A platform must decide whether to build or buy a capability. What is the primary test, and what is the pattern that usually works?

The primary test

Would you accept your closest competitor using the identical implementation?

If yes, it is not differentiating, and building means spending engineering on a problem others have solved and that no customer chooses you for. Buy identity, email delivery, payments processing, observability, most storage and delivery.

If no, building may be justified — but only if you can build it better, which is a separate and much harder claim than "it matters to us".

The signal that forces building

When the objective function is specific to your business model. A generic service optimises a generic objective; if what you need is a blend of factors only your business understands, no vendor has that objective — and the mismatch is not addressable by configuration.

The pattern that usually works

Buy the surrounding infrastructure; build the differentiating layer.

For a media platform: buy object storage, content delivery, compute and identity; build the rendering pipeline behaviour, the asset transfer layer and the collaboration model. For a marketplace: buy mapping and payment processing; build the matching and pricing logic.

This concentrates engineering where it differentiates and buys the parts where scale economics are unmatchable.

The secondary considerations

• Speed to first value. A common and sound path is to buy first and build later once the requirements are understood — and understanding them is frequently the main output of the buying period.
• The bought solution's ceiling, and whether you will reach it.
• Whether the capability improves with your data, since a vendor's model may improve from your usage while yours does not.
• Total cost including permanent operation, not just the build.
• Dependency risk, which is real and usually overstated relative to the cost of building.

The failure in each direction

Building everything spreads a limited team across problems others have solved. Buying everything means the differentiating capability is identical to competitors' and improves on a vendor's roadmap rather than yours.