Pay for what we really use, the trend of use-based pricing in the SaaS model

For years, the monthly package has remained an unbearable base of the SaaS. A stable line in income forecasts, psychological comfort for the customer as well as for the investor, a symbol of maturity in commercial structuring. But as products become more technical, distributed, API-first, that they integrate into complex stacks, and that adoption becomes progressive, this model shows its limits. The promise of a single subscription, supposed to reflect a universal value, no longer holds.

In 2025, pricing for us, long marginal, resurfaces, no longer as an anomaly reserved for cloud infrastructure or niche APIs, but as a real Income piloting alternative. For financial and product directions, it is a structural change in the way of defining, capturing and invoice the value.

Return to the economic basis of the product

The question is no longer “How much is my software worth?” “, But” when I offer really becomes useful, and for what intensity of use? ». Where a Seat-Based model supposes an artificial equivalence between a user and a monetary value, the use-based pricing starts from a different postulate, it is the real level of consumption that determines the generated value.

At Datadog, each metric collected, each second observation counts, at Snowflake, each byte stored or treated is measured, Openai invoices its customers to the Token. Nothing abstract, nor a future promise monetized in advance, We pay what we use.

But this apparent simplicity is based on a fundamental prerequisite, Know precisely what we measure, why we measure it, and especially how this measure is perceived by the customer. It is here that use-based pricing becomes a strategic design act, not a simple commercial tool.

Temptation and risks for SaaS companies

Many founders, CFO or CPO are seduced by the promises of the model, better retention, reduced commercial friction, facilitated adoption via low entrance tickets. These promises are often true, many are among those who have adopted it who observes that their net revenue retention is higher, that the sales cycles shorten, and customers feel more in control.

But behind this virtuous mechanism hide several traps. The first is that of theunpredictability. For the financial management, a purely variable model complicates the projection of the MRR and the cash flow. Gardement must be introduced, such as consumption levels, guaranteed minimum plans, or prepayment, like Amazon EC2 or Paddle, for example.

The second is that of theclient misunderstanding. UBP, badly formulated, can quickly look like an opaque tax. The customer discovers at the end of the month an invoice he does not understand, indexed to technical metrics that he does not control. Here, the role of the product becomes critical. Visualization of use, estimation calculators, overtaking alerts are not options but the central elements of contractual experience.

The third trap is that of technical instrumentation, We cannot invoice in use what we do not know precisely measure. This presupposes robust Metering infrastructure, adaptive Billing systems, close collaboration between engineering, product, finance and legal.

Gladia: a French SaaS who assumes the model for use

The startup Gladiaspecialized in Audio Transcription Assisted by AI, offers pricing at the second of treated audio. This choice is not anecdotal. In an environment where the unit cost (compute, model, bandwidth) is high and where adoption varies strongly depending on the use case, only invoicing to use makes it possible to offer an accessible input while capturing the actual value on a large scale.

Gladia relies on Hyperlinea French solution of Billing Use-Based, to manage both the instrumentation of use, the generation of invoices, and the customer interface. The system makes it possible to address both exploration developers and large groups with a high volume, without additional commercial effort.

Structuring solutions to deploy UBP

Adopting use-based pricing is not just a product or marketing decision. It is a Financial architecture decision. It implies the implementation of tools capable of measuring, visualizing and invoicing the use, in real time and without friction.

Several solutions are distinguished today:

• Hyperline (France) is designed for European startups. It offers a light, quick to integrate infrastructure, adapted to PLG or API-STE models. Widely used at IA startups, it allows you to charge by event, volume or request, with a clear UX side customer.
• Thriller (Sweden) is a complete solution to expose a use-based offer without technical overload. Allows you to define dynamic tariff rules, integrate monitoring into the product, and to view consumption for end customers.
• Metronome (United States): infrastructure tool for very high-scale companies. Often chosen by publishers with massive volumes of data and a need for personalization of the Billing (eg Openai, Vercel).
• Octobat (France): Local alternative closer to the European tax environment, integrated into conventional accounting stacks.

A structuring, but reversible decision

SaaS publishers who succeed in their transition to use-based do not make the mistake of switching everything at once. They test by use case (API, compute), integrate a fixed plan with overconsumption thresholds, or reserve invoicing for use for certain customer segments.

Pricing to use, when it is well thought out, makes it possible to reconcile commercial performance, product adoption, financial rigor and customer experience.