Usage-based billing

Usage-based billing ties the invoice to consumption. The customer is charged for what they used in the period, priced per unit, in tiers, or against a committed volume.

It has become the default expectation in infrastructure and AI-adjacent products, where cost of service scales with usage and a flat fee either overcharges light users or loses money on heavy ones.

What it requires that flat billing does not

Usage billing is substantially harder to operate, and the difficulty is rarely in the pricing model itself:

• Metering — reliable, idempotent collection of usage events, including late arrivals and corrections.
• Rating — applying tier, volume or banded pricing to aggregated usage, with the tier boundaries evaluated over the correct window.
• Mid-period visibility — customers on consumption pricing expect to see what they are accruing before the invoice lands. Surprise invoices are the fastest route to a disputed bill.
• Revenue recognition — usage revenue is generally recognised as consumed, which means it is billed in arrears and recognised in the same period rather than deferred.

Pure usage is rare in B2B

Most B2B SaaS companies do not sell pure consumption pricing. They sell a hybrid: a recurring platform fee billed annually in advance, plus metered usage billed monthly in arrears, often with a minimum commitment and overage above it.

That structure gives the vendor predictable revenue and the customer predictable baseline cost — but it means one subscription produces two billing streams on two schedules with two recognition treatments, which is precisely where billing systems built for flat subscriptions run out of road.