Bain & Company's latest research note carries a deliberately hedged headline: "Per-Seat Software Pricing Isn't Dead, but New Models Are Gaining Steam." For product managers running B2B SaaS businesses, that hedge is the headline — per-seat remains defensible in the right categories, and the wrong pricing model costs more than the right one ever saves.

Where per-seat still wins

Per-seat pricing ties cost to user count. It dominates where every licensed user generates roughly equivalent value: collaboration suites, CRM, and core developer tooling. In each, the per-employee line item fits how buyers procure and budget. Sales cycles stay short because reps can quote without usage modeling, and finance can forecast renewals from seat counts alone.

Where per-seat has already lost

Per-seat has long struggled where value scales with data volume, API calls, or automated workloads rather than human headcount. Data infrastructure, communications APIs, and developer-facing infrastructure have moved to consumption- or usage-based pricing for years. Snowflake's credit model, AWS instance-hours, and Twilio's per-message pricing capture meaningful revenue without per-seat billing.

What are the new models?

• Consumption-based charges customers for units consumed — compute hours, rows scanned, gigabytes stored.
• Usage-based charges for specific events — API calls, messages sent, records processed.
• Outcome-based ties cost to a measured result. Still rare outside healthcare revenue cycle and a handful of enterprise deals.
• Hybrid combines a platform fee or committed minimum with variable usage. Databricks and AWS have moved this way as their catalogs broadened.

Why is the shift accelerating now?

AI features have broken the per-seat math. A single user of an AI-augmented tool can produce output that previously required a department. Charging per seat undercharges heavy users and overcharges light ones. Salesforce's Flex Credits and Slack's AI add-on are early moves to thread this needle without abandoning per-seat entirely.

What does this mean for Monday?

For product managers evaluating pricing changes, the right question is not "should we move off per-seat?" but "does seat count still track value creation in our category?" If yes, per-seat remains the lowest-friction model. If no, usage or hybrid likely beats it on net revenue retention, even if it lengthens deal cycles and complicates RevOps.

The binding constraint on a pricing migration is rarely the new price list. It is the billing system, the sales compensation plan, and the customer-success playbook. PMs who own pricing should treat those systems as the real scope of the project, not the SKU table.

Where this is heading

Expect category-by-category bifurcation: per-seat stays default in human-collaboration software while hybrid and usage-based dominate wherever automation scales value past headcount. Bain's read is the right baseline — the pricing question for PMs running net-new features is no longer seat or usage but where on the spectrum this product lives, and what the billing system must support in 18 months.