Why Products Fail in Production — And How to Catch Problems Earlier

By the time a product fails in production, the failure usually started somewhere much earlier. A material assumption that was never checked. A tolerance that seemed fine in isolation and caused problems in assembly. A supplier selected for price rather than fit. A change made during sampling that wasn't formally documented.

Production failure is rarely sudden. It's the accumulation of small gaps — between design and engineering, between sample and specification, between what was agreed and what was understood.

Here's where those gaps most commonly appear, and what can be done about them.

The specification exists in someone's head

Design files communicate shape and proportion. They rarely communicate everything a factory needs to actually manufacture a product — surface finish standards, assembly tolerances, material grades, acceptable variation between units.

When that information lives in the designer's head rather than in a written specification, every factory interaction becomes a negotiation over unstated assumptions. What "natural" means for a wood finish. What "clean" means for a weld. These are real categories with real variation, and without explicit standards, a factory will make reasonable decisions that may not match what you had in mind.

Writing a proper manufacturing specification before engaging suppliers is tedious. It's also one of the most valuable things you can do, because it forces the conversation about ambiguity to happen before production, when it's cheap, rather than after, when it isn't.

The sample passed, but the brief changed

Sample approval is supposed to be the moment when design intent becomes manufacturing standard. In practice, it's often a moving target.

A designer sees the first sample and requests modifications. Those modifications are communicated informally — a photo with annotations, a voice message, an email thread. The factory implements what they understood. A second sample is produced. More modifications. By the time production starts, the specification has drifted from the original, and no one has a complete written record of what was agreed.

This isn't dishonesty on anyone's part. It's a documentation failure. Every change during development should be captured in a formal revision to the specification. Without that, the production run is built on a version of the product that exists only in an email chain.

The wrong factory for the job

Some production failures aren't about communication or documentation. They're about capability mismatch — a factory that accepted an order they weren't quite equipped to handle.

This is more common than it should be, partly because factories don't always know where their limits are until they've hit them, and partly because the pressure to take on orders is real. A factory that's excellent at high-volume stamped metal components may genuinely struggle with a lower-volume product that requires close tolerances and careful hand-finishing. The skill sets overlap but aren't the same.

Identifying the right factory before committing to production — not just a factory that says yes — is the step that prevents this category of failure.

Quality control happens at the end

In many production setups, quality control means inspection at the end of the run. Units are checked after they're made. Problems found at this stage — a dimension that's consistently off, a finish that's uneven across the batch — are expensive to fix. Rework is slow. Rejection means delay.

The alternative is building quality checkpoints into the process — checking materials before production begins, reviewing the first units off the line, inspecting at defined intervals during the run rather than only at the end. This approach surfaces problems when they're still cheap to address.

It requires more involvement in the production process, which is exactly why it tends not to happen when a brand is managing a factory relationship from a distance without someone on the ground.

What earlier actually looks like

Catching problems earlier means doing the harder work at the beginning: writing explicit specifications, documenting every change during sampling, selecting suppliers against specific criteria rather than availability, and building quality checkpoints into the production plan rather than hoping for the best at the end.

None of this is complicated. Most of it is simply the work that gets skipped when schedules are tight and budgets are under pressure.

933.work coordinates product manufacturing in China, with structured documentation and quality checkpoints built into every project. If you're heading into production and want a clearer process, we'd be glad to hear about it.