I Work FOR YOU, Not Factories.

Your prototype works. Your demo impressed investors. Your software runs flawlessly. And none of that matters if the factory building your product does not know who you are.

What Happens When Your Prototype Works But Your Factory Doesn’t Know You

By Leon Xu | Easelink Tech | Shenzhen, China

The Gap Nobody Warns You About

Here is a scenario I see repeatedly: an overseas hardware startup builds a working prototype. Maybe it is an AI wearable, a smart device, a robotics component. The prototype validates the technology. The team raises funding based on that validation. They send the design files to a factory in Shenzhen and wait for mass production to begin.

Then nothing goes right.

The first EVT build has problems the prototype never showed. The factory does not understand the firmware requirements. The BOM has components that are out of stock or discontinued. The DFM review — if it happens at all — reveals that the design cannot be manufactured at scale without significant changes. And the factory, which seemed responsive during the initial conversations, goes quiet for days at a time.

The team is confused. The prototype worked. Why is production so hard?

Because a working prototype proves your technology. It does not establish a manufacturing relationship. And without that relationship, you are just another email address in the factory’s inbox.

Why a Prototype Is Not a Production Handoff

A prototype is built by your engineering team — people who understand every decision behind the design, every tradeoff, every intentional choice. When something does not fit, they adjust. When a component behaves unexpectedly, they debug. When the firmware needs tweaking, they recompile.

Mass production is different. The people building your product did not design it. They are following instructions — a BOM, a Gerber file, an assembly drawing, a test procedure. If those instructions are incomplete, ambiguous, or wrong, the factory will either guess or stop. And when they guess, they guess based on their experience with other products — not yours.

This is where the gap becomes visible:

BOM gaps: Your prototype used specific components with specific tolerances. The factory’s BOM lists part numbers but not tolerances, temperature ratings, or acceptable substitutes. When a component goes out of stock, the factory picks an alternative — and it may not work.
DFM blindness: Your prototype was built on a CNC machine with 0.02mm tolerance. Mass production injection molding has 0.1mm tolerance. That 0.08mm difference can cause thermal deformation, assembly misalignment, or RF interference — none of which appeared in the prototype.
Firmware-hardware disconnect: Your engineering team wrote firmware for a specific hardware revision. The factory’s production line may produce slightly different hardware — different component batches, different PCB revisions — and the firmware may not handle those variations.
Test coverage: Your prototype was tested by engineers who knew what to look for. The factory’s test stations check what is written in the test plan — and if the test plan does not cover an edge case, that edge case ships to customers.

If you want to understand why prototypes lie about production readiness, I wrote about why hardware prototypes fail to predict mass production in detail. The short version: a prototype validates your idea. It does not validate your manufacturing plan.

What “Not Knowing You” Actually Means on the Factory Floor

When I say the factory does not know you, I do not mean they forgot your name. I mean there is no one on the factory floor who understands your product the way your engineering team does. And there is no one on your side who understands the factory’s constraints the way the production team does.

In Shenzhen, factories prioritize the customers they know. The ones who have been there for years. The ones who place repeat orders. The ones who have someone on the ground who shows up, reviews progress, and resolves issues in person.

A new overseas startup — even a well-funded one — is an unknown quantity. The factory will take the order. But when a production conflict arises — your order versus a larger customer’s order, your urgent change request versus a scheduled production run, your quality concern versus the factory’s shipping deadline — the factory will prioritize the relationship they trust.

This is not malicious. It is how every factory in the world works. But it means that without someone representing you on the ground, you are at the bottom of the priority stack — no matter how good your product is.

What I’ve Seen From the Production Side

I have been the person standing between an overseas startup and a Shenzhen factory. I have seen what happens when there is no bridge — and what changes when there is one.

One team sent their design to a factory and waited four weeks for an EVT build. When the units arrived, the speaker mesh was wrong, the button tactility was off, and the firmware could not connect to their app. The factory said: “We followed the drawings.” And they had — but the drawings did not specify mesh tension, button force curves, or BLE pairing procedures. No one had translated the prototype’s implicit requirements into explicit production specifications.

Another team had a BOM with 47 components. Twelve of those components were out of stock or had 20-week lead times. The factory knew this but did not tell the team — because the factory assumed the team would figure it out when production stalled. By the time the team found out, they had already committed to a launch date with their retail partners.

In both cases, the problem was not the factory. The problem was the gap between “my prototype works” and “the factory knows exactly what to build, how to test it, and who to call when something goes wrong.”

How Easelink Bridges the Gap

My role is to be the person the factory knows — on your behalf. Here is what that looks like in practice:

BOM validation before production: I review every line item in the BOM — checking availability, lead times, alternative sources, and compatibility. If a component is risky, I find substitutes before production starts, not after it stalls.
DFM review with the factory: I sit down with the factory’s engineering team and walk through the design — identifying tolerance issues, assembly challenges, and tooling risks. I translate your design intent into manufacturing instructions the factory can execute.
Test plan development: I work with your engineering team and the factory to define test stations, test parameters, and pass/fail criteria — so the factory knows exactly what “good” looks like for your product.
EVT/DVT/PVT management: I manage each validation build on the ground — reviewing units as they come off the line, debugging issues in person, and feeding real production data back to your engineering team. This is the process I described in my article about EVT/DVT/PVT for AI hardware.
Production priority management: Because I am physically present and have ongoing relationships with the factory, I can ensure your order gets attention — not just a promise, but actual line time and engineering resources.
Firmware-hardware coordination: I bridge the gap between your software team and the factory’s hardware team — making sure firmware revisions match hardware revisions, and that production-line variations do not break your product.

The Philosophy: Technology Validates, Relationships Execute

Overseas hardware teams bring real technical strength: product vision, AI capability, software engineering, market understanding. That strength is what makes the product worth building. But technology validates an idea. Execution requires relationships — with the factory, with the suppliers, with the people on the production floor who decide whether your product is built right.

China provides the manufacturing density and supply chain depth to execute at speed. But that execution does not happen automatically. It requires someone who is present, who understands both sides, and who represents your interests — not the factory’s.

The strongest hardware products come from combining overseas R&D with China-side execution that is actively managed. A working prototype is the start. A factory that knows you — through someone who represents you — is what turns that prototype into a shipping product.

Practical Takeaways

Do not assume a working prototype means production-ready. Validate your BOM, DFM, and test plan before sending anything to a factory.
Translate implicit requirements into explicit specifications. If the prototype’s button “feels right,” define what “right” means in grams of force, travel distance, and tactile response — or the factory will guess.
Check component availability before committing to a BOM. A part that was in stock during prototyping may have a 20-week lead time at production quantities.
Establish a point of contact at the factory who has engineering authority — not just a sales contact. If your only relationship is with the sales team, you have no relationship with production.
Have someone on the ground during EVT and DVT. These are the stages where prototype-to-production gaps become visible. If no one is there to catch them, they surface as customer complaints after shipping. If you want to understand why AI hardware startups specifically struggle here, I wrote about why AI hardware startups need Shenzhen execution.
Build the relationship before you need it. The factory will prioritize customers they trust. If you are new, you need someone who already has that trust — and who will lend it to your project.

Final Thoughts

A working prototype proves you have a product worth building. But the factory does not build your prototype — they build their interpretation of your instructions. If those instructions are incomplete, or if no one is there to clarify them, the result will not match your vision.

The gap between “my prototype works” and “my factory knows what to build” is where most hardware startups lose months and money. Bridging that gap is what I do.