SOFTWARE AS A SERVICE (SAAS)

Software as a Service is the delivery model in which software is hosted centrally and accessed via subscription rather than installed and owned by the customer.

It became the dominant enterprise software paradigm in the 2010s because it aligned incentives between vendor and customer: the vendor only gets paid as long as the customer finds value, creating structural pressure toward continuous improvement and customer success.

Key SaaS Metrics

The unit economics of SaaS are what made investors fall in love with the model. These are the numbers that matter:

• ARR (Annual Recurring Revenue): The annualized value of all active subscriptions — the primary top-line health metric for a SaaS business.
• MRR (Monthly Recurring Revenue): ARR divided by 12; useful for tracking growth velocity month-to-month.
• NRR (Net Revenue Retention): Revenue from existing customers at the end of a period divided by revenue from those same customers at the start — accounts for expansion, contraction, and churn. Above 100% means the business grows even without adding new customers.
• CAC (Customer Acquisition Cost): The total sales and marketing spend required to win one new customer.
• LTV:CAC (Lifetime Value to Customer Acquisition Cost): The ratio of a customer's total expected revenue contribution to the cost of acquiring them — the fundamental profitability signal for a SaaS model.

A healthy SaaS business typically shows NRR above 110%, CAC payback under 18 months, and gross margins above 70%. These benchmarks emerged from a decade of low-interest-rate financing — in the post-2022 interest rates environment, emphasis shifted from growth at all costs to efficient growth, making LTV:CAC and payback periods more scrutinized than ever.

AI's Impact on SaaS

AI is simultaneously an opportunity and a structural threat for SaaS businesses. The opportunity: AI features deepen product stickiness — a form of economic moats — justify pricing expansion, and create new upsell surfaces. The threat: AI agents and copilots may reduce the number of human seats required to accomplish the same workflows, directly compressing per-seat revenue.

The SaaS businesses best positioned for the AI transition are those with deep data moats and tight AI automation workflow integration — not those selling access to productivity surfaces that AI can replicate or bypass. Proprietary data accumulated through years of customer usage is becoming the durable competitive advantage that the software interface alone never was.

Pricing Model Evolution

The per-seat subscription model that defined SaaS is under pressure from usage-based pricing — charging for outcomes or consumption rather than licenses. AI accelerates this shift: when an AI agent does the work of ten human users, per-seat pricing breaks down as a value metric.

Companies like Snowflake, Twilio, and Stripe built on consumption pricing from the start. Legacy per-seat vendors are now grappling with a difficult choice: migrate their pricing model, cannibalizing near-term revenue, or defend it and risk competitive displacement by usage-based alternatives.

The transition to outcome-based pricing also changes the sales motion, the customer success model, and the revenue predictability that made SaaS attractive to investors in the first place — turning recurring revenue into something closer to variable revenue.

Open Questions

• Does usage-based pricing erode the predictable recurring revenue that justified high SaaS multiples, or does it unlock larger total contracts by removing seat-count friction?
• Which category of SaaS business — vertical software, horizontal productivity, infrastructure — is most exposed to AI agent substitution?
• If AI compresses headcount at customer companies, does the addressable market for per-seat SaaS shrink structurally, or does it expand as software reaches workers who previously had no tooling?
• Can legacy per-seat vendors transition their pricing model without triggering customer contract renegotiations that destroy near-term revenue?

Part of the knowledge graph at The Best Blog Ever — reference definitions for ideas that matter.