Data, Metrics & Experimentation
MRR, churn rate, LTV, CAC, NRR — the metrics framework and experimentation playbook for subscription businesses.
Most subscription businesses measure the right things the wrong way. MRR that includes annual contracts divided by twelve. LTV calculated from blended churn rates that mask cohort deterioration. CAC that counts marketing spend but ignores sales, onboarding, and infrastructure costs. The numbers look professional. The decisions they drive are wrong.
This chapter is the measurement and experimentation toolkit for subscription operators. Ross Williams starts with the uncomfortable premise that your current metrics are probably calculated incorrectly, then walks through each critical subscription metric with the precision required to make reliable decisions.
The metrics section covers MRR, ARR, ARPU, churn rate (gross and net), LTV, CAC, CAC payback period, and NRR. For each metric, Ross provides the correct calculation methodology, the most common errors he has seen (and made), and the specific decisions each metric should and should not be used to drive.
Cohort analysis gets a detailed standalone section. Ross teaches you to build and read cohort retention tables, spot the patterns that indicate whether your business is genuinely improving or just growing its way past deteriorating fundamentals, and use cohort data to forecast future revenue with meaningful accuracy.
The experimentation section covers A/B testing methodology for subscription businesses, including sample size requirements, test duration, and the specific pitfalls of testing in recurring revenue environments. Ross introduces the ICE prioritisation framework (Impact, Confidence, Ease) for deciding which experiments to run first.
Want to read this chapter in full?
Pre-register now to be first in line when Subscribe & Conquer launches.
Subscription operators and data teams who want to ensure their metrics are calculated correctly and their experiments are designed to produce reliable results. Especially valuable if you have never audited your metric definitions or if you run experiments without clear methodology.
The chapter provides the precise methodology, including how to handle annual contracts, discounts, trials, and mid-month changes. The most common error is dividing annual contract revenue by twelve and adding it to monthly revenue.
ICE stands for Impact, Confidence, and Ease. It is a prioritisation framework for deciding which experiments to run first, scoring each potential test on these three dimensions.
Longer than you think. The chapter explains why subscription A/B tests require extended durations because the true revenue impact of a change may not be visible for weeks or months.
Because blended averages can hide deteriorating performance. A growing subscriber base can mask the fact that recent cohorts are retaining worse than earlier ones. Cohort analysis reveals whether your business is genuinely improving or just outrunning its problems.
Ready to build unstoppable recurring revenue?
Pre-register now to be first in line when Subscribe & Conquer launches.
Or explore another chapter →