Track and grow your ARR →
Unlock CFO-grade clarity into your SaaS metrics with Equals.
The components of ARR tell you the direction of the trend at any given time: is it increasing, decreasing, or staying the same?
Cohort analyses help you understand how customers perform as they age and how that performance compares across different customer vintages i.e., cohorts.
OK, but what does that mean, and why is it important?
Cohort analyses are the most straightforward way to see if your business is acquiring sticky customers and whether or not those customers are increasing their spend over time. Demonstrating an ability to do both is the secret to unlocking sustained, durable growth.
This kind of enduring growth has been described in many ways. One of our favorites comes from Jeff Bezos’s annual shareholder letter in 2000.
Net Revenue Retention (NRR) shows how the relative total spend of a customer or set of customers has changed over time.
NRR analyses can be run on a cohorted or uncohorted basis, each requiring slightly different methodologies.
This “left–justified” view cohorts a set of customers based on their subscription start date (y-axis) and compares their total spend across time (x-axis) to their initial spend (Gross New ARR). The cohorts and aging are typically grouped by months, but as you have more data, it’s common to also group on a quarterly basis.
Here’s how to read the NRR of a given cohort compared to the starting period period (e.g., month one):
< 100% | 100% | > 100% |
---|---|---|
Relative spend decreased | Relative spend remained flat | Relative spend increased |
There are two things to be mindful of when looking at Cohorted NRR.
First, how retention compares across customer vintages at fixed age intervals. Said another way, are newer cohorts retaining worse, the same, or better than older cohorts at month 3/6/9? This tells you if the stickiness of your customer base is changing, which could come down to the quality of your product changing and/or the quality of a given cohort of customers.
Second, the slope of retention curves over time should be compared to see if the pace of retention changes with aging cohorts. This will help us understand point-in-time differences and whether lines across vintages converge or diverge over time.
A common case to consider is if you changed pricing or introduced annual contracts. Maybe your initial ACV is higher in more recent cohorts, which results in less relative expansion in the first ~3 months, but the lines converge by months 9-12 as the drop off from Month-to-Month (M2M) customers weighs down the legacy cohorts.
Your business is constantly changing as an early-stage startup. You can think of Cohorts as your map to understand how experiments impact results.
This is the more commonly discussed of the two calculations. It compares the total spend of a fixed set of customers over a fixed window of time, typically 12 months, to show how much ARR is retained. It’s important to note that the same set of customers are evaluated, meaning spending related to any Gross New customers acquired in the same fixed window (e.g., 12 months) is not included in the calculation.
The example below highlights the relevant cohorts used to calculate NRR in April 2023:
Following this methodology ensures the same customers are compared between the April 2023 and April 2022 periods.
Almost every public SaaS company reports on NRR in public filings, but be mindful when comparing against peers—we cover why later in Benchmarking ARR.
Gross Revenue Retention (GRR) is similar to NRR but only considers Churn and Contraction. Removing the additive revenue from Expansion and Restart gives a clearer picture of how much spend erosion takes place over time.
A “left-justified” view benchmarks customers based on their age, typically in months, and compares their spend at any given time to their initial Gross New ARR.
The “right-justified” view compares the spend of the entire customer base over a defined window—typically 12 months—to understand how much ARR a business should expect to lose from its existing customers.
The example below highlights the relevant cohorts used to calculate NRR in June 2023:
Following this methodology ensures the same customers are compared between the June 2023 and June 2022 periods.
Next topic