The Economics of Pricing Leverage
In their classic pricing study published in the Harvard Business Review, McKinsey & Company analyzed 1,000 global companies and revealed a startling economic truth: a 1% improvement in price yields an 11.1% increase in operating profits, assuming sales volume remains constant. By contrast, a 1% reduction in fixed costs improves profits by only 2.3%, and a 1% increase in customer acquisition volume improves profits by just 3.3%.
Despite this massive leverage, most software companies spend less than 10 hours per year evaluating their pricing architecture. They copy competitors, guess round numbers, and leave millions of dollars of enterprise consumer surplus on the table.
Comparative Framework: The 3 Core Pricing Archetypes
| Model | How It Works | Primary Advantage | Critical Risk |
|---|---|---|---|
| Per-Seat / User Licensing | Fixed monthly fee per employee account | Predictable recurring monthly revenue (MRR) | Customers share logins or restrict rollout to save budget |
| Pure Usage / Consumption | Billed per gigabyte, API call, token, or query | Zero friction to start; scales directly with customer usage | Revenue volatility; quarterly budget unpredictability for enterprises |
| Hybrid Value-Tiered | Base platform annual fee + usage tiers & SLAs | High predictable floor + uncapped expansion upside | Requires sophisticated metering & billing infrastructure |
Why the AI Era is Killing the Pure Per-Seat Model
For two decades, software margins were insulated by seat-based models. A customer buying 500 Salesforce licenses paid for 500 humans using the tool. However, modern autonomous AI agents and automated workflows do the work of human operators.
If an enterprise deploys an AI agent that accomplishes the output of ten analysts, a per-seat software vendor would see their licensing revenue collapse from 10 seats to 1 seat, even though the enterprise extracted 10x more productivity! Forward-thinking SaaS providers are rapidly migrating toward Outcome-Based and Workload-Based Pricing, charging based on completed workflows, resolved customer tickets, or compute transactions rather than human heads.