but how much of the surrounding work they absorb: design for manufacturability feedback, component sourcing and lifecycle management, test fixture development, regulatory pre-compliance support, and the new product introduction (NPI) discipline that turns a working prototype into a repeatable product.

Choosing badly costs more than money. The wrong shop for your volume will quote well, then deprioritise your build behind larger customers, and every schedule slip lands on your launch. The sequence below is the vetting order we use with clients, from a 10-15 shop longlist down to a pilot build with two finalists.

Volume fit is the single filter that removes most of a longlist. A prototype house that thrives on 25-board runs will struggle at 20,000 units a year; a high-volume plant will quote your 500-unit run at a price that reflects how little they want it.

Before a single call, screen the longlist against the technical requirements of your board. Most shops publish enough on their capability page to eliminate half the list in an afternoon.

Every shop says yes on a capability page. These questions get you an answer that is either specific or evasive, and the difference tells you what you need to know.

Quotes are only comparable once they are decomposed. A shop that looks 12% cheaper on the headline number is often more expensive after non-recurring engineering (NRE), test development and freight.

Your schematic, layout, firmware and test procedures are the asset. Put the protections in the manufacturing agreement, not in a side email.

Award the first production run to two finalists rather than one. A 50-100 unit pilot at each shop costs less than one delayed launch and tells you what no audit can: how they communicate when something goes wrong.

Score both shops on the same sheet, then award volume to the winner and keep the runner-up warm with a small annual order. A second qualified source is the cheapest supply-chain insurance you will ever buy.

For a moderately complex board, expect $300-$1,500 in setup and stencil charges, $1,500-$20,000 for functional test development if you need it, and an assembly cost of roughly $0.01-$0.03 per placement at low volume. The bill of materials usually dominates: on most consumer-scale products it is 45-70% of the landed cost per unit.

Below roughly 5,000 units a year, domestic or nearshore usually wins once you price in freight, duty, travel for audits, and the cost of a two-week feedback loop on every engineering change. Above that, offshore unit-cost advantages start to outweigh the coordination overhead — provided your design is stable and your test coverage is good enough to catch defects without an engineer on the floor.

You should, and it must be written into the quote or manufacturing agreement. Fixtures you funded but do not own are the most common reason a company finds it cannot leave a contract manufacturer without rebuilding several thousand dollars of test hardware.

Plan on 8-14 weeks from longlist to a completed pilot build: two weeks of screening and calls, two to three weeks for quoting and DFM review, four to six weeks for the pilot run including component lead times, and one to two weeks to review results and negotiate terms.

ISO 9001 is the baseline for any regulated or professional product. Add ISO 13485 for medical devices, AS9100 for aerospace, and IATF 16949 for automotive. Ask for IPC-A-610 Class 2 workmanship as a minimum and Class 3 for products where failure carries safety or field-service consequences.

Once two or three shops have quoted and built a pilot, score them on the same sheet rather than on how the last phone call went. Weight the categories before you see the results, so a low price cannot quietly outrank a quality problem.

Award the volume to the winner, keep the runner-up qualified with a small annual order, and revisit the scorecard yearly. A second qualified shop is the cheapest insurance a hardware company can buy.

Work with LA NPDT: if you are moving from here to execution, start with our product design services or talk to us about industrial design and development.