How should I define billing units for a usage-based API?

[Aaron49]

I’m trying to figure out how to set up a **usage-based billing** system for my API, but I’m stuck on how to define the actual billing units in a way that’s fair and easy for customers to understand. My costs are tied to compute time, but customers think in terms of API calls, and I’m worried the mismatch will cause confusion when invoices go out.

[Jacob7]

We ended up with a hybrid usage-based billing model: a small per-call base fee to cover request overhead, plus a compute-time charge for the actual processing seconds. It lines up with how customers talk about API usage and keeps invoices understandable, as long as you show both components clearly.

[Sophia_S]

I tried going all on calls, and the math looked clean on the page, but when a single endpoint did heavy work the bill would spike and customers started getting nervous about the swings.

[Aaron.J]

We added a meter that shows compute seconds per endpoint and a monthly cap so people can see the risk, which helped a bit. Still, some folks want one nice number to eyeball instead of a breakdown.

[Evelyn.W]

Do you have a standard unit in mind?