# How to Read a Cohort Retention Chart

> How to read a SaaS cohort retention chart, including rows, columns, triangle shape, curve flattening, comparing cohorts and common interpretation mistakes.

Source: https://comppound.com/blog/how-to-read-a-cohort-retention-chart (published 2026-08-24)

A cohort retention chart shows how groups of customers behave over time. Each cohort is a group of customers who started during the same period, such as everyone who signed up in January. The chart tracks what share of each cohort is still active, or still paying, in each period after they started.

Cohort charts are one of the most useful tools a SaaS team has, because they answer questions that a single churn number cannot. Is retention getting better or worse? When do customers tend to leave? Do customers who stay past the first few months keep staying? Did the onboarding changes we made in spring actually help? This article explains how to read a cohort chart, what patterns to look for, the difference between customer and revenue retention, and common mistakes in interpreting them.

As an example, we will use a SaaS company selling bookkeeping software to freelancers and small agencies. It bills monthly and has been tracking cohorts since it launched its paid plans.

## The basic layout

A cohort chart is usually a table, often shaded like a heat map so that higher retention appears darker. Here is a simplified example with illustrative figures.

| Cohort | Month 0 | Month 1 | Month 2 | Month 3 |
|---|---|---|---|---|
| January | 100% | 82% | 74% | 70% |
| February | 100% | 85% | 78% | |
| March | 100% | 88% | | |

Each row is a cohort, the group of customers who started in a given month. Each column is time since the cohort started, measured in months in this case. Month 0 is the starting point, and every cohort begins at 100 percent because, by definition, everyone in the cohort is present at the start. The values in the table are the percentage of the cohort retained in each later month. Retention can be measured in customers, meaning the share of customers still active or paying, or in revenue, meaning the share of the cohort's starting revenue still being collected.

## Why the chart looks like a triangle

Newer cohorts have less history than older ones. In the example, the March cohort has only been around long enough to have a Month 1 value; its Month 2 and Month 3 cells are empty because those months have not happened yet. As a result, the filled cells form a triangle, with the oldest cohort at the top having the longest row and the newest cohort at the bottom having the shortest.

This shape is normal. It just means that comparisons at later ages rely on fewer cohorts, which is worth remembering when drawing conclusions.

## Read across rows to follow one cohort

Reading across a row shows how a single cohort retains over time. In the example, the January cohort starts at 100 percent, drops to 82 percent after one month, 74 percent after two and 70 percent after three. That row tells the story of January's customers: nearly a fifth left within the first month, and the losses slowed after that.

Reading rows helps you understand the shape of retention for a typical cohort. Where is the biggest drop? How quickly does the decline slow? For the bookkeeping company, the biggest drop happens between Month 0 and Month 1, which suggests that many customers decide quickly whether the product is for them.

## Read down columns to compare cohorts at the same age

Reading down a column compares different cohorts at the same point in their life. In the example, Month 1 retention rises from 82 percent for January to 85 percent for February and 88 percent for March. That is a meaningful pattern: newer cohorts are retaining better in their first month than older ones did.

Column comparisons are how you see whether retention is improving over time, and they are the best way to judge the effect of changes. If the bookkeeping company redesigned its onboarding in February, the improvement in Month 1 retention for the February and March cohorts would be consistent with that change helping. It does not prove it, since other things may have changed too, but it is far stronger evidence than a blended churn rate that mixes old and new customers.

## Look for the curve to flatten

Healthy SaaS retention curves usually drop early and then flatten. Customers who are a poor fit, or who never got set up properly, leave in the first few months. Customers who stay past that point have found lasting value, and they tend to keep staying. On the chart, this shows up as retention values that fall quickly in the first few columns and then level off, changing only slightly from month to month.

Flattening is one of the most important things to look for. If a cohort's retention keeps declining steadily, month after month, without leveling off, customers may not be finding lasting value at all. Eventually, every cohort trends toward zero, and growth depends entirely on acquiring new customers faster than old ones leave. That is a much harder business to run.

For the bookkeeping company, the January cohort's decline slows noticeably after Month 2, dropping only four points between Month 2 and Month 3. If that pattern holds in later months, it suggests the product has a stable core of long-term users.

## Customer retention versus revenue retention

Cohort charts can measure customers or revenue, and the two can look quite different.

Customer retention counts how many customers from the cohort remain. It treats every customer equally, so it reflects how broadly the product retains its users.

Revenue retention measures how much of the cohort's starting revenue remains. It captures downgrades and, if you choose, expansion. A cohort's revenue retention can actually exceed 100 percent if customers who stay expand their spending enough to more than offset those who leave or downgrade. Charts that include expansion are often called net revenue retention cohorts.

Looking at both is valuable. If customer retention is declining but revenue retention is stable or rising, the customers who leave are mostly small while the ones who stay are growing. If revenue retention falls faster than customer retention, larger customers or downgrades are driving the losses.

## Segment your cohorts

A single cohort chart for all customers can hide important differences. Build separate charts for different segments: plans, acquisition channels, company sizes, industries or use cases. The bookkeeping company might find that agency customers retain far better than solo freelancers, or that customers acquired through referrals retain better than those acquired through paid ads. Those differences can guide decisions about which customers to target, how to price and where to spend on acquisition.

## Watch out for small cohorts

When cohorts are small, retention percentages can swing a lot because of just a few customers. A cohort of twenty customers losing three in a month shows a 15 percent drop; losing one more makes it 20 percent. Be cautious about drawing conclusions from small cohorts. Consider grouping by quarter instead of month if monthly cohorts are too small, and look for patterns that persist across several cohorts rather than reacting to a single unusual one.

## Define "retained" carefully

The meaning of a cohort chart depends on what counts as retained. For paid retention, a customer might count as retained if they are still paying at the end of the period. For usage retention, they might count as retained if they performed a core action during the period. The two can tell different stories. A customer might still be paying but no longer using the product, which is a warning sign of future churn that paid retention alone will not show.

Choose a definition that matches the question you are asking, document it and keep it consistent. Changing the definition partway through makes cohorts incomparable.

## Use cohort charts to make decisions

Cohort charts are most valuable when they drive action. If the biggest drop happens in the first month, focus on onboarding and early value. If retention flattens late or not at all, look at long-term value, feature adoption and customer success. If certain segments retain much better than others, consider focusing acquisition on those segments. If newer cohorts are retaining better, look at what changed and do more of it. If they are retaining worse, look at what changed in acquisition or product and investigate. Changes in acquisition mix often show up here: a shift toward channels that bring better-fit customers, such as high-intent search traffic built through an [SEO growth program](/services/seo-growth), tends to appear as stronger early retention in newer cohorts.

## Avoid the common mistakes

The mistakes in reading cohort charts are consistent. People compare cohorts at different ages, such as an old cohort at Month 6 with a new cohort at Month 1. They overreact to small cohorts. They ignore the difference between customer and revenue retention. They mix definitions of retention. They look only at blended charts without segmentation. And they treat cohort charts as a reporting exercise rather than a guide to decisions.

## Bringing it together

A cohort retention chart groups customers by when they started and tracks what share remain over time. Read across rows to see how a single cohort behaves. Read down columns to compare cohorts at the same age and see whether retention is improving. Look for the curve to flatten, compare customer and revenue retention, segment where it helps, be careful with small cohorts and define retention clearly. For the bookkeeping company, a few minutes with its cohort chart shows that onboarding changes are improving first-month retention and that long-term customers are stable, which tells the team exactly where to keep investing.

## Frequently asked questions

### What is a cohort retention chart?

A table showing how groups of customers who started in the same period retain over time.

### Why do cohort charts look like triangles?

Newer cohorts have less time history, so later periods are empty.

### What does a flattening retention curve mean?

It suggests customers who stay past early periods tend to remain long term.

## About Comppound

Comppound runs SEO and AI search for B2B SaaS companies: a person sets the plan and reviews samples and every high-risk page, and AI speeds up production.

- [See the comparisons](https://comppound.com/compare): Weighing who should run your SEO?
- [Book a call](https://comppound.com/contact): Most of this works better when buyers can find you first