Metered billing: how it works and where it goes wrong

What is metered billing?

Metered billing charges customers for what they actually used during a billing period, such as API calls, messages sent or gigabytes stored, rather than only a fixed subscription fee.

How metered billing works

  • Every billable event is recorded as it happens, with the customer, the quantity and the time.
  • At the end of the period, the usage records are added up per customer.
  • The price plan is applied: per unit, tiered, or an included quota plus overage.
  • An invoice is produced, with a breakdown the customer can check.

A metered billing example

A plan costs $20 a month and includes 50,000 calls, with overage at $0.80 per thousand. In March the customer makes 90,000 calls: $20 plus 40 × $0.80, a total of $52. In April they make 30,000 calls and pay just $20.

Metered vs subscription billing

A flat subscription is simple and predictable, but customers who use a little subsidise those who use a lot. Metered billing ties price to value and lets small customers start cheaply, at the cost of less predictable bills. Many products combine them: a base subscription with included usage, then metered overage.

Metered billing challenges

  • Counting accurately under load, without double-counting retries.
  • Deciding what is billable: failed calls usually should not be.
  • Making bills explainable, so disputes can be settled from the records.
  • Warning customers before usage turns into a surprise invoice.

Micro SaaS API Marketplace counts every call per key, records failed and rate-limited calls without billing them, and re-derives each payout statement from that log so it can always be checked.

Last reviewed 2026-11-01.

All glossary terms →