Cohort study for Saas

A cohort study is one of the best-known tools for calculating retention, but not everyone knows how to calculate them correctly. In this article, we will explain how to implement a cohort study for SaaS, with customer retention cohorts, user cohorts, and customer cohort analyses. We will also address the specific use case for a SaaS.

What is a cohort study?

A cohort is a group of customers acquired at a specific point in time; for example, the “September 2023” cohort assesses the behavior of customers acquired in September 2023.

These cohorts typically measure user and customer behavior, and what percentage of them remain in the subsequent months.

This analysis allows us to “easily” benchmark with other SaaS startups and helps estimate some specific business metrics, such as the Lifetime Value and other unit economics like the LTV/CAC ratio we mentioned earlier.

Common uses of cohort analysis for saas and Ecommerce

As we mentioned, the most common metrics used in SaaS cohort analysis are:

• Churn.
• MRR movements. Upgrade, downgrade…
• Conversion from user to customer.
• Retention rate, whether for users, customers, or MRR.
• Lifetime Value.
• Order repeat rate in an e-commerce…

But…

Why can cohort analysis help with Saas retention?

Let’s take an example.

Suppose we changed our software onboarding in January, and in February we noticed an increase in inactive users.

Why did 70% of users acquired in February become inactive the following week, while in January it was only 30%?

Is the new user onboarding launched at the end of January causing issues?

Cohort analysis can help us answer these questions and react promptly, making this analysis highly actionable and allowing us to identify a problem and take steps before it’s too late.

Why is the cohort study so crucial for SaaS retention?

When a SaaS is in its growth phase, it’s essential to identify and address pain points in the customer lifecycle. Cohort analyses allow us to do precisely this.

Instead of looking at high-level metrics for customer loss or retention, you can focus on specific months and tailor retention efforts accordingly.

Which of these statements is more “actionable”?

• “Our churn rate has increased by 3%”.
• “Churn is significantly higher in the second month of the customer’s lifecycle”.

In the second example, we can focus our efforts on understanding what’s happening in the second month, engage with the customer cohort at that time, and implement additional actions to retain them at this stage.

Note: Cohorts work best with monthly subscriptions (or shorter intervals). It’s essential not to mix annual subscriptions with monthly cohorts.

Apart from looking at the number of customers, we can also check:

• A cohort’s MRR retention, to see if churning customers are from bigger plans, primarily small accounts, or a mix of both. This could help us narrow down the reasons behind increased cancellations.
• We can also compare the same cohorts by plan to see if customers are being acquired more quickly in one of them.
• We can also see if they are more likely to cancel in a specific plan. This could help the sales and marketing team double down on promoting the better-performing plan.

Okay, Rubén… all this sounds great, but… how do I do it?

Let’s dive in 😉

How to build a customer cohort analysis

For cohort analysis, we need all the records of the following fields:

• Order id.
• User id or customer id.
• Plan id.
• Revenue plan.
• Order date.

It’s crucial to have a well-organized pricing plan to have consolidated information.

If we don’t have all the pre-configured fields, with order id, user id, and order date, we can start a simpler cohort analysis.

This tool is very handy for calculating your cohorts:

Build your free cohort analysis.

How to analyze a cohort analysis for churn calculation

Now that we’ve seen how to work with the data, let’s see how to analyze it.

The most common form of cohort analysis is presented as a table, with some standard features:

• Each row represents a user cohort, with the cohort’s name in the first column (e.g., “Feb 2014”).
• Each column represents a month after the cohort’s creation (with the zero month being the enrollment month).
• Each cell’s value is usually the churn or retention rate compared to the previous month.

Why are some cohort cells blank?

Because these cells refer to a point in the future. This kind of chart is extremely helpful for quickly identifying problematic months concerning customer loss or retention.

The color shading (green = low churn, red = high churn) immediately shows problem areas. In this example, the second month showed a churn spike in earlier cohorts but seems to have improved starting with the August 2014 cohort.

Other ways to visualize cohort analysis

You can also view a cohort analysis as a stacked line chart:

I hope you found this article on how to build a cohort analysis for SaaS helpful. If you’re unsure how to implement it or want to automate it, you can use our subscription and metric analysis tools.