every director who calls us wanting to “outsource the technology” is really trying to outsource a pain. And pain can't be outsourced — it has to be diagnosed. When the vendor gets the pain without understanding the business behind it, they hand back code. They rarely hand back a solution.
IT outsourcing works. Access to rare talent, speed, variable cost instead of fixed — all real. The problem was never the model. It was treating the external team as a task machine instead of people who need to understand what's at stake.
§ 01 / ThesisExecution is outsourced. Business judgment is not.
There's a clear line. On this side of it: what to do, which KPI to move, the real operation, what can break. That's yours — and you can't delegate it outside without losing control of your own business. On the other side: the engineering, the speed, the technical skill to turn the decision into a system that runs.
Whoever confuses the two outsources the decision along with the execution — and the vendor ends up deciding, by omission, things that should belong to the owner. The right partner executes with technical autonomy and zero authority over the business. They ask “why” before “how,” and hand back a reasoned no when the request doesn't add up.
§ 02 / When it makes senseThree situations where the external team wins.
Not every problem calls for an external team. But three almost always do:
- Rare, one-off competence. You need someone who has done it ten times — not to hire, train, and then have nothing for them to do.
- Speed against a window. There's a market deadline, and building an internal team from scratch doesn't fit the calendar.
- A spike that won't become structure. The demand is real now but doesn't justify permanent headcount. Models like EOR exist precisely to hire talent — even abroad — without setting up a legal entity and without losing compliance.
In all three, the point isn't “cheaper.” It's the right capacity, at the right time, without the liability of maintaining what you won't use later.
§ 03 / The real riskOutsourcing can't become dependency.
The classic failure of outsourcing isn't the external team delivering badly. It's them delivering well and walking away with the knowledge. Six months later, no one in the house knows why the system works the way it does — and you're locked to the vendor not by contract, but by ignorance.
A good partner leaves the house more capable. A bad one leaves it more dependent.
That's why knowledge transfer isn't an end-of-project deliverable. It's continuous practice: an internal owner named from day one, documentation that lives next to the code, decisions recorded with their why. Nothing we deliver should need us to keep running.
§ 04 / ClosingThe vendor you want is the one who thinks like a partner.
The market is full of people who bill by the hour and deliver what they're told. Cheap on the budget, expensive in the operation. What moves the needle for a mid-to-large company is the opposite: someone who enters the conversation with business judgment, executes with seniority, and leaves the operation stronger than they found it.
Outsource the execution. Keep the problem — and keep whoever truly helps you solve it.
end · field note #48 · noûs / jun 26