Why Consumption Pricing Makes Forecasting Harder with Devavrat Shah

Why It Matters

AI‑enabled consumption forecasting turns opaque usage data into reliable revenue insights, allowing firms to price, allocate resources, and scale usage‑based models with confidence.

Key Takeaways

• Consumption pricing simplifies overall revenue forecasting but complicates regional sales predictions
• AI can aggregate cohort data to improve granular consumption forecasts
• Token‑based API pricing mirrors traditional consumption models, shifting forecasting challenges
• Sales reps face “smooth volume vs. umbrella” forecasting dichotomy
• Trust‑focused AI tools reduce finger‑pointing in forecasting disagreements

Summary

The Revenue Builders podcast revisits a conversation with MIT professor Devavrat Shah, who examines how consumption‑based pricing—exemplified by token‑based API calls—reshapes revenue forecasting.

Shah notes that while total consumption volume is readily modeled, predicting where that usage will occur—by region, sales rep, or channel—creates a “smooth water versus umbrella” problem. He likens it to a manufacturer receiving irregular, large orders that aggregate into a predictable quarterly total but remain opaque at the individual level.

He argues AI can bridge the gap by treating sales reps and channels as cohorts, learning patterns across them, and integrating data from CRM, metering, and booking systems. Co‑host John adds that seasonality, new bookings, and customer‑specific cycles further complicate forecasts, reinforcing the need for advanced analytics.

For CFOs and go‑to‑market leaders, AI‑driven consumption forecasting promises more accurate revenue planning, reduced internal blame‑games, and a clearer path to scaling usage‑based business models.

Original Description

Consumption pricing puts pressure on the forecast in places traditional SaaS models rarely exposed. Total usage may be easier to model from the CFO’s seat, but the field still has to answer harder questions: which customer, which channel, which rep, and when. In this segment we’re revisiting this week, Devavrat Shah explains how AI can help teams learn across cohorts, spot patterns in uneven data, and create more trust in a forecast that would otherwise depend on isolated judgment calls.

Devavrat Shah is an MIT professor, director of MIT’s Statistics and Data Science Center, and co-founder and CEO of Ikigai Labs. He brings a data science and operator’s perspective to forecasting, consumption pricing, and enterprise AI.

🔗 Connect with Devavrat:

➡️ LinkedIn - https://www.linkedin.com/in/devavrat-shah-63b59a2/

🎙️ Catch the full episode here:

➡️ Understanding AI Through History and Practical Application with Devavrat Shah - https://www.forcemanagement.com/understanding-ai-through-history-and-practical-application

🎙️ ABOUT REVENUE BUILDERS PODCAST

Hosted by five-time CRO John McMahon and Force Management Co-Founder John Kaplan, the Revenue Builders podcast goes behind the scenes with the sales leaders who have been there, done that, and seen the results.

This show is brought to you by Force Management. We help companies improve sales performance, executing their growth strategy at the point of sale.

Connect with Us:

➡️ LinkedIn - https://www.linkedin.com/company/revenue-builders-podcast/

➡️ YouTube - https://www.youtube.com/@UCxS34OZ40_RgmyteJiTFBHQ

➡️ Force Management - https://www.forcemanagement.com/