Poyar State of B2B Monetization 2026
Kyle Poyar's Growth Unhinged annual survey on B2B monetization. Published May 13, 2026. N=230 SaaS/AI companies. The clearest signal: hybrid pricing won, credits are exploding, and AI margins are half what SaaS margins were.
Survey Demographics
- N=230, fielded April-May 2026
- ARR: 22% <$1M, 28% $1-20M, 24% $20-150M, 25% >$150M
- ACV: 28% <$5K, 27% $5-25K, 30% $25-100K, 16% >$100K
- Product type: 32% mostly SaaS, 15% AI-native, 43% hybrid SaaS+AI, 10% other
Key Findings
1. Hybrid pricing dominates (37%)
37% of respondents use hybrid pricing, up from 25% twelve months ago. The shift is not uniform by stage:
- Early-stage (<$5M ARR): 37% still on flat fees
- Large (>$150M ARR): 29% still per-seat
Investor preference diverges further from current adoption: 35% prefer hybrid, 26% outcome-based, 24% usage-based, 10% flat-fee, 5% seat-based.
2. Biggest complaint: not enough expansion revenue
Most cited challenge across all respondents. Particularly acute for flat-fee and seat-based models. Seat-based companies specifically worry the model is "under threat and not future-proof." Usage/outcome-based companies report the inverse problem: forecasting difficulty.
#### Challenges by Pricing Model
Respondents could select up to two challenges.
| Challenge | Flat-fee | Seat-based | Usage-based | Outcome-based | Hybrid |
|---|---|---|---|---|---|
| Not enough expansion $ | 38% | 36% | 13% | 20% | 13% |
| Not future proof | 26% | 36% | 17% | 0% | 18% |
| Expensive for small customers | 26% | 18% | 30% | 10% | 18% |
| Hard to forecast/predict | 10% | 9% | 37% | 30% | 23% |
| Difficult to explain | 10% | 6% | 23% | 20% | 35% |
| Not aligned with value | 20% | 30% | 13% | 0% | 9% |
| Customers self-police | 4% | 21% | 20% | 30% | 13% |
| Too much complexity | 4% | 6% | 7% | 20% | 14% |
| Low margin customers | 12% | 12% | 7% | 10% | 6% |
The hybrid paradox: best expansion + future-proofing scores but hardest to explain (35%).
3. Pricing changes accelerating
75% changed pricing or packaging in the past year. Largest companies are making the most changes. Named companies with significant early 2026 changes: Salesforce, Anthropic, HubSpot, OpenAI, Clay, Figma, Canva, SAP, GitHub.
4. AI credit wave
- Current adoption: 29% (up from 13% in H1 2025 — 126% YoY)
- 33% plan to add AI credits within 6-12 months
- >$50M ARR: 1-in-2 plan AI credits this year
- Projected: 62% by H1 2027
- Named H1 2026 adopters: Airtable, Atlassian, GitHub, Apollo, Figma, HubSpot, Salesforce
Poyar: "credits are a lifeline — just not the AI pricing endgame."
5. AI margins are 50%
Target gross margin distribution across respondents:
| Target margin | % of respondents |
|---|---|
| Breakeven/negative | 7% |
| 1-20% | 7% |
| 21-40% | 16% |
| 41-60% | 31% |
| 61-80% | 26% |
| 81-100% | 12% |
Only 12% aim for SaaS-like 80%+ margins. Internal costs are the #1 factor when pricing AI features (54%). Secondary: market/competition (36%), productivity gains (30%).
6. AI cannibalizing SaaS budgets
70% said AI spending comes from existing tech/software budgets (75% for SaaS companies adding AI). AI-native companies more likely to tap services (35%) or headcount (15%) budgets. Sequoia's thesis: AI-native services will create the next $1T company.
7. Multiple pricing models offered
29% let customers choose between pricing models (up from 21%). Most common among usage/outcome-based companies. Salesforce leads with 4 concurrent models: per-conversation, flex credits, per-user add-ons, Agentic ELA.
Disruptive AI Pricing Changes (Apr 2026)
| Company | Change |
|---|---|
| Salesforce | Headless MCP — platform as APIs, no seat required |
| Clay | Dual-track: platform fee + token pricing |
| HubSpot | Outcome-based pricing (pay per lead, per resolved conversation) |
| HubSpot | $0.50/resolved conversation (half the cost of Intercom Fin) |
| Figma | AI credit add-ons available across all plans |
| Fin (Intercom) | AI analytics add-on at $99/1K conversations/month |
| Notion | Credit model for custom agents (starts May 4, 2026) |
| Anthropic | Lowered ENT seat pricing, shifted to usage-based at API rates |
| Anthropic | Tested Claude Code as Max-plan-only (prompted backlash) |
| OpenAI | Pay-per-click ads in ChatGPT |
| OpenAI | New $100/mo Pro tier (down from $200) |
| Lovable | ENT pricing: platform fee + credits |
| ServiceNow | 50% of net biz revenue from non-seat pricing |
| Canva | "AI Pass" monthly add-on with 40x more AI usage |
| SAP | Shift from SaaS to AI consumption pricing |
Data Labs Validation (May 14, 2026)
Same-day lint-sources run verified 126 companies against live pricing pages. Results validate — and sharpen — Poyar's findings:
ai_packaging drift confirms bundling wave. 11 companies had stale ai_packaging fields, almost all in the same direction: what was recorded as "add-on" is now "bundled." Notion, GitLab, PagerDuty, Asana, Deel, ChiliPiper — all bundled AI into base plans between H2 2025 and H1 2026. This isn't a sampling artifact. The industry moved, and companies that still sell AI as add-on are increasingly outliers.
Credit adoption tracks Poyar's 29% exactly. Our dataset shows credit-based or credit-hybrid models at Atlassian (Rovo credits), Figma (AI credits), HubSpot (Breeze credits), Salesforce (Flex Credits), Clay (token pricing), Notion (agent credits May 4), GitHub (Copilot credits June 1). These are the same names in Poyar's chart — not coincidence, this is the same wave viewed from two angles.
Price movement is real, not just packaging. Docusign raised all plans 10-20%. Atlassian Jira Premium up 14%. Clay dropped 10%. These aren't AI-driven changes — they're companies repricing while attention is on AI packaging. The cover of an AI transition makes traditional price increases less visible.
The 18 blocked companies reveal a data quality ceiling. 14% of our dataset (18/126) returned 403 or JS-rendered pages — unreachable by automated verification. This includes Salesforce, HubSpot, OpenAI, Canva, Zoom — exactly the largest, most-analyzed companies. The irony: the companies everyone writes about are the hardest to verify programmatically.
What This Means for Pricing Teams
The hybrid tax is real. Hybrid pricing solves expansion (only 13% complain vs 38% flat-fee) but creates an explanation problem (35% say "difficult to explain"). Companies adopting hybrid need to invest in pricing page clarity, billing transparency, and sales enablement — not just the model itself.
Credits are a transitional architecture. 29% today, projected 62% by H1 2027. But credits-as-commodity (Poyar: "the ultimate commodity metric") means they can't sustain differentiation. The companies that win will be those who make credits legible to buyers — credit calculators, usage dashboards, predictable bundles. Steven Forth (ValueIQ) noted in the LinkedIn discussion that credit calculators will become table stakes.
Expansion revenue is a product problem, not just a pricing problem. The data shows the fix isn't "switch to usage-based" — that just trades expansion for forecasting pain (37% complaint). The real lever is identifying upsell triggers: usage approaching limits, features gated by tier, credit depletion velocity. Structured data on these triggers doesn't exist yet.
AI margins compress the business model. At 50% target margins (vs 70-80% SaaS), AI features need 40-60% more revenue to deliver the same gross profit. Companies passing through inference costs (Autify's model-based credit pricing, Anthropic's unbundled API rates) are being transparent about this. Companies absorbing it (bundled AI) are subsidizing adoption from SaaS margins — viable short-term, unsustainable at scale.
Frameworks
CAMP framework for evaluating outcome-based pricing readiness (Poyar):
- Consistency of outcomes — can the AI reliably deliver the promised result?
- Attribution — can you prove the AI caused the outcome?
- Measurability — can both sides agree on what happened?
- Predictability — can the buyer forecast their spend?
Outcome-based is at 5% adoption. The gap between aspiration (26% investor preference) and reality (5% adoption) suggests CAMP failures are common. Most AI products fail on consistency and attribution today.