What you described is likely usage-based, in the sense that the (presumably cloud) vendor usually has to reserve a certain amount of resources per user, “just in case”, because they don’t know / can’t predict your activity pattern. Same reason VMs are still charged for when started but idle: they reserve their CPUs.

What people really want, when they say “usage-based billing”, is outcome-based billing. They want to get charged money whenever they hit the button in your software that makes them money (or, for a cost-center, saves them money.)

Think of e.g. tax prep companies. (For the average Joe employee), they don’t charge you money up-front; instead, they take a part of the net-positive return they fully expect to find you. They make you happy, then take a slice of your happiness at the exact point that they’re making you happy. Outcome-based billing.