glossary · strategy
Product monetization definition
The set of decisions that determine how a product captures a share of the value it creates: which customers pay, for what unit, at what price, and under what conditions.
Monetization is a product design decision, not a finance one. It determines which customers pay, for what unit, at what price, and under what conditions, and those choices shape user behavior as much as any feature does. A PM who treats monetization as a downstream pricing call has already made a mistake: the model you choose constrains what users feel safe doing with your product.
The three models that matter
Subscription. A flat recurring fee for access. Predictable revenue, easy to budget for enterprise buyers, and it hides usage cost risk on your side. The downside: subscription hides dissatisfaction too. A customer who barely uses the product keeps paying until renewal, which masks churn signals until they arrive as a surprise.
Usage-based (consumption). Customers pay for what they use: API calls, seats active in a period, messages sent, tokens consumed. This aligns the bill with value delivered, and it drives expansion revenue as usage grows. The product risk is bill anxiety. When users can see a cost per action, they hesitate. For AI products specifically, usage-based pricing at the prompt level suppresses experimentation and discovery, the exact behaviors that generate habit and retention. Stripe, Snowflake, Twilio, and AWS all use consumption pricing because their value scales directly with usage volume and their customers are businesses that can model the cost. That context matters when you’re choosing.
Outcome-based. Customers pay for results: per resolved ticket, per matter closed, per hire made. Harvey charges per matter, not per prompt. Highest alignment (vendor and customer incentives match), but it requires the PM to define the outcome upfront, instrument it, and defend the measurement when the model is wrong. For agentic AI, where a hallucination can void the “outcome,” that’s a product scoping problem before it’s a contract term.
Most mature products blend: a predictable subscription floor that covers enterprise budget cycles, plus variable usage billing that captures upside. OpenAI runs this playbook: ChatGPT Plus at a fixed monthly fee creates a consumer revenue floor, and API pricing captures developer and enterprise consumption above it.
Unit economics in the interview
Healthy SaaS targets LTV:CAC of 3:1 with payback under 12 months. For AI products, the real unit is LTV:(CAC + inference cost per customer lifetime). If your inference cost per active user is $8/month and your subscription is $10/month, you have $2 of gross margin to cover acquisition, support, and overhead. That math fails at scale. Interviewers in 2026 expect you to name this number and know whether your model survives it.
Cohort matters too. Slack’s enterprise customers carry 5-7x the LTV of SMB. A single model applied uniformly across segments destroys margin in one cohort and leaves value on the table in another.
What “how should this product monetize?” actually tests
The interviewer wants to see whether you can identify a value metric: the unit that scales with customer value. “Per active seat” means you’re selling access to a tool. “Per API call” means you’re selling infrastructure capacity. “Per resolved ticket” means you’re selling outcomes. Naming the value metric is the first move; everything else follows from it.
Strong answer structure: name the value metric, select the model that aligns it with how the product creates value, check the LTV:CAC math including inference cost for AI products, name the biggest behavioral risk (bill anxiety, churn masking, measurement disputes), propose how you’d validate the choice, and name your north star: net revenue retention for subscription, expansion revenue per cohort for usage-based, cost to measure and verify for outcome-based.
Weak answers list model names without a value metric and pick one without justifying it against unit economics or user behavior. Interviewers call this model memorization.
The 2026 shift
Feasibility is cheap now. That moves the hard problem from “can we build this?” to “what outcome will a customer pay for repeatedly, and do our unit economics survive the inference cost of delivering it?” Monetization is a constraint that shapes what you scope and what you promise. Viable (someone pays, at margin, repeatedly, at scale) is what separates a product from a demo.
Related: LTV, CAC, product-led growth, proving viability, LLM unit economics, price an AI product.