Metrics, metrics, metrics. As soon as you launch your product, you’re virtually swimming in data.

The geographies where your customers are based. Their purchasing habits. Their job titles. The number of times they log in and how often they open your emails and their retention rate ...

The data itself will be specific to your business. But the sorting of data is something every startup should learn how to do – and do well.

So today we’re gonna focus on one particular way of sorting data: cohort analysis.

First, let’s break down what a cohort is.

In simplest terms, a cohort is a group of people.

In the case of your product, they may be users or customers. When we talk about cohorts, we’re talking about breaking up that big bucket of “users” into smaller buckets.

You can create a cohort along lots of different axes (depending on the data you have) – age, gender, income bracket, industry, and very importantly, when they became your user or customer.

🚨 What matters when we talk about cohort analysis is that you are keeping track of any given cohort over time.

The “over time” piece is so important here because it allows you to effectively learn what is and isn’t working for your users. In other words, you analyze what a cohort does, and voilá! You have a cohort analysis.

What does cohort analysis get you?

Well, for one thing, it helps you learn.

Learn how to keep customers, and learn how to keep customers buying. Learn what marketing has been effective or what product changes have been ineffective.

Most companies are paying to acquire customers through a variety of channels, but getting them is different than retaining them. In the long run, the most effective way to make money is to keep your existing customers as you add new ones.

That’s where cohort analysis helps you learn. It’s a way to debug your marketing process.

See, cohort analysis allows you to look at a group of users and compare them to a different group of users.

For example:

How does retention compare when looking at users who joined in January vs. users who joined in April? What changes did we make in that time that could have affected retention, and how can we make sure we’re making smart decisions to continue to increase retention over time.

Or …

How does lifetime value (LTV) compare when looking at customers we found via advertising vs. customers we found by hosting a webinar? Which of those audiences are more valuable to us, and therefore where should we be spending our marketing budget?

Or …

How does app usage compare when we look at users with X job title vs. users with Y job title?

The point of cohort analysis is to allow your data to help you work smarter.

So what does a cohort analysis actually look like?

We’re so glad you asked! Elizabeth Yin, one of our co-founders, took the time to do a sample cohort analysis at minute 21:00 of this discussion of How to Drive Revenue by Increasing Customer Retention.

Check it out to watch her in action!

This article was written by Carolyn Abram, a freelance writer with a passion for technology. She also writes fiction and teaches writing classes in her home of Seattle. You can learn more about her at her website or on LinkedIn.