Sooner or later every technology leader faces the same decision about a function: do we build and run this ourselves, or hand it to a partner? The debate usually collapses into a spreadsheet argument about cost per head — which is exactly the wrong place to start.
Here's a cleaner way to decide.
Ask what should define you
Some work is your competitive edge — the systems and capabilities that make you you. That work belongs in-house, owned by people who live your mission. Other work is essential but undifferentiated: it has to run well, but doing it yourself wins you nothing. That's the work a managed-services partner is built for.
Sort your functions along that line before you touch a cost model. It clarifies most of the decision on its own.
Count the real cost of in-house
The salary is the visible cost. The hidden ones — recruiting, management attention, coverage for holidays and departures, the opportunity cost of your leaders running an operation instead of building — are where in-house quietly gets expensive. A fair comparison includes all of it.
Judge a partner on outcomes, not bodies
The failure mode of managed services is renting warm bodies and calling it a solution. A good partner owns an outcome — a service level, a delivery, a running capability — and is accountable for it, freeing your leaders from the day-to-day entirely. If the arrangement still needs you to manage it closely, it isn't managed.
You don't have to choose once, or for everything
This isn't all-or-nothing. Many teams keep the differentiating core in-house and hand the steady, well-understood operations to a partner — then adjust the line as the business changes. The right answer is a portfolio, not a doctrine.
Build-versus-outsource stops being hard the moment you ask the right question. Own what defines you; let a trusted partner run what simply needs to run well.
If you're weighing where a function belongs, let's think it through.