The 3 B2B Pricing Models Replacing Seat Licenses. Pick One Before You Get Picked Off.

Enterprise software pricing has been stuck on seats for decades. The logic was elegant: more users means more value, which means more licenses. AI has broken the logic by enabling smaller teams to deliver larger outputs — making seat counts an increasingly unreliable proxy for the value software delivers.

The Framework

SaaStr’s analysis opens with data that explains why the shift is accelerating: B2B software costs rose from $7,900 per employee (2023) to $9,100 per employee (2025), while application counts stayed flat. Enterprise spend is $55.7M annually on average. CIOs are cutting traditional software to fund AI — 45% of AI budgets come from existing software allocations. And 79% of IT leaders experienced price increases at renewal, while 78% encountered unexpected AI feature charges.

The old unit is broken: seats measure headcount, not output. Three new models are replacing it.

Consumption Pricing

What it is: Customers pay for what they use — credits, tokens, API lookups, records processed, or compute runs.

Who uses it: Stripe, Twilio, Snowflake, Databricks, MongoDB. Nearly every infrastructure-adjacent product has moved here.

The advantage: Direct alignment between cost and value. Customers with variable workloads pay proportionally; high-use customers pay more.

The catch: Bill shock. Customers struggle to forecast variable costs, and surprise invoices create churn risk even for satisfied users.

The fix: Hard caps, pre-purchased blocks with rollover, threshold alerting, and budget APIs that agents can access for real-time spend visibility. The consumption model works when vendors help buyers forecast and cap exposure; it fails when the variance is invisible until month-end.

Outcome Pricing

What it is: Payment triggered by defined business results, not usage or access.

Who uses it: Palantir (contracts anchored to measured dollar impact), Sierra (per-resolved case, expanding to multi-week business goals).

The advantage: Converts the software purchase into a measurable business case with explicit ROI. The vendor has skin in the outcome, not just the license.

The catch: Attribution complexity. Proving that the software caused the outcome — rather than the team, the market, or seasonal factors — requires organizational alignment on measurement methodology. It also often requires restructuring how success is tracked across the customer’s business.

Resolution Pricing

What it is: The simplified version of outcome — bill when one countable thing happens or doesn’t. A support ticket resolved. A data record matched. An API call completed successfully.

Who uses it: Intercom’s Fin ($0.99 per outcome), Sierra (~$1.50), HubSpot ($0.50 per resolution).

The advantage: “When both sides look at the same number and agree on it, the pricing argument mostly disappears.” Resolution pricing works when the outcome is unambiguous, countable, and agreed upon. Support resolution rates meet that bar more cleanly than most enterprise outcomes.

The catch: Only works for processes that have that clean, countable unit. Professional services, complex analysis, or strategic advisory work don’t compress to a single resolution event.

Where It Doesn’t Apply

Seat pricing isn’t dying universally. Software that controls access rather than generating output — identity management, compliance tooling, security platforms — often still makes sense on a per-user or per-resource basis because the value is access, not throughput. Early-stage companies without meaningful usage data to meter may also need to start with seats before they understand what unit their customers care about.

As Stack Archive has covered, the pressure on enterprise software revenue from agentic AI makes the seat model unsustainable where agents handle work that users once did. The new models aren’t a response to that pressure — they’re the path through it.

Source: SaaStr

Content created with AI assistance and reviewed for accuracy.