↓ Download as PDF

Most companies outsource manufacturing for a reason that is entirely sound. Running your own factory means investing in equipment, people and process technology, and then re-investing to keep them current. That is a large fixed cost and a large piece of inflexibility. In a volatile industry — semiconductor equipment is the one I know best — the ability to scale up and down quickly is not a nice-to-have. It is an existential skill, and a contract manufacturer is often the only way to get it.

So the decision to farm out production is usually right. What follows the decision is where the trouble starts.

Because one morning you realise something has quietly inverted. You are the customer. You are paying for the product. And yet the company you are paying now controls your ability to stay in business. You are caught between your own customer and your factory, with less visibility than either of them, and the leverage you thought the relationship gave you has drained away somewhere you did not notice.

This is not a rare failure. It is a well-travelled road. Which means the pitfalls along it are known, and avoidable, if you go in understanding one thing: outsourcing execution is not the same as abdicating responsibility for your product. They are your factory now. You still own what comes out of it.

The contract is not the control system

The first thing people reach for is the contract. Get the MSA right, the thinking goes, and the relationship is safe.

The contract is just the beginning. It sets the terms; it does not run the relationship. And here is the hard truth underneath it: once production starts, the power shifts — especially if you have a single source. If it ever gets to the point where either party is reaching for their lawyers, it is going to end poorly for both of you, whatever the contract says.

A contract is a floor. It is what you stand on when everything else has failed. What you actually manage the relationship with is something the contract cannot contain: high-level relationships with people you trust, built before you need them. Stuff will go wrong — that is not pessimism, it is planning. When it does, you want to be sitting across the table from a business person who has as much interest in the outcome as you do, not an administrator and not a lawyer. Talk about the hard things early: forecasting risk, inventory risk, what happens in a pandemic, what happens when the next Lehman Brothers goes under, what happens when a container falls off a ship or a product has to be recalled. Have those conversations while everyone is still friends.

They are your factory. Act like it.

If you take one idea from this, take that one. Everything else follows from it.

Right-size the relationship before you start the courtship. The core of every outsourcing relationship is the amount of money at stake. Do you want to be a big fish in a small pond, or a small fish in a big one?

Own the product even though you have outsourced its making. Own the full bill of materials, the component choices and the supplier choices. Own the data — insist on a data definition designed to document product and production, structured so you can actually get it out. Understand their supply chain, because it is your supply chain now; they can manage the purchasing, but you should own the supply chain management. And understand that the discussions, the workflows, the decisions and the production results are part of your intellectual property. Maintain visibility at a reasonable level of detail as the product is built, and set that visibility up from the start — because your manufacturing partner may reasonably regard detailed production data as their proprietary information, and that is a much harder argument to win after the fact than before.

Document the product until it is impossible to get wrong. Getting production right when it is halfway around the world means documenting your product — really documenting it. Down to every nut and bolt. If the CM supplies the French-language operating manual, that manual is a line on the bill of materials. Get the BOM complete, and make it impossible to build from an old revision. Keep a historical record of revisions so you can trace everything — connectors, tooling, fixtures, verification and validation. Be very clear about which items cannot be substituted without your approval, spell out on the BOM which parts need serial-number tracing, which can be lot-traced and which the CM may substitute at will, and have validation plans so nobody quietly bends the rules. What is obvious to you is not obvious to a manufacturer who did not live through your development and prototyping. Assume nothing is understood that is not written down.

Forecasting is how you keep your credibility

Forecasts are uncertain — it is hard to make predictions, especially about the future. But a forecast does two things that matter enormously, and neither is about being right.

It gives the contract manufacturer visibility of the revenue and profit they can make from your business, and that drives the caliber of the resources they assign to you and their level of commitment to your account. And it defines the risk they will be exposed to, which in turn is the risk they will expose you to. A forecast is not a promise about the future. It is the foundation of your credibility with the people building your product. Once a CM loses trust in your ability to forecast, you will work much harder to rebuild it than you ever spent getting it right the first time.

Make forecasting a collaborative, cross-functional process with your manufacturing partner — not a number you throw over the wall.

Communication is a policy, not a friendship

Communicating well is the single hardest part of working with a contract manufacturer, and the mistake is to treat it as a matter of goodwill. Goodwill helps. But you need policies, not just relationships. Engineering change is the obvious example: agree, implement and enforce a disciplined change process, because uncontrolled revisions during a new product introduction are where control quietly leaks away.

And have someone you trust on the ground, early and often. Two lines I have earned the hard way:

If your relationship runs on late-night escalations, you do not have a relationship. You have a recurring emergency.

Plan for the divorce while you are still friends

Two risks sit at the far end of the road, and both are cheapest to handle at the start, when nobody thinks you will need to.

The first is disintermediation — the real possibility that your contract manufacturer becomes your competitor. Ask IBM how defining, publishing and then losing control of the PC interface worked out for them long-term. Have an intellectual-property strategy against it before you need one: a component only you fit at commissioning; a software element that needs a key; splitting the product so no single CM has all of it; retaining ownership of the tooling and fixtures. And define your exit path early — write the operational details of leaving the relationship into the contract, while leaving is the last thing on anyone’s mind.

The second is inventory. When things go wrong and the divorce lawyers start circulating, it always — always — comes down to who paid for what, and what it is worth now. Decide who carries the inventory risk early. The CM will, in hindsight, have spent money in ways that look unwise. You will have provided a forecast, and forecasts are always wrong, especially for highly configured products with a lot of options — so they will blame your forecast when things slip. Be prepared for that. And if you supply consigned inventory, understand that it will occasionally cause a delay, and that the delay will always be a convenient explanation — especially the moment you start claiming contractual redress. None of this is cynicism. It is knowing where the argument goes before you are having it.

And have more than one

Plan to have more than one contract manufacturer. The complexity of managing suppliers rises with roughly the square of their number, so this is not free — but a single source is a single point of failure on the one thing that keeps you in business. Frame the second source as resilience, not as a threat, and build it in before you are desperate for it.

---

Handled well, none of this is a reason to keep manufacturing in-house. The opposite: awareness of the common pitfalls is exactly what lets you turn outsourced manufacturing into a competitive advantage — the scale and flexibility without the fixed cost, and without losing the product.

But notice what all of it has in common. Right-sizing, forecasting discipline, BOM control, data ownership, change process, IP strategy, exit terms, second-source — none of these is a clause you sign once. Each is a structure you have to stand up and run, inside a live relationship, usually while the ramp is already happening and the pressure is already on. Knowing they matter is the easy half. Building them, in the right order, on the access you actually have and against a schedule that will not wait, is the work.

That is the work I do.