SaaS metrics: the six that matter and the ones that flatter
Recurring revenue, net retention, gross margin, payback period, activation rate and churn by cohort. Everything else is either derived from these or designed to look good.
Key takeaways
- Recurring revenue, net retention, gross margin, payback period, activation and cohort churn describe the business.
- Net revenue retention above 100 per cent is the most informative single indicator of product-market fit.
- Separate involuntary churn from failed payments; it is a fixable operational problem worth real revenue.
- For an early product, activation rate comes first; poor activation makes every later metric look like churn.
Six numbers describe the health of a software business: monthly recurring revenue and its growth, net revenue retention, gross margin, customer acquisition payback period, activation rate, and churn measured by cohort. Almost every other metric is either derived from these or exists to make a deck look better than the business is.
The six
- Recurring revenue and growth rate. Contracted, repeating revenue only. One-off implementation fees are not recurring revenue and including them is the most common way early figures mislead.
- Net revenue retention. What last year's customers pay this year, including expansion, contraction and churn. Above 100 per cent means the business grows without new customers, which is the single most informative number about product-market fit.
- Gross margin. Revenue minus the cost of delivering the service: infrastructure, third-party services, and support. Software businesses with margins in the fifties are usually service businesses that have not admitted it.
- Payback period. Months to recover what it cost to acquire a customer. Under twelve is comfortable, over eighteen means growth consumes cash faster than it produces it.
- Activation rate. The share of signups reaching the first moment of real value. It is the leading indicator of every retention number that follows.
- Churn by cohort. Aggregate churn hides that customers acquired in one quarter behave completely differently from another. Cohorts show whether the product is improving.
What flatters
Total registered users counts people who signed up once and never returned. Page views and sessions measure nothing about value delivered. Annual contract value averaged across everyone hides that two customers carry the business. And any metric quoted without a denominator, particularly growth expressed only as a percentage from a small base, is a presentation choice rather than a measurement.
Measuring churn honestly
Count both customer churn and revenue churn, since losing ten small customers and one large one are very different events that look similar in a single figure. Separate voluntary churn from involuntary churn caused by failed payments, because the second is a fixable operational problem often worth several per cent of revenue. And measure churn at the cohort level over time rather than as a monthly rate, which smooths away the pattern you need to see.
The number to watch first
For an early product, activation rate. If people sign up and never reach the point where the product does something useful for them, no amount of acquisition spending will produce a business, and every downstream metric will be poor for a reason that looks like churn but is actually onboarding. Fix activation before spending on growth.
Fix activation before spending on acquisition. Otherwise you are buying people to disappoint.