SaaS pricing debates get religious quickly. Per-seat teams accuse usage pricing of surprise bills. Usage teams accuse seats of punishing collaboration. The right model maps to how value accrues and how the buyer wants to forecast spend. Everything else is packaging theater.
The three families
How value, sales motion, and risk line up
| Model | Best when | Buyer fear | Churn mechanic |
|---|---|---|---|
| Per seat | Value scales with people in the workflow | Shelfware seats | Contraction at renewal |
| Usage / consumption | Value scales with jobs run, GBs, API calls | Unpredictable invoices | Sudden spike → rage-quit |
| Hybrid | Platform + bursty consumption | Complexity / two meters | Misaligned packaging |
Seats still win many B2B categories
If a human must live in the UI daily—CRM, design, support—seats are intuitive. The trap is charging for viewers who only approve. Offer a cheap or free approver role or you will train customers to screenshot their way around your meter.
Usage works when the unit is obvious
Messages sent, nightly builds, records processed, tokens consumed: if the customer already thinks in that unit, usage can expand with success. Publish a pricing simulator. Enterprise buyers will build one anyway, and they will assume the worst if you do not.
Hybrid without confusing the market
- One primary meter in the headline price.
- A secondary meter only for costly infrastructure (AI tokens, SMS, storage).
- Committed spend with burst overage, not three overlapping SKUs.
- A true-up language a procurement team can explain to their CFO.
Conclusion
Choose seat pricing when collaboration is the product, usage when the job is metered and beloved, and hybrid when a platform fee must protect gross margin on bursty AI or messaging. Optimize for forecastability as much as willingness to pay. In 2026, surprise is a churn feature—not a growth hack.