Monetizing AI — Alvaro Morales, Orb

Key Concepts:

• Direct vs. Indirect Monetization
• Value Metric Selection (Resource-based, Proxy-based, Outcome-based)
• Continual Pricing Experimentation and Evolution
• Orb Simulations (Backtesting Pricing Strategies)

1. Should You Monetize AI? (Direct vs. Indirect Monetization)

• Main Point: The initial framework focuses on deciding whether to monetize AI features directly (as a standalone product or add-on) or indirectly (driving upsells or bundling).
• Framework: Simon Kutcher's framework is used to determine if monetization should be direct or indirect.
• Examples:
  • GitHub Copilot: Launched as a separable, monetizable add-on. This is suitable when the new feature doesn't add value to the entire audience.
  • Notion AI: Initially launched as an add-on, but later bundled into business and enterprise tiers with price increases to encourage adoption at higher tiers.
  • Expedia: Launched an AI-powered feature (turning Instagram reels into bookable trips) for free, absorbing costs to incentivize more travel bookings. This is an example of indirect monetization.

2. Value Metric Selection

• Main Point: Selecting the correct value metric is crucial for AI pricing. The spectrum ranges from resource-based (token-based) to outcome-based pricing.
• Spectrum:
  • Resource-Based (Token-Based): Pricing based on granular capabilities (e.g., tokens).
  • Proxy-Based: Pricing based on proxies of value, such as steps in a workflow or entire workflows.
  • Outcome-Based: Pricing based on achieving specific outcomes.
• Examples:
  • Vercel's V0: Uses a token-based model, aligning with developers who want to understand the capabilities they are using.
  • Zapier: Uses task-based pricing, where tasks in a workflow (Zap) are the unit of value.
  • Intercom's Finn: Charges 99 cents per successful customer support ticket resolution (outcome-based).
• Outcome-Based Pricing Challenges: Requires alignment between customer and vendor on the definition of the outcome and objective measurement. It's easier in customer support (e.g., ticket resolution) but harder in creative or less directly measurable areas.
• Quote: "Outcome-based pricing is not a proxy for ROI. It's literally ROI."

3. Continual Pricing Experimentation and Evolution

• Main Point: Pricing strategy needs continual evolution and experimentation, especially with the rapid changes in AI (e.g., model cost changes).
• Argument: The traditional SAS model of infrequent pricing updates (every 1-3 years) is no longer sufficient for AI due to rapid changes in model costs and capabilities.
• Emphasis: The entire pricing experience (feature packaging, rate limits, custom terms, communication) contributes to the alignment of price to value.

4. Orb Simulations: A Tool for AI Pricing

• Context: Orb is a billing platform for AI and SaaS companies. Customers were using Orb to simulate pricing structures without actually billing.
• Orb Simulations: A product designed to backtest alternate pricing strategies on top of product usage data.
• Functionality:
  • Define a simulation period and a cohort of customers.
  • Build scenarios for alternate pricing models (e.g., add-on, token-based, tiered).
  • Generate reports showing the impact on topline revenue, the lowest average change to existing customers, and the revenue mix.
  • Provides a detailed view of revenue impact and percentage change for individual customers.
• Goal: To help teams make data-informed pricing decisions and reduce uncertainty.
• Process:
  1. Define Simulation: Set the time period and customer cohort.
  2. Build Scenarios: Create alternate pricing strategies (e.g., fixed fee, token-based).
  3. Run Simulation: Generate a report with insights on revenue impact and customer changes.
• Quote: "AI builders should always simulate first before putting something out to market."

5. Conclusion

• Main Takeaway: Finding the right AI pricing strategy is challenging but crucial for revenue and reaching the right audience.
• Key Steps:
  1. Determine whether to monetize directly or indirectly.
  2. Select a value metric that aligns with the perceived value and desired customer behavior.
  3. Continuously experiment and evolve pricing strategies.
  4. Use tools like Orb Simulations to make data-informed decisions.
• Final Note: Achieving the right pricing strategy can lead to significant revenue results.