Cohort Retention Curve
CRM & RetentionAlso: Retention Curve · Cohort Curve
Quick definition
A cohort retention curve is a chart that tracks what percentage of customers who signed up in the same period are still active over time. Instead of averaging all customers together, it groups them by their start date so you can see how a single group actually behaves as it ages.
How it varies across Australia
Australian subscription businesses typically see the steepest drop in the first month, then a flattening if the product has found real fit. Categories with weaker onboarding show curves that keep sliding well past month six instead of levelling out. The shape matters more than any single retention rate figure.
See retention benchmarks across Australian industries →What it actually means
Picture a bucket with a hole in it. Averaged churn rate tells you the water level dropped. A cohort retention curve tells you exactly when the water left and how fast the leak is slowing down, which is the difference between fixing the bucket and just topping it up forever.
The curve plots one signup group, say everyone who joined in March, and tracks what percentage of them are still active at week one, week four, month three, month six. Every business loses people early. What separates a healthy business from a leaky one is whether the curve flattens into a stable plateau or keeps sliding toward zero.
This is different from looking at churn rate or retention rate as a single number for the whole customer base. A blended number hides the fact that your March cohort might behave completely differently to your September cohort, especially if you changed your onboarding, pricing, or acquisition channel in between. Cohort analysis exposes those differences. It's also the clearest way to see whether lifetime value assumptions are grounded in reality or just optimistic maths.
A retention curve that never flattens isn't a retention problem. It's a product problem wearing a marketing costume.
How to calculate it
Retention at time T = Customers from cohort still active at T ÷ Total customers in cohort at signup
Worked example. 100 customers signed up in March. At month one, 62 are still active (62% retention). At month three, 41 are still active (41%). At month six, 38 are still active (38%). The steep drop between month one and three, followed by a flatter line into month six, is the shape you're looking for.
The Australian context
Australian SaaS and subscription businesses often build cohort curves off financial-year signup groups rather than calendar quarters, which makes sense internally but complicates comparison against global benchmark data published on calendar-year cohorts. When comparing your curve to an overseas benchmark, check which calendar convention it uses before assuming the shapes are directly comparable.
Where people get this wrong
Cohort Retention Curve vs Churn Rate
| Cohort Retention Curve | Churn Rate | |
|---|---|---|
| What it shows | How one signup group changes over time | A single point-in-time loss rate across all customers |
| Format | A curve or chart | A single percentage |
| Best for | Diagnosing when and why people leave | Reporting overall health in one number |
| Hides cohort differences? | No, exposes them | Yes, blends everything together |
Related terms
Common questions
What's the difference between a cohort retention curve and retention rate?
Retention rate is usually a single number for a fixed period. A cohort retention curve tracks how that number changes over the life of one specific signup group. The curve shows you the pattern behind the number, not just the number itself.
How many cohorts do I need before the curve is reliable?
You want enough customers per cohort that random churn doesn't distort the shape, generally at least a few dozen, and enough historical cohorts (three to six) to see whether the shape is consistent or improving over time.
What does a healthy cohort retention curve look like?
It drops in the early period as people who were never a good fit leave, then flattens into a plateau. A curve that keeps sliding downward with no flattening point signals an ongoing product or onboarding problem, not just normal early churn.
Can cohort curves predict lifetime value?
Yes, and more reliably than a blended churn rate. Once a cohort's curve flattens, you can project forward with more confidence, which is exactly the assumption most lifetime value calculations depend on getting right.
Debrief
Get the next one
No spam. No fluff. Just the next article, straight to your inbox.
Keep exploring
About New Rebellion
New Rebellion is a marketing intelligence consultancy. We build tools, score Australian businesses on how their marketing actually performs, and publish Debrief every day. This dictionary is part of how we work in the open.
How we think →