Recurring-revenue businesses revolve around a few core top-line metrics. Annual Recurring Revenue (ARR) is the annualized value of all recurring contracts at a point in time — the canonical top-line metric for SaaS unit economics. Monthly Recurring Revenue (MRR) is the monthly version, with the simple relationship \( ARR = MRR \times 12 \). Both can be analyzed on a cohort basis to expose acquisition-channel quality over time and to anchor all of the retention and growth metrics covered in earlier chapters.
Pricing model choice shapes everything downstream. Seat-based pricing charges customers per user or seat with a flat subscription, producing predictable revenue that is easy to forecast but misaligned with actual value consumption. Usage-based pricing charges customers per unit of consumption — API calls, gigabytes stored, messages sent, transactions, generated tokens — aligning revenue with both cost and value delivered but introducing volatility and harder forecasting. The unit in usage-based pricing is any countable resource that becomes the billable metric. Tiered pricing offers 2–4 product packages at distinct price points, typically Free, Pro, and Enterprise, to capture different willingness-to-pay segments. The trade-off is structural: seat-based is predictable, usage-based is aligned but volatile, and tiered is segmenting.
Freemium deserves special attention because free users are not free. Hosting, support, abuse, fraud, and dev time amortize per free user, often $0.10–$5/month depending on the product — a fully loaded cost that must be measured and tracked. The free tier's role is to drive top-of-funnel volume for CAC efficiency, but the blended LTV of the free cohort must exceed the CAC that funded them, including the cost of serving them over their lifespan. Conversion rate from free to paid is a small number — 1–5% is common — but multiplied by huge free-user volume drives paid growth. The break-even conversion rate is the minimum free-to-paid conversion at which the blended LTV of the free cohort equals the CAC; below this, growth destroys value regardless of how fast the funnel looks.
Variable costs in subscription businesses — hosting, payment processing, transaction fees, per-user support time — must be carefully tracked because they scale linearly with units. Payment processing fees, typically 2.5–3.5% per transaction charged by Stripe, Adyen, and others, are a variable cost that must be included in unit-economics math. Land-and-expand motions rely on expansion ARR — net new ARR added by existing customers in a period via upsell, cross-sell, or seat increases — to produce negative net churn, which is why usage curves (graphs of how a customer's usage changes over time) become central forecasting tools for predicting both expansion opportunity and churn risk well before either shows up in headline retention numbers.