The way SaaS companies charge for their products is undergoing the most significant structural shift since the move from on-premise licences to subscriptions. 61% of SaaS companies now use usage-based pricing components, up from 45% in 2021 (Chargebee). Pure per-seat pricing — once the default for everything from CRMs to project management tools — has collapsed to just 8% of the market, down from 47% in 2020.
This is not a trend. It is a structural reset, driven by three converging forces: AI agents that do the work of multiple human seats, buyers who refuse to pay for shelfware, and vendors who realised that aligning cost to value is better for long-term retention than front-loading revenue in a sales motion.
Whether you are building a SaaS product and debating your pricing architecture, or buying SaaS tools and trying to understand what you are actually signing up for, this guide gives you the numbers and the framework to make the right call.
The Four Models: How They Actually Work
Before benchmarking, it helps to be precise about what each model means in practice. "Usage-based" is an umbrella term that gets applied to very different billing approaches.
| Model | How It Works | Pros | Cons | Best For |
|---|---|---|---|---|
| Seat-based | Fixed price per named user per month or year | Predictable revenue; easy to quote and invoice | Users share logins; paying for shelfware; penalises growth | Collaboration tools with per-person value |
| Usage-based (pure) | Billed on consumption — API calls, events, tasks, rows | Cost scales with value delivered; low barrier to start | Revenue volatility; hard to forecast; customers hate surprise bills | Developer tools, AI APIs, data pipelines |
| Flat-rate / tiered | Fixed monthly fee for a feature tier, regardless of users or usage | Simple to communicate; easy to budget | Doesn't reward power users; poor fit for high-variance usage | SMB tools with relatively homogenous use cases |
| Hybrid | Base subscription (seat or tier) + usage overage or credit system | Balances predictability with alignment; captures upside | Complexity in billing; harder to explain; requires good metering | Growth-stage B2B SaaS |
43% of SaaS companies already use hybrid models, and that figure is projected to reach 61% by the end of 2026. This is where the market is heading, and for good reason.
Why Seat-Based Pricing Is Losing Ground
The seat model made sense when software was used by a defined set of named employees. You licensed Microsoft Office for 50 employees. You paid for 50 seats. Simple.
Three things have broken this logic.
First, AI agents. A single agent doing research, drafting, and scheduling is functionally replacing 3–5 human seats. You cannot charge per seat when the seat is a bot. IDC projects that 70% of vendors will move away from pure per-seat pricing by 2028 because the unit economics simply do not survive AI-native workflows.
Second, buyer sophistication. Enterprise procurement teams now run software audits. Shelfware — software purchased but unused — is a line item they actively cut. Pure seat pricing creates shelfware almost by design, because procurement locks in 50 seats for the year and only 30 are ever activated.
Third, competitive pressure from usage-based entrants. When Zapier charges per task and Stripe enables metered billing infrastructure that any developer can wire up in an afternoon, the barrier to offering usage-based pricing has dropped to near zero. New entrants default to it, forcing incumbents to follow.
The Revenue Trade-Off: Growth vs Predictability
Usage-based companies achieve 38% faster revenue growth than pure seat-based companies (Chargebee, 2025 cohort analysis). But faster growth comes with a cost: annual contract values are 23% lower with pure usage-based billing.
This is the central tension. You expand faster because customers can start small and ramp. But each individual contract is smaller, and revenue is harder to predict, which matters to your CFO, your investors, and your ability to model headcount.
The median free-to-paid conversion with usage-based pricing is 7.2%, versus 4.8% for flat-fee models. That 50% difference in conversion rate often more than offsets the lower ACV at the top of funnel — but only if you have sufficient volume and a product with genuine expansion mechanics.
The hybrid model is the pragmatic answer for most B2B SaaS companies in 2026. You lock in a predictable base (seat licence or tier fee), then capture upside through usage overages or credit consumption. HubSpot has been shifting toward this structure in its professional and enterprise tiers. Amplitude uses a monthly tracked users (MTU) model that blends a base commitment with consumption.
Choosing by Stage: Which Model Fits Where You Are
The right pricing model is not universal. It depends on where your company is in its development, what your buyer looks like, and how much operational complexity you can handle.
| Stage | Recommended Model | Why | Risk |
|---|---|---|---|
| Pre-product / MVP | Flat-rate or freemium | Reduces friction; focus is on learning | May undercharge power users |
| Seed / early PMF | Usage-based (simple) | Aligns cost to value; lowers barrier to first paid conversion | Revenue unpredictability |
| Series A / scaling | Hybrid (base + overage) | Balances predictability and expansion revenue | Billing complexity requires proper metering |
| Series B+ / enterprise | Hybrid with annual commits | Enterprises need to budget; usage upside via ELAs | Sales motion becomes more complex |
| Mature / broad horizontal | Tiered hybrid | Large buyer heterogeneity needs multiple entry points | Risk of tier confusion |
If you are at seed stage and debating this, the answer is almost always: start simpler than you think you need to. The operational overhead of metered billing — handling disputes, building dashboards, managing overages — is significant. Get to Product-Market Fit first, then layer in the complexity.
"When a single AI agent can do the work of 3–5 human seats, per-seat pricing creates an immediate commercial absurdity: the more efficient your product becomes, the less revenue it generates."— SaaSTweaks Editorial, 2026
AI Agents: The Pricing Model Killer
Gartner estimates that 40% of enterprise SaaS spend will shift to usage, agent, or outcome-based pricing by 2030. That is not a distant forecast. It is already happening in the categories adjacent to AI workflows.
When a single AI agent can do the work of 3–5 human seats, per-seat pricing creates an immediate commercial absurdity: the more efficient your product becomes, the less revenue it generates. No vendor can survive that incentive structure.
| Model | Impact of AI Agents | Likely Outcome by 2028 |
|---|---|---|
| Seat-based | Directly undermined — agents replace seats without generating additional per-seat revenue | Largely abandoned outside legacy enterprise contracts |
| Usage-based | Well-suited — agent actions (API calls, tasks, inferences) are natural billing units | Dominant model for AI-native SaaS |
| Flat-rate / tiered | Vulnerable — AI capabilities increase value delivered without increasing billing | Pressure to migrate to hybrid or usage tiers |
| Hybrid | Best positioned — base commit captures floor, usage captures AI-driven expansion | Default model for growth-stage B2B SaaS |
| Outcome-based | Early stage — charge for results | May reach 15–20% of enterprise deals by 2028 |
Analytics tools that track AI agent activity — events fired, sessions initiated, actions completed — are becoming essential infrastructure for any company moving to usage or outcome billing. Without instrumentation, you cannot bill. Without billing data, you cannot optimise your pricing.
What This Means If You Are Buying SaaS
If you are on the buying side, the pricing model is as important as the feature set. A tool that looks cheap on a per-seat basis may become very expensive at scale if it carries significant usage overages you did not model.
Three questions to ask any SaaS vendor before signing:
- What is your overage rate, and is it capped? Uncapped overage billing is a budget risk. Ask for a cap or a notification trigger.
- How is usage defined, and who controls it? Some vendors define "active users" in ways that catch buyers off guard.
- What happens if our usage drops? With pure usage-based pricing, this is fine. With a hybrid model carrying a base commit, you may be paying for capacity you no longer need.
Workflow automation tools are a particular watch-out here. Task-based billing — as used by Zapier — can balloon quickly if you build automations that run frequently or trigger unexpectedly. Always model your expected task volume before committing.
FAQ
What is usage-based pricing in SaaS?
Usage-based pricing charges customers based on how much they consume — API calls, tasks run, events tracked, data processed, or seats actively used. Cost scales with consumption rather than being fixed upfront. It is the dominant pricing direction for SaaS in 2026, with 61% of companies now using some form of usage-based component.
Is usage-based pricing better for buyers or vendors?
Both benefit, but in different ways. Buyers benefit from lower barriers to entry and costs that align with the value they receive. Vendors benefit from expansion revenue and higher conversion rates. The trade-off for vendors is revenue volatility — annual contract values are 23% lower on average with pure usage-based models compared to seat-based equivalents.
Which pricing model is best for a seed-stage SaaS startup?
Start with a simple flat-rate or basic usage-based model. Do not introduce billing complexity before you have product-market fit. The operational overhead of metered billing — infrastructure, dispute handling, dashboard tooling — is significant and will distract from the more important work of validating your product.
How do AI agents change SaaS pricing models?
AI agents fundamentally break the seat model because they replace multiple human users without generating per-seat revenue. Vendors are responding by shifting to task-based, inference-based, or outcome-based billing. Gartner projects 40% of enterprise SaaS spend will shift to usage, agent, or outcome-based pricing by 2030.
What is a hybrid SaaS pricing model?
A hybrid model combines a fixed base fee — typically a seat licence or a tier subscription — with a variable usage component for consumption above a threshold. It gives vendors predictable revenue while allowing expansion billing as customers grow. 43% of SaaS companies already use a hybrid structure, with that figure projected to reach 61% by end of 2026.
How does seat-based pricing work?
Seat-based pricing charges a fixed amount per named user per billing period. It was the dominant SaaS model from roughly 2008 to 2020. Today, pure seat-only pricing accounts for just 8% of the market, down from 47% in 2020. It remains appropriate for human-centric collaboration tools but is structurally unsuited to AI-augmented workflows.
