Build, buy, or borrow — role by role, not by policy

For a long time the sourcing model was a centre-level stance: this GCC builds in-house, that one leans on partners. What we’re watching now is that stance dissolve into a per-role judgment. Within the same team, one capability is built permanently, the next is bought as a ready-made pod, the one beside it is borrowed through a contract specialist for as long as it’s needed. The unit of decision has moved from the org to the seat.

The reason is that the three options answer different questions. Building owns the muscle but costs time and commitment. Buying skips the learning curve but you inherit someone else’s way of working. Borrowing gets you moving today but leaves nothing behind when it ends. No single policy is right across a whole charter, because different roles genuinely want different answers — and pretending otherwise leaves speed or ownership on the table.

The question that decides it is simple to ask and honest to answer: is this capability core to what the centre exists to own, and do you intend to still own it in three years? If yes, build it — hire permanent, absorb the ramp, keep the muscle inside. Renting your core hollows out the reason the centre was chartered, however tempting the speed. This is especially true in ER&D, where the deep engineering judgment is usually the whole point of the centre and the thing least sensible to borrow.

If the capability is real but not core, buy or borrow with intent. Buy when a mature, well-defined function is worth acquiring whole to skip the curve; borrow when the need is scarce, spiky, or temporary and there’s no reason to carry it permanently. The discipline is to decide each role against that test deliberately, rather than defaulting the whole centre to one habit and discovering later you built what you should have borrowed — or rented what you needed to own.