Pricing is the single highest-leverage product decision in SaaS, and it's the one teams change least often. The model you pick — per-seat, usage, tiered, or hybrid — sets a ceiling on growth, the volatility of your revenue, and the kind of customers you'll attract. This guide walks through the trade-offs and where each model breaks.
Per-seat pricing
How it works: Charge per active user, usually per month. Slack, Notion, Linear all use it.
Why it's popular: Predictable for buyers, easy to model on their side, expansion happens automatically as headcount grows.
Where it breaks:
- If your product gets more valuable as fewer users use it (admin tools, automation), per-seat actively penalizes you for delivering value.
- Customers de-provision unused seats aggressively — your "expansion" can reverse on a quarterly cleanup.
- You're capped by the buyer's headcount; you can't earn more from a customer who finds 10× the value but has the same team size.
Usage-based pricing
How it works: Charge per unit of consumption — API calls, documents processed, GB stored, messages sent. Snowflake, Twilio, OpenAI.
Why it's popular: Aligned with value delivered, friction-free adoption (start small, grow with the customer), naturally compounding.
Where it breaks:
- Revenue volatility: customer drops a project, your MRR drops with it. NRR can swing wildly month to month.
- Buyer anxiety: enterprises hate unpredictable bills. You'll need to offer commit contracts to win the deal.
- Forecasting becomes harder for finance and sales planning.
- Free riders: usage forecasting is hard, and a subset of customers will engineer their workflow to minimize your meter.
Tiered (good/better/best)
How it works: Three or four bundled plans at fixed price points. Most B2B SaaS still leads with this.
Why it works: Clean buying experience, easy to communicate, anchors price expectations.
Where it breaks:
- Tiers eventually misalign with usage — power users on the cheap tier, light users on enterprise.
- Hard to capture value from outliers without adding seats or usage on top.
- Plans accrete features over time and become unintelligible to new buyers.
Hybrid: where the market is going
Most modern SaaS is converging on a hybrid: a base platform fee (often per-seat or tiered) plus a usage component that scales with value. HubSpot, Notion AI, and the new generation of AI-enabled SaaS all use some version of this.
The shape that's working:
- Base: per-seat or tiered, captures the core platform value.
- Usage layer: consumption-based for high-value/AI-heavy features (queries, automations, generated content).
- Enterprise floor: annual minimum commit that gives finance teams predictability while preserving usage upside.
Model your blended ARPU and projected MRR in our SaaS MRR & ARR calculator.
The pricing-power test
Independent of model, ask one question every six months: can we raise prices 15% on new customers without losing more than 10% of the pipeline? If yes, you have pricing power and you're under-priced. If no, the issue isn't pricing — it's positioning, ICP, or product depth. Pricing changes won't fix those.
The unit-economics check
Whichever model you choose, the LTV:CAC math has to work. A clever pricing model with terrible unit economics is still a bad business. See our LTV:CAC benchmarks guide and run the numbers in the LTV:CAC calculator before you commit to a pricing redesign.