Cohort Analysis - the best tool for tracking customer retention

We all know that getting new customers is vital to keep the business growing. What’s equally important is retaining the customers you already have acquired by using your valuable time and resources. But to improve your customer retention, you need to know how to measure it. Meet the cohort analysis! We’ll go through what the cohort is all about and how you can use it to improve customer experience and sales. At the end of the article, you will find more thorough instructions on how to use the cohort analysis in Custobar.

Cohort Analysis

What is a Cohort analysis?

Cohort analysis is the best way to track customer retention. It shows you how many customers are left at the end of each month after they initially purchased from you or were active in another way, for example, signed up for your loyalty program. In other words, it’s a tool you can use to observe and measure the behaviour of your customers over time.

How to read the cohort analysis?

The cohort analysis can be a powerful tool, but for that, one needs to be able to understand what they are looking at.

Let’s take the example here to find out how to read it.


Those two highlighted numbers can be read as follow:

Out of the 128 customers who’s first Event was on April 2019, 2.3% are still having events 12 months later.

Out of the 81 customers who’s first Event was on February 2020, 32.1% did not have an Event beyond the first month and have probably churned.

Please note: Those figures are updated in real-time based on the available data in Custobar, and will update every time a customer has a Sale or Event. Also, the total of a row is higher than 100%. The reason for that is a customer who has not churned and is still engaging during month 12 will also be counted as “on board” during months 1 through 11 as well since they have not churned during those months.

What’s the benefit?

Good customer retention is a critical success factor. It is more cost-effective to take care of the ones who already trust you i.e. your current customers than attract new ones with big marketing budgets. You already have data-based knowledge about your customers, which makes them less costly to serve. Loyal customers are also less likely to turn to your competitors. The cohort analysis helps you to analyze the behaviour of your customers and act at the right time. You can act based on facts, not guesses. Is there a specific time after you start losing your customers? What can you do better at that pain point to keep them?

You can track whether your marketing campaigns are worth the money in the long run. For example, if you have a Black Friday campaign and promote a loss leader (a product sold at loss to attract new customers), the campaign might first cause you more costs than profits. However, with the cohort analysis, you can track whether these customers from the Black Friday campaign will buy from you also in the future. Using Custobar audiences you can compare different campaigns with each other and see which campaign brought you more loyal customers. In other words, cohort analysis is a tool that helps you to experiment with your campaigns and find what works best for you. It will also help to prove the effectiveness of your marketing campaign.

Cohort analysis in Custobar

In Custobar, the cohort analysis breaks down the entire customer base or a selected audience in groups based on the time of their first purchase or activity event and presents the retention rate of those groups based on the time passed in months since that first action.

The cohort analysis can be displayed for the entire customer base, or for a saved audience. Audiences pinned to the dashboard will be showing as an option in the drop-down menu:

Cohort Audience

But it is also possible to access the cohort analysis of all other audiences from the audience folders:

Audience folders

The cohort analysis can be configured to break down the audience based on the time since the first Sale, or the time since the first Event. In the case of Events, only activity events count. Those can be MAIL SUBSCRIBE, BROWSE, MAIL OPEN, etc.

The cohort can also be set to display the breakdown in percentage, or the actual customer counts for each row.

Absolute numbers

If you have any questions or concerns, please contact


  1. Coussement, Kristof. “Improving Customer Retention Management through Cost-Sensitive Learning.” European Journal of Marketing 48.3/4 (2014): 477–495. Web.
  2. Launch Day by Experience Commerce Finland
Back to top