All CollectionsGetting StartedRetention
Understanding Cohort Analysis
Understanding Cohort Analysis

The cohort tables within your Retention tab measure your customer retention over time.

Written by Abby Garland
Updated over a week ago

# Introduction

A cohort is a group of customers sharing a specific characteristic over a period of time. With cohort analysis, you can identify common behaviors on a specific group of people (cohorts). With these new identified behaviors, you can adjust your communications, product offers, promotions.

The cohort tables within your Retention tab measure your customer retention over time, rather than for individual months. You can utilize your cohort tables to answer the following questions:

• What’s my retention rate 6 months after the first purchase?

• How is my retention improving over time?

The video below details how to understand and utilize the cohort tables within your Retention tab.

# How-to: Read the Cohort Tables

• Keep in mind that Polar calculates each cohort as the group of customers that purchased for the first time within 30 days of a given month (starting from the date they made their first purchase). See the section below for an example.

• Horizontally (in red)

• Let's take customers by month for August 2021. In August 2021, there were 2 129 new customers acquired. 3 months after (in November 2021), 4,09% of customers who bought for the first time in August did another purchase in November 2021. If you look at non-cumulative data, it will just tell on that month the percentage of customers that came back that month. If a customer buys twice for example, in November 2021 and in January of 2022, that customer will be counted twice in the non-cumulative cohorts.

• Vertically (in green)

• Evaluate if there is a time when you have a higher repurchase rate. For example, here we can see in month 1 and 2 after the first purchase, customers are buying again more than in month 5 and 6.

• Diagonally (in pink)

• This is all your customers who bought in March. Usually, you can see interesting data around the holidays or if you have a special offer, you will see that more customers come back at that time.

• You can toggle between seeing the absolute and percentage values within your cohort tables in the top righthand corner of each table.

• Keep in mind that the cohort analysis only takes total sales from new customers (not total sales for the month).

• If you're on our Plus Plan, you get access to additional filters to analyze your cohorts even further.

# How-To: Read the Cohort Evolution Graph

The Cohort Evolution graph shows you how historical cumulative LTV changes over time for one specific cohort of customers. The "Month 0" is calculated as the customer's first 30 days, starting from the date they first made a purchase.

By hovering over each data point, you can see how the LTV evolves over time for this set cohort of customers (and how many total customers are included in the cohort). Keep in mind that the total number of customers included in the cohort does not change as you view more data points in this graph, but rather, the LTV data will shift in alignment with your customers' lifespan.

# Differences between Shopify & Polar Cohort Analysis

Polar calculates each cohort as the group of customers that purchased for the first time within 30 days of a given month (starting from the date they made their first purchase). We consider this to be the most accurate representation of new customer cohorts, as it counts the "full" months in the evolution of cohorts.

For example, if a customer purchased on November 15th 2022, they are considered in the November 2022 cohort and their "Month 0" is November 15th to December 14th. Their "Month 1" is December 15th to January 14th, their "Month 2" is January 15th to February 14th, and so on for the remaining months set in your date range.

• In the example above, if this customer also makes a second purchase on December 1st 2022 and a third purchase on January 28th, they are counted as returning:

• at Month 0 and Months 2, 3, 4, 5, etc, in the cumulative view

• at Month 0 and Month 2 only, in the non-cumulative view

Alternatively, in the example above, Shopify would consider November as "Month 0", December as "Month 1", and January as "Month 2", regardless of which day in the month the customer first made a purchase.

For this reason, you'll notice slight differences in the percentages of Shopify's cohort tables versus Polar. If you'd like, you can recreate the Shopify cohort tables within Custom Reports on Polar by adding a "Customer First Order Date" filter, and adjusting your date range to look at only the orders made in one month.

In this screenshot, you'd be viewing Shopify's "Month 1" cohort from the example above.