A depth map throws away colour, texture and identity and keeps only spatial arrangement. Conditioning on one gives you the same room, the same camera position and the same object placement across an entire set, with the surface treatment free to change.
It is the cheapest way to hold a composition across variants. Nine ads that need identical staging and different products are one depth map and nine generations, not nine compositions negotiated one at a time.
What it does not hold is identity. A depth map will keep a bottle in the same place and will not keep it the same bottle, which is why depth conditioning and a master plate are complementary rather than alternatives.
What does a depth map control?
Spatial arrangement: where things are relative to the camera and to each other. It does not control colour, texture or identity, so it holds staging without holding the subject.
When is depth conditioning the right tool?
When several assets need identical staging with different content, or when a client has approved a composition and everything else is still moving.