Net retention convexity in SaaS - by ned - cranberry blog
In this post, I assume the reader has limited prior exposure to net retention but is somehow still interested in the metric. As a consequence, this post is probably going to annoy everyone.
I use this piece to motivate a SaaS metric of my own invention, net retention convexity, which I hope fills a void of context in interpreting net retention.
What context do I think is missing from net retention? I frequently hear folks in the industry interpret net retention incorrectly. Sometimes they interpret net retention incorrectly while writing very large checks and without realizing that they’re making a mistake. I hope we can all agree that this is suboptimal.
The most common error is to project trailing net retention as the underpinning for some forward-looking statement. For example, you’ll most commonly hear someone say “This business has 141% net retention. That means if they didn’t acquire any new customers, they’d still grow topline by 41%.” This result does not logically follow, and it is frequently incorrect in practice.
So why is this interpretation incorrect? Because it ignores two important details — first, the representative expansion path of a customer cohort; and second, the changing cohort composition of the business over time. Retention convexity helps check these dynamics.
