Send a drawing to a competent Chinese factory and it will come back marked up. Wall thicknesses adjusted, corners rounded, a fastener swapped, a tolerance relaxed. This is DFM — design for manufacturing — and it is a good sign, not a bad one. But inside the same list of revisions sit two very different kinds of change: those that make your product possible to produce reliably, and those that make it cheaper to produce while quietly making it worse. This article is for companies putting their own product into production in China, from around 1,000 units per order, and explains how to read that list.
What DFM actually is
DFM is the adaptation of a design to the process that will manufacture it. A drawing that is geometrically valid can still be unmanufacturable, or manufacturable only with high scrap rates. The factory's engineers know their machines, their moulds and their material behaviour better than any external designer — which is why the DFM round belongs before tooling is cut, at the stage described in how manufacturing in China actually works.
Skipping DFM does not avoid the changes. It just moves them to after the tooling exists, where every modification is paid for in steel.
Changes worth accepting
- Uniform wall thickness. Thick sections cool unevenly and produce sink marks and warping. Ribs deliver the same strength without the mass.
- Draft angles. A part with no draft will not release cleanly from the mould. This is physics, not preference.
- Radii instead of sharp internal corners. Sharp corners concentrate stress and crack in service; they also shorten mould life.
- Fewer separate parts. Combining two moulded pieces into one usually improves both cost and reliability.
- Standard fasteners and standard components. A locally available standard screw is cheaper, more reliable and replaceable; an exotic one becomes a supply problem on every reorder.
- Tolerances relaxed where nothing depends on them. Precision costs money on every single unit produced; pay for it only where it does something.
- Gate and parting-line placement. The factory knows where the marks will be. Choosing where they show is better than discovering it.
Accepting these usually improves the product. Refusing them out of attachment to the original CAD file is how importers pay a premium for a higher defect rate.
Changes that cheapen the product, not the process
- Thinner walls than the specification, beyond what structural logic requires.
- A different resin or "equivalent material", including recycled content, where the grade was specified for a reason.
- Thinner plating or coating — the microns you cannot see are the ones that decide corrosion resistance.
- A cheaper component brand for motors, cells, bearings or electronics, presented as functionally identical.
- Removing a metal insert, reinforcement or seal that "is not necessary."
- Reducing test requirements because a test "takes too long."
None of these are automatically dishonest — sometimes they are genuinely sound. But each one moves the product, not the process, and each must be an explicit decision by you rather than a line item you nodded through. This is the same substitution mechanism that later produces a fine sample and a defective batch.
Three questions that separate the two
1. Does it change the function, strength, lifetime or appearance the customer sees? If yes, it is a product change, not a manufacturing change. 2. Does it change a material, grade, component or coating named in the specification? If yes, it needs written approval, no exceptions. What belongs in that specification is in the technical spec sheet for a Chinese factory. 3. Who captures the saving? A genuine DFM change lowers the quoted price or the defect rate. If the change saves cost and the price stays the same, ask why.
What a proper DFM report contains
- Marked-up drawings showing each proposed change, not a text list
- The reason for each change: mouldability, tool life, cycle time, assembly, cost
- The cost impact per change — unit price and tooling both
- The effect on lead time
- A clear split between "required to manufacture" and "suggested to reduce cost"
- A sign-off line, so approval is documented and dated
Ask for it in this form. A factory that cannot produce a structured DFM report is telling you how it will handle the rest of the project — which is why the DFM round doubles as an engineering capability check, alongside the checks in our factory verification protocol.
Freeze the design before the tooling
Once steel is cut, every change costs money and weeks: mould modification, re-sampling, sometimes a new insert. Set an explicit design freeze at the end of the DFM round, and state in the contract who pays for changes after that point — yours if you change your mind, theirs if the tooling does not meet the approved drawing. The commercial side of this belongs in tooling and moulds: cost and ownership and in what a contract with a Chinese factory must contain.
A factory that sends back no DFM comments is not agreeing with your design. It is telling you that it will build exactly what you drew, ship exactly what comes out, and hand you the defect rate as your problem.
Reviewing DFM with someone on your side
We review DFM reports on our clients' behalf: separating manufacturability from cost-cutting, checking that nothing in the specification changed silently, pricing each revision, and locking the design before tooling. Seven years, 3,500+ deliveries and 340+ verified suppliers of doing this — real projects are in our case studies.
Start with product sourcing or private label manufacturing if the design is already yours.
