
Usage-based billing: the complete guide to consumption pricing
The customer pays for what they use. You still have to measure it right.
Cloud, energy, telecoms, connected devices, artificial intelligence: more and more offers are no longer sold as flat fees but on actual consumption. The customer pays per gigabyte, per kilowatt-hour, per API request, per trip or per document processed.
Usage-based billing promises a price aligned with value: the customer starts small, the invoice grows with their activity, and the provider grows with them. It also lets you pass on costs that have become variable, especially with AI.
The shift is clear. According to the OpenView survey cited by TechCrunch, 61% of SaaS companies surveyed used some form of usage-based pricing in 2022, up from 34% in 2020.
But moving to usage is not just a commercial decision. It is a deep change to the billing chain: you must collect millions of events, link them to the right customer, price them under each contract and produce invoices the customer understands.
This guide is for product leaders, CFOs and CIOs considering the model, or who have launched it and want to make it reliable. It covers definitions, pricing models, industries, how usage-based billing works, the pitfalls, choosing a metric, the steps to switch and how to choose usage-based billing software.
Contents
- What is usage-based billing?
- Usage vs subscription vs hybrid
- The 7 usage-based pricing models
- Which industries use usage-based billing?
- Why usage-based billing is growing
- How usage-based billing works
- The 6 pitfalls of usage-based billing
- How to choose a usage metric
- Switching to usage-based billing in 7 steps
- How to choose usage-based billing software
- Frequently asked questions about usage-based billing
- Billing on usage starts with measuring right
What is usage-based billing?
Usage-based billing is a model in which customers pay according to their actual consumption of a service: minutes, kilowatt-hours, gigabytes, API requests, trips or transactions. The amount varies from one period to the next. It is also called consumption-based billing, metered billing or pay as you go.
The principle is old. Electricity, water and phone calls have been billed on consumption for decades.
What is new is its spread to industries that used to sell licenses or flat fees: software, cloud, connected devices, digital services and, now, artificial intelligence.
Two notions are often confused:
- usage-based pricing: the commercial decision to set a price per unit consumed,
- usage-based billing: the operational ability to measure that consumption, rate it and invoice it.
The first is decided in a meeting room. The second is built in systems. This guide focuses on the second; for pricing design, see our guide to pricing models.
Usage vs subscription vs hybrid
A subscription bills a fixed amount per period, whatever the usage. Usage-based billing bills actual consumption. A hybrid model combines a subscription base with a variable part above a threshold. Subscriptions maximize predictability, usage maximizes alignment with value, and hybrid tries to get the best of both.
| Criterion | Subscription (flat fee) | Usage-based billing | Hybrid model |
|---|---|---|---|
| What the customer pays | A fixed amount per period | Actual consumption | A fixed base plus usage above it |
| Revenue predictability | High | Low to medium | Good |
| Alignment with value | Low | High | Good |
| Barrier to entry | Medium | Low | Low to medium |
| Billing complexity | Low | High | High |
| Main risk | Over-billed customers who churn | Volatile revenue, disputed invoices | Offers that are hard to explain |
Subscriptions stay simple to sell and forecast. Their weakness shows when usage diverges: two customers pay the same price for very different consumption. Our guide to recurring billing covers this model.
Pure usage appeals to customers who want to start without commitment. It worries finance, which must steer variable revenue, and sometimes the customer, who fears bill shock.
Hybrid is often where companies land: a subscription that secures recurring revenue, a variable part that captures growth, tiers that reward volume.
The 7 usage-based pricing models
Usage-based billing relies on seven main pricing models: unit price, volume pricing, tiered pricing, overage, minimum commitment, cap and prepaid credits. They often combine in one offer, which is what makes invoice calculation complex.
The examples below are illustrative: they show the mechanics, not market prices.
| Model | Principle | Illustrative example |
|---|---|---|
| Unit price | A fixed price per unit consumed | $0.02 per API request |
| Volume pricing | The unit price drops for all units once total volume crosses a threshold | $0.02 per unit below 10,000; $0.015 per unit for all units above |
| Tiered pricing | Each band of consumption has its own price | First 1,000 units at $0.02, the next at $0.015 |
| Overage | A plan includes a volume; the excess is billed per unit | 50 GB included, then $2 per extra GB |
| Minimum commitment | A floor amount is billed even if usage is lower | $500 minimum per month |
| Cap | The invoice never exceeds a maximum | Usage billed, capped at $2,000 per month |
| Prepaid credits | The customer buys volume in advance and draws it down | 10,000 credits bought, decremented with usage |
The difference between volume and tiered pricing is one of the most misunderstood. With volume pricing, crossing a threshold lowers the price of every unit. With tiers, only the units in the new band get the new price. The same consumption can produce very different invoices.
On top of these come conditional discounts, per-event prices (an activation, an alert) and, in partner ecosystems, revenue sharing.
Which industries use usage-based billing?
Usage-based billing is historic in telecoms and energy and is spreading to cloud, SaaS, artificial intelligence, IoT and mobility. In every industry, it starts with a unit of value the customer understands: the minute, the kilowatt-hour, the gigabyte, the request, the connected device or the trip.
| Industry | Typical usage unit | What to watch |
|---|---|---|
| Telecoms | Minute, text, gigabyte, session | Massive event volumes, bundles, roaming |
| Energy and utilities | Kilowatt-hour, cubic meter, capacity | Grid operator data, missing or estimated readings |
| Cloud and infrastructure | Compute hour, GB stored, bandwidth | Reserved capacity plus usage beyond |
| SaaS | Active user, transaction, document processed | Gradual move from licenses to hybrid |
| Artificial intelligence | Request, token, task completed | Variable costs to pass on, hard-to-predict usage |
| IoT and M2M | Active device, message, data | Fleets of thousands of devices, partner resale |
| Mobility | Trip, mile, rental minute | Fragmented usage, many sales channels |
Telecoms and energy are far ahead: their systems process event volumes few other industries know. Software vendors are discovering these constraints as they move to consumption. See our guides to telecom billing, IoT billing and AI billing.
The open source project behind Opencell was born in telecoms in 2007. Our references now include operators such as Eir, energy suppliers such as Engie and cloud providers such as OVHcloud.
Why usage-based billing is growing
Usage-based billing is growing for three reasons: it lowers the customer's barrier to entry, it lets revenue grow with usage without renegotiation, and it passes on costs that have become variable, especially with AI. Yet it rarely replaces subscriptions entirely.
Customers want to pay for value, not promises
A customer hesitates to commit to a large flat fee for a service they have not tried. A usage price lets them start small and scale as they see value.
Revenue grows with the customer
With a flat fee, charging more to a customer who uses more requires a renegotiation. With usage, the invoice follows the customer's activity. Revenue expands naturally with successful accounts.
Costs have become variable
When every request processed by an AI model, every gigabyte stored or every kilowatt-hour bought has a cost, unlimited offers become risky for the provider. Usage-based pricing keeps margin aligned with consumption.
But usage does not replace everything
As the TechCrunch headline puts it, usage is rising without replacing other models. In practice, most companies combine it with a subscription.
How usage-based billing works
Usage-based billing follows a six-step chain: a usage event is produced, collected, cleaned and linked to the right customer (mediation), priced under the contract (rating), aggregated into an invoice, then checked before release. Invoice accuracy depends on every step.
1. The usage event
Everything starts with a measurable fact: a call made, a request received, a meter read, a trip ended. Events are produced by network equipment, meters, applications or devices, often in large numbers.
2. Collection
Events flow to the billing system, continuously or in batches. Collection must be complete (no missing feed) and on time (no data arriving after the period closes).
3. Mediation
Mediation turns raw data into billable data: removing duplicates, formatting, checking consistency, linking each event to the right customer and contract. See our guide to billing mediation.
4. Rating
Each event gets a price under the customer's contract: unit price, tier reached, included volume, discount. This is the heart of the billing engine.
5. Invoicing
Rated events are aggregated with subscription fees, discounts and taxes to produce the invoice. To be accepted, it must let the customer understand what they are paying for.
6. Control
Before release, invoices are checked: unusual changes versus previous periods, zero or excessive consumption, customers with no usage despite an active service.
These steps sit within a broader process that runs from quote to cash collection, detailed in our quote-to-cash guide.
The 6 pitfalls of usage-based billing
The six most common pitfalls are incomplete or late usage feeds, events linked to the wrong customer, hard-coded pricing rules, invoices customers cannot understand, hard-to-forecast revenue, and data volumes the system cannot handle. Each has a technical or contractual answer.
| Pitfall | Consequence | Answer |
|---|---|---|
| Incomplete or late feeds | Usage not billed, corrective invoices | Completeness checks before rating |
| Events linked to the wrong customer | Usage billed to the wrong account, or not at all | Rigorous mediation, single customer reference |
| Hard-coded pricing rules | Each new offer becomes an IT project | Configurable catalog, combinable pricing models |
| Unreadable invoices | Disputes, credit notes, late payment | Readable usage detail, line-by-line explanations |
| Unpredictable revenue | Worried finance and customers | Minimums, caps, usage alerts, hybrid model |
| Volumes not handled | Slow processing, delayed close | Architecture designed for volume, scalability |
The most underestimated point is feed completeness. A calculation error shows: the customer complains. A feed that never arrives does not show: the usage is simply never billed. This is a classic source of revenue leakage.
The second is invoice explainability. A variable invoice the customer does not understand becomes a disputed invoice. And a disputed invoice is a late payment.
How to choose a usage metric
A good usage metric meets five criteria: the customer understands it, it can be measured reliably, it tracks the value the customer gets, the customer can anticipate it, and it cannot easily be gamed. A metric that misses one of these creates disputes or revenue loss.
The metric is the unit the whole model rests on. It deserves more time than the price grid.
- Understood: the customer must be able to relate the metric to their activity. "Per document processed" is clearer than "per compute unit".
- Measurable: the data must be available, reliable and verifiable, by you and by the customer.
- Value-linked: when the customer gets more value, the metric goes up.
- Predictable: the customer must be able to estimate their invoice in advance, or they will limit usage.
- Hard to game: a metric that can be reduced without reducing value invites workarounds.
In telecoms or energy, the metric is set by the service (minute, kilowatt-hour). In software and AI, it is chosen, and that choice often decides the offer's success.
Switching to usage-based billing in 7 steps
Moving to usage-based billing takes seven steps: choose the metric, design the offer (often hybrid), check usage data availability, configure pricing models, test on real invoices, launch gradually, then steer with KPIs. The price grid comes after the data, not before.
- Choose the metric: with the five criteria above. Test several candidates on a few customers' consumption history.
- Design the offer: subscription base, included volume, price beyond, minimum, cap. Simulate the effect on existing customers' invoices: who would pay more, who less?
- Check usage data: where is each event produced? Does it arrive complete, on time, linkable to a customer? This diagnosis conditions everything else.
- Configure pricing models: in an offer catalog, without custom development. The goal: launch an offer variant in days, not months.
- Test on real invoices: replay a past period under the new model and compare the invoices. Have customer care review them: can they be explained?
- Launch gradually: start with new customers or one segment, plan a period of parallel calculation, support existing customers with invoice simulations.
- Steer: track revenue per customer, consumption, disputed invoices and missing feeds. Adjust the offer based on real usage.
How to choose usage-based billing software
Usage-based billing software must collect and rate high event volumes, combine every pricing model without development, check feed completeness, produce detailed, understandable invoices and integrate with CRM and ERP. The deployment model and data hosting matter too.
Billing tools for small businesses rarely handle usage. Generic ERPs do it with custom development. Specialist platforms differ widely in their ability to handle volume and complex pricing.
| Criterion | Questions to ask |
|---|---|
| Volumes | How many events per day can it rate? Real time or batch? |
| Pricing models | Unit, volume, tiers, overage, minimum, cap, credits: can they combine in one offer, without development? |
| Mediation | Does it clean, deduplicate and link events to the right customer? |
| Completeness | Does it detect a missing feed before billing? |
| Readability | Does the invoice detail consumption? Can support explain every amount? |
| Integration | Documented APIs, CRM, ERP and payment connectors? |
| Deployment | SaaS, cloud or on premise? Where is data hosted? |
| Agility | Can you launch a new offer without an IT project? |
Ask to see your own cases in the demo: a mid-month plan change, an overage, a feed arriving late.
That is the promise of the Opencell platform: a mediation, rating and billing engine built for high volumes, where flat fees, unit prices, volume, tiers, overage, minimums and caps combine in one offer. Our AI agents complete it: one flags a missing usage feed before rating, another catches a usage anomaly before the invoice, a third explains a billed amount back to the source data.
Frequently asked questions about usage-based billing
What is the difference between usage-based billing and usage-based pricing?
Usage-based pricing is the commercial decision to set a price per unit consumed. Usage-based billing is the operational ability to measure that consumption, rate it and invoice it.
What is metered billing?
Metered billing is another name for usage-based billing: the customer is billed according to a measured quantity, like a meter reading, an API call count or data volume.
Does usage-based billing work in B2B?
Yes. It is the norm in B2B industries such as telecoms, energy and cloud. In B2B, it often takes a hybrid form, with a minimum commitment and usage billed beyond it.
How do you forecast revenue with usage-based billing?
By combining three levers: a hybrid model with a recurring base, minimum commitments, and consumption data available as it happens. Customer usage alerts also reduce surprises on both sides.
What is rating in billing?
Rating is the step that assigns a price to each usage event under the customer's contract: unit price, tier reached, included volume, discount. It sits between mediation and invoicing.
What is billing mediation?
Mediation turns raw usage data into billable data. It removes duplicates, formats data, checks consistency and links each event to the right customer and contract before rating.
Should you drop subscriptions to move to usage?
Rarely. Most companies keep a subscription base and add a variable part. This hybrid model secures recurring revenue while capturing growth.
Billing on usage starts with measuring right
Usage-based billing aligns price with value. It appeals to customers, grows revenue with them and passes on costs that have become variable.
But it moves the difficulty: from the price grid to the data.
Key takeaways:
- usage-based billing charges actual consumption, often on top of a subscription,
- seven pricing models combine, and their calculation differences are often misunderstood,
- the chain runs from event to invoice, through mediation and rating,
- feed completeness and invoice readability are the two most underestimated points,
- choose the metric and the data before the price grid.
Selling on usage is a promise. Billing it right is a craft.
Go further with Opencell
Opencell is an agile, open source monetization platform that automates quote-to-cash, from quote to cash collection, for subscription and usage-based services. It manages €14 billion in revenue for around fifty customers in telecoms, energy, cloud, IoT, mobility and services.
Moving to usage-based pricing, or making your consumption billing reliable?



