Metering
MEE-ter-ing
Tracking and measuring customer usage in real time for billing purposes. The plumbing that makes usage-based pricing work.
Metering is the system that counts what customers use. API calls, compute seconds, storage bytes, messages sent, tokens processed. Every usage-based pricing model depends on accurate metering. Get it wrong and you either undercharge (losing revenue) or overcharge (losing trust).
Metering sounds simple but the engineering is hard. You need real-time or near-real-time tracking, aggregation across distributed systems, idempotent event processing (so retries do not double-count), and a billing pipeline that converts raw usage into invoices. Companies like Amberflo, Orb, and Metronome exist specifically to solve metering.
The metering system also powers the customer-facing usage dashboard. Customers need to see their consumption in real time so they can manage costs and avoid surprises. If your usage dashboard lags by 24 hours, customers cannot react to spikes. That leads to surprise invoices and angry support tickets.
Examples
An API company meters requests.
Stripe logs every API call with a timestamp, endpoint, and response code. At the end of each billing cycle, the metering system counts successful API calls per account, applies the pricing rules, and generates an invoice. If a webhook fires twice due to a retry, idempotency ensures the call is counted once.
A compute platform meters by the second.
AWS Lambda meters in 1-millisecond increments. A function that runs for 150ms is billed for 150ms, not rounded up to the nearest second. This granularity required rebuilding their metering pipeline from seconds to milliseconds. The change saved customers 30-50% on compute costs.
Metering powers a real-time usage dashboard.
Vercel shows bandwidth consumption in real time on the project dashboard. A developer deploys a new feature and watches bandwidth spike. They see they are approaching their plan limit and add caching before the overage kicks in. Real-time metering turned a surprise bill into a proactive optimization.
In practice
Read more on the blog
Frequently asked questions
Should I build or buy metering?
Buy. Metering is infrastructure, not a differentiator. Companies like Amberflo, Orb, and Metronome have solved the hard problems: real-time aggregation, idempotent processing, and billing integration. Building metering in-house typically takes 2-3 engineering months and requires ongoing maintenance. Buy it and focus on your product.
How granular should metering be?
As granular as your pricing requires. If you charge per API call, meter per call. If you charge per GB, meter per byte and aggregate. If you charge per hour, meter per second and round up. The golden rule: meter more granularly than you bill, so you have flexibility to change pricing without rebuilding metering.
Related terms
Charging customers based on how much they consume. Pay for what you use. The model behind Snowflake, Twilio, and AWS.
A pricing structure where revenue scales directly with customer usage. Broader than usage-based pricing, it includes credits and commitments.
Charges that kick in when a customer exceeds their plan's included usage limits. The bill that surprises people.
A virtual currency customers buy upfront and spend as they use your product. The bridge between subscriptions and pure usage-based pricing.
Want the complete playbook?
Picks and Shovels is the definitive guide to developer marketing. Amazon #1 bestseller with practical strategies from 30 years of marketing to developers.