Once you’ve created a killer product and acquired your first clients, you’ve reached your first significant milestone. Even so, long-term business success does not just mean attracting someone to download an app or buy a subscription once, but rather getting them to keep returning to your product – increasing customer retention. Surely, you also want a stable, continuous flow of new users coming in. To achieve this, data analysis is crucial. It’s important to remember your users aren’t all the same, so your data analysis shouldn’t strive to treat them equally.

To boost your business’s performance, you need to go beyond vanity metrics like download counts, purchases, subscription numbers, etc. You need to dig deeper into your data and find the patterns of the best (or worst) performing target groups; in accomplishing this, retention cohort analysis will serve you best.

How does cohort analysis work? What are the different approaches to conducting cohort analysis? Why is it so important? Read on for more.

What is retention сohort analysis?

First things first. What is retention?

Customer retention rate is an indicator of how many customers stay with your business for the long term. It’s a percentage-based metric that shows the amount of users retained by the end of a given time period.

CRR formula

What are cohorts?

A cohort is a group of users with common profile traits, behaviors, or both, including:

  • Usage time
  • The number of goals completed
  • Users who own iOS devices
  • Users who own iOS devices, who logged in every day last week…

Put simply, retention сohort analysis is a tool for measuring customer retention over time. Its results will help you reveal how different groups behave over a set period of time, pinpoint their patterns, reveal which products or services to promote, and indicate what customers are the most valuable. Handling different cohorts with unique marketing strategies will greatly increase your retention rate and revenue.

Let’s polish up your knowledge with an example:

ARR Cumulative Cogort Analysis

Let’s say, we’re analyzing the retention rate for a language app. In the first column of the table below, we see how many people have downloaded the app. The remaining columns show what part of those users opened the app during the next six weeks. In this case, we can see the cohort retention rate has changed (actually decreased) from week to week: in the second week, only 4.47% of them returned to the app; after two weeks, 2.32% returned, and so on. 

Retention rate analysis
Retention rate analysis, example

Just at a glance, it’s clear: Houston, we have a problem. Therefore, the next step should be determining your cohorts, conducting a detailed analysis, and searching for patterns.

Types of retention cohort analysis

There are two common cohort types that can help you understand customers’ behavior. They are:

  1. Acquisition cohorts: Groups are split depending on when they signed up for your product.
  2. Behavioral cohorts: Groups are split depending on their behaviors and actions with your product.

Those are their basic definitions, but, that’s not enough to decide which to use for your project (spoiler: use both). Let’s take a closer look at these two types.

Acquisition cohorts

Acquisition cohorts are used to monitor new users and see how long they continue using the app after their first interaction – i.e., the length of your customer’s lifetime. The acquisition event covers purchasing a product, downloading an app, and signing up.

This type of cohort typically can give you answers to questions like Who’s buying the product? and When did they make the first purchase (or subscription)?

To estimate the success of a newly launched application, you can break down the number of users buying Premium subscriptions into cohorts by day or week. This allows you to see the number of people and the average period for in-app purchases, therefore revealing when the retention rate starts to drop. Is it after the first week or month? These numbers narrow down the causes that might cause customer churn (the percentage of your users who cancel or don’t renew their subscriptions during a given period).

Acquisition cohorts are great for showing you trends and telling you when churn rate declines.
Acquisition cohorts show numbers and statistics only. It can’t explain why your users are leaving. For that kind of information, you need to learn from another type of cohort: behavioral cohorts.

Behavioral cohorts

Behavioral cohorts are a custom group of people from your audience based on any combination of past behaviors or user profile features. With this, you’re able to monitor what people do or don’t do with your product through special event triggers you can learn to understand the behavior of demographically different users.

In behavioral cohorts, users are grouped depending on the actions they take in a given time frame after acquiring the product.

Behavioral cohorts help you understand how users engage with your product and what impact user engagement has on the retention rate.
Behavioral cohorts require you to be attentive and creative to identify the correct groups.

Why cohorts?

“In God we trust, everybody else brings data.”

W. Edwards Deming

The major advantage of cohort analysis is that it helps you understand your customers and therefore make better decisions. How does it do this? By showing you how users act over time.

Cohort analysis is crucial because metrics like daily or monthly active users (DAU and MAU) are distorted by growth. If your app is growing rapidly, new user signups will hide decreases in your existing users. It doesn’t matter how effective your acquisition channels are if you lose users faster than you attract new ones.

That’s why performing a retention cohort analysis is one of the most important methods to check and improve the health of your business.

Retention cohort analysis
Retention cohort analysis, example

The results of these analyses allow you to make specific decisions about your product. When you group your users into cohorts (e.g., by month or quarter of sign up), over time, you can see which groups have the highest retention rates or longest engagement times, among other valuable metrics. Once you pinpoint a group of users with high retention rates, you can dig into that group’s traits and behaviors to find patterns. A deeper knowledge of user patterns and cohort behaviors could significantly expand your ability to find key product metrics and hidden correlations in retention and acquisition. Furthermore, you can also make use of information about habits and trends you saw in prior cohorts.

But this technique opens up even more opportunities to improve your business, a handful of which we will discuss below:

Save marketing costs

Retention is cheaper than acquisition. The main reason for this significant difference in cost is that customers will buy from brands they trust. Over half of consumers have said they would spend up to 57% more on a company when they feel loyal to that company. According to Grant Thornton’s research, there is a “50/20” rule that says, on average, 50% of a company’s sales come from the top 20% of its users.

That’s why conducting a retention cohort analysis is one of the most effective methods to improve the health of your business. Loyal clients who make repeat purchases help a business build stable revenue while offsetting the costs of attracting new customers.

Help manage customer churn

Churn rate – the percentage of users who stop using a product within a given period – is best when it is as low as possible. Cohort analysis helps you decrease churn rates because it allows you to explore why groups or segments of users stop buying your products. If you can figure out what’s making people leave, you can avoid high customer churn rates.

In addition, you can test assumptions about the relationship between particular customer actions and churn rates (e.g., whether sign-ups related to specific promotions cause greater churn or how the exact channel’s churn rate grows over time).

Improve your customer LTV

Customer lifetime value (CLTV) is the total amount of money a user is expected to spend on your app or products during their lifetime. For example, if the price of a monthly subscription to the language app is $10 over an average lifetime of 16 months, the CLTV would be calculated as: $10 x 16 = $160.

Cohort analysis and CLTV allow you to estimate exactly how users in a group behaved over their lifetimes (by subscription type, size, age, etc.) and how much money a particular cohort brought you as revenue. After analyzing the overall data, certain evaluations may be true for some customers and wrong for others. Without this information, you’re prone to decisions that can cause significant harm to your business.

Optimize your conversion funnel

Which channels bring the best users?

Where should we invest more in the marketing budget, and where should we make cuts?

Conversion funnel analysis will give you answers to these questions. And the cohort analysis will help you improve your conversion funnel optimization.

It’s a good idea to compare customer groups who were acquired in different funnel stages and via different sales funnels, since this can help you analyze the significance of the marketing funnel to the output value of customers overall.

Such data is useful for realizing how different marketing strategies impact the conversion from the top of the funnel to the bottom.

Increase customer engagement

Vanity metrics, whether they’re the amount of the monthly users, app downloads, mentions, or page views – can be pleasant to look at, but they’re not always informative in the long run. If our aim is to grow the business through making better decisions about the products and services you make, there are better metrics more worthy of our attention.

When you learn how customers interact with different marketing tactics and certain features of your product or service, you can take further steps that will stimulate all your customers to take action (e.g., including cross-selling and upselling options in your business model).

As you can surely tell from the above descriptions, retention cohort analysis can give you a vision of the future development of the product. I also provides you with an understanding of long-term relationships with a certain user cohort and their influence on unit economics. This will also provide you with an effective tool for forecasting marketing budgets and targeting user structure, product weaknesses, and strengths, along with their projected influences on business.

Using retention cohort analysis in financial modeling

Retention cohort analysis can be extremely useful in financial modeling and future performance projection, as it shows how key metrics are related and how they influence future financial results.

Predicting the performance of each specific group allows you to handle them separately and properly.

Retention cohort analysis may be useful in analyzing the following metrics:

  • User acquisition
  • Churned users
  • User retention
  • LTV estimate
  • New/Lost MRR

Retention cohort analysis can help to define the future development of each metric and show how to manage them and unit economics. No type of cohort analysis is superior to another. Businesses should merge two or more cohorts to get a better understanding of how their products are being used.

As an example, we can look at forecasted cohorts for user retention, churned users, and retained Monthly recurring revenue (MRR). Individually, they are all less important, but together, they can give a broader understanding of mutual reliances and possible directions for further analysis.

Forecasted cohorts for user retention
Forecasted cohorts for user retention, churned users, and retained Monthly recurring revenue (MRR), example

From that simple analysis, we can discover how the user base and MRR are going to develop over time.

Factors that affect particular metrics are the monthly churn rate and price per one user. However, a lot of other drivers can be found if we dig deeper, such as marketing budgets and channels of new users acquisition, subscription plan diversification, possible improvements in churn rate and prices, month to roll out the product, and others.

Remember, when any of the mentioned variables are in a start-up financial model, they must be precise so they can be easily managed, altered, and analyzed separately.

How to read and interact with cohort analyses

If that was too much information all at once, let’s clear it up with a real-life example of how to read a cohort table. Take a look at how we, at Waveup, use retention cohorts projection in our models:

Retention cohorts projection
Retention cohorts projection, example

The cohort table shows app user retention during the first 12 months.

Key elements of the table:

  • Vertical axis (cohorts): new users
  • Horizontal axis (time period): a month
  • The numbers inside (development of the cohorts over time)

First, we choose which common features or types we analyze (user acquisition, time period, purchase time, subscription plan, etc.). Secondly, we identify key drivers and factors that influence the cohort (churn rate, average contract length, a time horizon of analysis, etc.). Then, we build separate cohorts for user retention metrics to analyze.

The first acquisition month was used to insert a historical user base (2,701,970 users).

The churn rate (3+ months) was identified based on historical data, but we can also use industry benchmarks or other applicable assumptions.

In addition, we have to account for other specific features of our product and industry. For instance, a particular cohort was built for app-based product analysis, and there is an additional increased churn rate for the first 1-2 months.

Finally, if we apply defined churn rates to the users acquired in each month, we will get the number of retained users, allowing us to observe how it will evolve in the future. As we see, cohort analysis can be extremely useful – not only for the standard historical analysis but also in making the forecast looking forward.

Furthermore, we can multiply retained user numbers by current/estimated prices and determine the ability of each cohort to generate revenue over time.

How to interpret the results of cohort analysis?

After we acquire the data, we have a few things to analyze:

Key elements of the table:

  • How the product or each segment (cohort) will develop in the future
  • What financial results it will bring with the current level of acquisition, retention, and churn

We hope this example clarified how useful it is to use the results of cohort analysis through the financial model, linking the retention metrics directly to cohorts. Such an approach is more solid and makes it possible to dig deeper into the numbers.

Retention cohort analysis in different businesses

As was mentioned before, different business models or industries may have their own specific features in their analyses. The features chosen can have a dramatic influence on retention cohort analysis, either on key metrics, segments, or key drivers.

Cohort analysis is extensively used in the following verticals:

SaaS app

A successful SaaS app has one important feature: stickiness. Since SaaS is a subscription-based business model, user retention is crucial for gaining profit. Cohort analysis is a reliable method for understanding potential reasons for churn and for discovering what features, offers, and pricing limits drive longer-term retention for SaaS companies. By looking at those metrics, you can tell if your product and marketing strategies are working well or not.

For SaaS companies, working with a customer means either growth or stagnation. When user engagement starts to decline, quickly running a cohort analysis can help avoid churn. The product team can pinpoint common traits and behaviors among less-active user groups and re-engage them later with new marketing campaigns or feature improvements based on specific customer behavior.

Cohort analysis allows you to learn the exact number of users you retain in each month of their user journey. Other important metrics to include:

  • LVP
  • Consecutive renewal periods
  • Different platforms’ performances (e.g., AppStore vs. GooglePlay)

For example, an analysis can compare AppStore and GooglePlay, showing that, after the price increase, the GooglePlay users stopped buying Premium accounts.


E-commerce companies can use cohort analysis to track customer behavior and see which products, demographics, and seasonal patterns most correspond with repeat purchases and higher lifetime value.

Conversion is usually the top metric of cohort analysis for e-ommerce brands. Typically, user groups who showed certain engagement at an earlier point – browsing specific pages, leaving the products in the shopping cart, or reading reviews – tend to buy specific items. But, to discover those exact behaviors for your e-commerce website or app, you must conduct a cohort analysis and research deeper if an interesting pattern appears.

Important metrics to include:

  • Purchase frequency
  • Repeat rate
  • AOV
  • CAC

With cohort analysis, you can clearly understand how much revenue customers are generating per certain time interval. Below is one example of how to analyze these numbers, showing how much revenue a cohort generates over time.


In a product-led growth (PLG) company, the product itself is your main way to engage, convert, and drive users to come back. Tracking the right PLG metrics will help you line up your product-led approach with broader business goals and make data-driven decisions about what to work on next.

Important metrics to include:

  • Time-to-value
  • Emphasis on behavior analysis

For instance, let’s say an expansion MRR cohort analysis shows the following:

MRR cohort analysis
MRR cohort analysis, example

The table above shows that users aged 25-45 upgrade their subscription plans more often than other age groups. By contrast, users over 45 years old used to downgrade their subscription plan.

Wrapping up

In the end, businesses are all about customer relationships. If you do not place customer satisfaction first when developing your product and services, you significantly decrease your chances of success.

Using retention cohort analysis in financial modeling can give you comprehensive knowledge about what works best for engaging, converting, and retaining customers. Find the most applicable combination of cohorts for your business, and plan the correct steps for your future growth.

Instead of jumping into big product changes, A/B tests on your problem cohorts will give you data to understand what works and what doesn’t. That way, you can make data-backed changes that are likely to reduce the churn rate and increase the retention rate. Once you’ve successfully improved your retention rate, all that’s left is to upgrade and repeat.

3 posts


Senior Analyst

With 5+ years of total experience, first at financial advisory then at venture builder companies, my current activities are focused on the financial modelling and consulting for startups. Here at Waveup, I write my articles to share the experience I have obtained. I hope that you will find them interesting and helpful!