Seat vs. Usage: How to Choose a Pricing Model for B2B SaaS in 2026
Choosing a pricing model is one of the most consequential decisions you'll make as a founder. For B2B SaaS companies that process payments, enable transactions, and orchestrate workflows for their clients, the wrong choice doesn't just leave money on the table — it distorts your metrics, caps your NRR, and makes your revenue story harder to underwrite. The three main options each have a distinct fit. Most companies above $3M ARR end up with a hybrid of two of them.
How seat-based pricing works
You charge a fixed monthly or annual fee per user. It's the classic SaaS model: Salesforce, Notion, Figma.
Seats are easy to sell, easy to forecast, and easy for procurement to approve. The problem is structural: as AI reduces headcount and automates tasks, the number of people logging into your product goes down — even if the value you deliver goes up. Microsoft's Satya Nadella put it plainly on a recent earnings call: seats are now "just entitlement to some consumption."
Works for: collaboration tools, CRMs, communication platforms — products where value scales with human users.
Breaks for: payments infrastructure, AI agents, transaction processing, API-first products.
The hidden risk: if a customer automates away two of their five seats, you lose 40% of that contract with no recourse.
How usage-based pricing works
Customers pay based on what they consume — API calls, transactions processed, tokens used, emails sent. The more they use, the more they pay. Snowflake, Stripe, Twilio, and OpenAI all run this model.
The upside is structural: you grow with your customers automatically. A SaaS platform processing $10M/month today that hits $50M/month in 18 months pays you 5x more with no upsell required. Intercom's Fin.ai demonstrated this flywheel — NRR climbed from 112% to 146% after launching per-resolution pricing, reaching $100M ARR in three years.
The downsides are real too:
- Revenue is harder to forecast without consumption history
- Sales comp gets complicated — do you pay on commit or on actual usage?
- Enterprise buyers in fintech and healthtech push back on open-ended consumption because their finance teams can't budget for it
For payments and fintech specifically: per-transaction or per-API-call pricing is the most natural fit. Your revenue literally scales with your customer's GMV. That's a defensible value story. OpenView's research on usage-based pricing shows that usage-based companies consistently command higher NRR and faster expansion revenue than seat-based peers.
How flat-fee unlimited pricing works
One price, unlimited use. Simple to sell, simple to understand, no billing surprises.
The problem: flat-fee pricing has no expansion motion. As Kyle Poyar of Growth Unhinged writes in The State of B2B Monetization in 2026, flat-fee is "structurally broken" for companies that need revenue growth beyond new logos:
- Small customers get priced out
- Large customers pay too little
- Your top 10% of power users — who drive 70-80% of AI token consumption — get subsidized by everyone else
- There's no natural path to NRR above 100%
When it makes sense: early-stage customer acquisition, before you have enough usage data to price dynamically. Riya Grover, CEO of Sequence, points to early-stage companies using flat-fee to build their customer base first, then layering in usage pricing once patterns emerge.
Seat vs. usage vs. flat-fee: side by side
| Seat-based | Usage-based | Flat-fee | |
|---|---|---|---|
| Revenue predictability | High | Low–Medium | High |
| Expansion potential | Capped by headcount | Uncapped | None |
| Investor preference (2026) | Low (5%) | Medium (24%) | Low (10%) |
| Works for payments/fintech | Rarely | Yes | Early stage only |
| Works for vertical AI | No | Yes | No |
| Forecasting complexity | Low | High | Low |
| Enterprise buyer acceptance | High | Medium | High |
| Finance stack complexity | Low | High | Low |
The model most companies actually land on: hybrid
In Kyle Poyar's 2026 State of B2B Monetization survey of 230 companies, 37% use a hybrid model — the most common structure, up from 25% a year earlier. The three most common hybrid patterns:
1. Base commit + overage
A monthly subscription that includes a usage allowance, with overage billed per unit. GitHub and Zapier both use this structure. Buyers get budget predictability; you keep the upside.
2. Platform fee + consumption
A flat access fee plus separate billing for token usage or transaction volume. Clay introduced this in early 2026, explicitly separating "value" (the platform) from "cost" (the tokens). Anthropic followed by cutting Enterprise seat prices while shifting more aggressively to consumption billing.
3. Tiered subscriptions with usage gates
Packages priced by feature set, with usage limits that push customers up tiers. The key is that limits exist — unlimited plans kill the expansion motion.
For B2B SaaS companies that process payments and enable transactions, the natural hybrid is a monthly platform fee plus per-transaction pricing. Shopify is the canonical example: subscriptions account for less than 30% of revenue; the rest is transaction fees and FinTech revenue that scales automatically with merchant volume. Forbes outlines the proven playbook for shifting to usage-based pricing and confirms that most companies land on this hybrid structure after their first attempt at pure consumption pricing.
3 questions to ask before you decide
1. Does your value scale with usage or with users?
If your product processes transactions, generates outputs, or performs work autonomously, usage-based fits. If value is about team collaboration or workflow access, seats make more sense.
2. How much consumption history do you have?
Steve Love, a six-time CFO who advises growth-stage companies, notes that pure usage-based pricing is hardest to run without consumption data. Start with a hybrid that includes a minimum commit — it's easier to loosen constraints later than to tighten them.
3. What does your buyer's budget process look like?
Enterprise fintech and healthtech buyers can't hold open-ended contingency budgets for unknown consumption. A base commit with consumption on top clears procurement; pure PAYG often doesn't.
What this decision does to your finance stack
This is the part most founders underestimate. Your pricing model isn't just a revenue decision — it restructures how you close books, recognize revenue, and forecast.
- Seat-based: predictable, low overhead, straightforward rev rec
- Usage-based: requires metering infrastructure, triggers ASC 606 variable consideration treatment, makes MRR an unreliable metric, complicates CAC payback calculations
- Hybrid: inherits usage-based complexity plus the overhead of managing base commits, thresholds, and overage calculations across your entire customer base
The billing infrastructure that follows this decision is where a lot of companies get into trouble — but that's a separate article.
