Key Takeaways
- A contract manufacturer (CM) builds to your spec. An EMS provider adds DFM, test engineering, and NPI process.
- Tier 3–4 EMS providers ($10M–$100M) are where hardware startups actually get served. Tier 1 will not take your 5,000-unit order.
- The transition from CM to EMS costs 8–16 weeks and should happen before you scale, not after quality fails.
- An ODM owns the design. A CM/EMS executes your design. If you have proprietary technology, ODM is not an option.
- The five signs you need to graduate: inconsistent quality, component sourcing gaps, no test engineering, volume ceiling, no retailer compliance support.
The Vocabulary Problem
When a hardware founder says “we need a contract manufacturer,” they usually mean one of three things:
- A job shop that builds exactly what you specify. You hand them the BOM, the Gerbers, the assembly drawings, and the test protocol. They produce it. That is a contract manufacturer.
- An electronics manufacturing services (EMS) provider that adds design-for-manufacturability review, test fixture development, component sourcing, and a formal new product introduction process. They bring engineering, not just capacity.
- An original design manufacturer (ODM) that owns the product design and lets you put your brand on it. They design it, they build it, you sell it.
These are three different business models with different capabilities, different cost structures, and different IP implications. A founder who treats them as interchangeable usually discovers the problem 3–6 months into the relationship, when the production process cannot do what they assumed it could.
For a deeper look at the full co-manufacturer selection framework, see our guide. This post focuses specifically on the CM vs. EMS decision for consumer electronics and hardware brands.
What a Contract Manufacturer Does
A contract manufacturer executes to your specification. You provide:
- Bill of materials (BOM) with approved vendor list (AVL)
- PCB Gerber files, assembly drawings, and pick-and-place data
- Test procedures and acceptance criteria
- Packaging specifications
- Quality standards and inspection protocol
The CM owns the production equipment, the floor labor, and the assembly process. They charge per unit plus tooling and NRE (non-recurring engineering) for fixtures and setup. You own the design, the IP, and the QC standards.
Good for: You have a fully validated product design, a finalized BOM, production-ready test procedures, and you need reliable production capacity at a known cost per unit.
The risk: A CM optimizes for their production line, not your product. If the design needs manufacturing optimization — and it almost always does in the transition from prototype to production — you either do that engineering yourself or it does not happen. CMs do not typically flag DFM issues proactively. They build what you send them.
What an EMS Provider Does
An EMS provider adds engineering capability on top of production capacity. The difference matters most during the transition from prototype to production — the phase where most hardware brands discover that their design works in a lab but does not work on a production line.
An EMS provider typically offers:
- Design for manufacturability (DFM). They review your PCB layout, component placement, and assembly sequence for production-line compatibility. This catches problems like components that are too close for automated placement, thermal issues that only appear at line speed, and test point accessibility that matters during production but not during prototyping.
- Test engineering. They develop functional test fixtures, in-circuit test programs, and final QC procedures. For consumer electronics, this often includes burn-in testing, firmware verification, and calibration sequences. A CM expects you to provide this. An EMS provider builds it.
- Component sourcing. They manage the approved vendor list, negotiate component pricing at production volumes, handle allocation constraints, and qualify alternates when primary components go on allocation. Lead times for semiconductors, displays, and connectors run 12–26 weeks. An EMS provider manages this pipeline.
- New product introduction (NPI). A formal process for bringing a new product from engineering release to production — including pilot runs, process validation, yield analysis, and production ramp. This is where the production process gets debugged.
Good for: Post-prototype, pre-mass-production. The product design is stable but the manufacturing process has not been validated at scale. The brand needs manufacturing engineering as a service, not just capacity.
The cost: EMS NRE charges typically run $5,000–$25,000 for test fixture development, DFM analysis, and NPI setup. The per-unit premium over a bare CM is typically 10–25%, but this frequently pays back through higher yield, fewer field returns, and faster ramp to stable production.
The EMS Tier Structure
Not all EMS providers are the same. The industry has a clear tier structure, and picking the wrong tier is the second most common mistake hardware startups make (after confusing CM with EMS in the first place).
| Tier | Revenue | Examples | Typical MOQ | Startup fit |
|---|---|---|---|---|
| Tier 1 | $1B+ | Foxconn, Flex, Jabil, Celestica | 50,000+ units | Not viable. They will not take your call at 5,000 units. |
| Tier 2 | $100M–$1B | Regional specialists | 10,000–50,000 units | Possible but you compete with larger clients for line time. |
| Tier 3 | $10M–$100M | Startup-friendly EMS | 1,000–10,000 units | Best fit. Set up for NPI at startup scale. Willing to invest in your ramp. |
| Tier 4 | <$10M | Small shops, regional | 100–5,000 units | Fast and flexible. Limited capacity. May not support your growth beyond 10K units. |
Most venture-backed hardware brands doing their first production run should target Tier 3 or Tier 4. The right provider at Tier 3 gives you EMS-level engineering at startup-compatible volumes, with a path to scale if the product succeeds.
Tier 1 and 2 become relevant after the product has proven retail demand and production volumes consistently exceed 10,000–50,000 units per run.
CM vs. EMS: The Decision Framework
Use this table to evaluate which model fits your current stage. The answer often changes as the product matures — many hardware brands start with a Tier 4 CM for prototyping, graduate to a Tier 3 EMS for first production, and eventually move to a Tier 2 for scale.
| Dimension | Contract Manufacturer | EMS Provider |
|---|---|---|
| Design support | None. Builds to your spec. | DFM review, layout optimization, component alternates |
| NPI process | Informal. First run is the production run. | Formal: pilot run, process validation, yield tracking, ramp plan |
| Component sourcing | You source, they assemble (consignment) or limited procurement | Full procurement with AVL management, allocation handling, alternate qualification |
| Test engineering | You provide test fixtures and procedures | They develop ICT, functional test, burn-in, and calibration |
| MOQ | Often lower (100–1,000 units at Tier 4) | Higher (1,000–5,000+ at Tier 3) due to NRE investment |
| Cost structure | Lower per-unit, higher risk of yield loss | 10–25% per-unit premium, offset by fewer field returns and faster ramp |
| IP risk | Lower — you control every file and specification | Moderate — they see your design internals during DFM review. NDA essential. |
| Scalability | Limited by their equipment and capacity | Built for ramp — Tier 3 EMS can typically 3–5x volume within existing capacity |
The pattern: Most hardware startups start with a CM because it is cheaper and faster for the first run. They discover they need an EMS provider when quality becomes inconsistent, component sourcing becomes a bottleneck, or retailer compliance requires manufacturing traceability the CM cannot provide.
Five Signs You Need to Graduate from Your Current Manufacturer
These signals usually appear between the second and fifth production run. By then, the prototype-stage manufacturer has served its purpose and the product needs production-stage manufacturing.
1. Quality is inconsistent across runs
The first run was acceptable. The third run had solder defects. The fifth run had component placement errors. This is not a quality problem — it is a process control problem. CMs without formal process documentation tend to rely on operator skill, which varies. An EMS provider documents the process, tracks defects by station, and implements corrective action when yield drops.
2. They cannot source the components you need
Semiconductor lead times run 12–26 weeks. Display panels, specific connectors, and custom ICs can run longer. If your CM expects you to supply components on consignment and you are spending 20 hours a month managing allocation, alternate qualification, and supplier negotiations, the manufacturing relationship is absorbing engineering time that should be spent on product development.
3. No test engineering capability
Your prototype was tested manually. At production volumes, every unit needs functional test, and many need in-circuit test, firmware verification, and calibration. If your manufacturer does not have test engineering capability, you are either shipping untested product or hiring a test engineer to build fixtures — neither of which scales.
4. They are not set up for your target volume
A shop that runs 500-unit batches efficiently may not be able to run 5,000-unit batches without bottlenecks, overtime, and quality compromises. Ask your current manufacturer: at 3x your current volume, what changes in their process? If they do not have a clear answer, they have not planned for your growth.
5. They cannot support retailer compliance
Retail channels — Best Buy, Amazon, Target — require manufacturing traceability, lot tracking, specific labeling, and in some cases factory audits. Many small CMs do not have the documentation systems, certifications, or audit history to pass a retailer compliance review. If you are targeting retail within 12 months, verify this now — not when the buyer asks for the factory audit report. See our Retail Readiness Scorecard for the full compliance checklist.
The ODM Option: When You Do Not Own the Design
An ODM (Original Design Manufacturer) owns the product design. You buy their product and apply your brand, your packaging, and your marketing. The ODM handles all of the manufacturing, testing, and component sourcing because it is their product.
ODM works when:
- You want to enter a product category without investing in product development
- The product is relatively commoditized (generic Bluetooth speakers, basic IoT sensors, standard power banks)
- Differentiation comes from brand, software, or channel strategy — not hardware design
ODM does not work when:
- You have proprietary technology, firmware, or industrial design
- Differentiation is in the hardware itself
- You need control over the BOM, component quality, or manufacturing process
- You plan to file patents on the product design
If you have raised venture capital for a hardware product, you almost certainly have proprietary technology that makes ODM inappropriate. The decision is between CM and EMS — not ODM.
How to Run the Transition Without Disrupting Product Availability
The worst time to discover your manufacturer cannot support your next phase is during a retail launch. The transition from one manufacturing partner to another should be planned, not reactive.
The sequence:
- Qualify the new partner while the current one is still producing. Never single-source your transition. Run the evaluation, NRE, and pilot run in parallel with ongoing production at the current manufacturer.
- Run a full pilot with 100% inspection. The pilot run proves the new partner's process, not just their proposal. Every unit in the pilot should be inspected against your production specification.
- Bridge inventory. Build 60–90 days of safety stock before the cutover. This covers the gap between last shipment from the old manufacturer and first production release from the new one.
- Transfer documentation. BOM, Gerbers, test procedures, firmware, packaging specs, supplier contacts, and any process notes from the current manufacturer. If this documentation does not exist, create it before the transition — not during.
- Cut over and monitor. First 3 production runs at the new partner get enhanced QC. Track yield, defect categories, and cycle time. Compare against the baseline from the old manufacturer.
Proof point: When Logic managed a supplier transition for Audio Enhancement, every unit in the pilot run was individually inspected. 100% pilot-run QC, zero customer-facing disruption, and 2x manufacturing capacity after the transition. The transition ran from qualification to cutover without a single customer-facing disruption.
For the full operational timeline from prototype to production scale, see our Prototype to Production guide. For the broader operational model, see our consumer electronics industry page.
How Logic Runs Manufacturing Transitions
Logic manages the operational bridge between manufacturing partners. For funded hardware brands:
- Manufacturer evaluation. We assess production capability, certifications, component sourcing networks, test engineering capacity, NPI process maturity, and retailer compliance readiness.
- Pilot-run QC. 100% inspection on the first production run from any new manufacturing partner. Defect categorization, yield tracking, and corrective action before scale.
- Transition management. Parallel qualification, inventory bridging, documentation transfer, and cutover monitoring. The goal is zero customer-facing disruption.
- Packaging coordination. Retail packaging design, structural engineering, and compliance documentation run in parallel with manufacturer qualification. See our packaging sourcing guide for the full framework.
- Ongoing production oversight. Run-to-run quality monitoring, supplier scorecards, and escalation management on a monthly retainer.
Retainers for hardware brands start at $5,000–$7,000/month for active management. Advisory engagements start at $2,500/month. All month-to-month, no annual contract. For a full cost comparison against hiring a VP of Operations, see the Logic vs. VP Ops guide.
For working capital planning around production transitions, see the hardware section of our Working Capital Playbook.
