3 Steps To Better Holiday Insights With Cohort Analysis

Holiday planning season is in full swing, which brings up some big questions around customer retention. How do holiday customers compare with your core customer base? Are they more likely to be “one-and-done,” or do they make more repeat purchases? Are they fading away faster or slower?

Cohort analysis is one of the most powerful tools available to marketers for assessing long-term trends in customer retention. It’s particularly useful when taking a look at holiday customers and trying to better understand their behavior (and value) over time.

What Is Cohort Analysis?

A cohort is a group of customers who all performed an action at the same time. Whether that means everyone who made a first purchase in the same month or everyone who subscribed to your email list in the same week, a cohort will be a group of similar individuals, with membership criteria defined as needed.

The idea of cohort analysis is to follow this group of customers and watch how their behavior changes over time.

According to Alistair Croll and Benjamin Yoskovitz in their book “Lean Analytics: Use Data to Build a Better Startup Faster,” cohort analysis lets you “see patterns clearly across the lifecycle of a customer (or user), rather than slicing across all customers blindly without accounting for the natural cycle that a customer undergoes.”

Watching different cohorts at comparable points in their lifecycle can help you pinpoint differences in your customer retention over time — rather than looking at changes in a metric like Average Revenue Per User (ARPU) for your whole customer base over time, which can be skewed by new customer acquisition.

For example, you might look at the cohort of shoppers who made their first purchase in Q4 2014 and those who made their first purchase in Q2 2014. By studying each cohort by looking at their behavior in regard to time that’s elapsed since their first purchase, you can make an “apples-to-apples” comparison of their habits.

Here’s how you can get started with cohort analysis:

1. Define Your Cohort Groupings

Depending on your use case, you may want to vary your definitions of “joined” or “at the same time.” For long-term trends, it usually makes sense to group cohorts on a quarterly basis.

If you want to know the performance of customers acquired from a specific campaign or channel, it makes more sense to group by week or month.

For holiday season comparisons, you’ll want to check out customers who made their first purchase in the fourth quarter.

2. Observe Cohort Behavior

Many analytics platforms have cohort reports built in. But if you don’t have a simple database query, grouping customers by first transaction date will do the trick. You will probably want to choose one metric to follow and note changes over time.

For instance, we might look at the Q4 2014 cohort and observe the ARPU of that cohort three quarters later, compared with the Q3 2014 cohort.

3. Compare Different Cohorts Over Time

To understand how a cohort is unique, you’ll need to compare it with another cohort. If you wanted to study the difference between your Q4 2014 cohort and your “normal” cohorts, you might look at the ARPU two quarters after first purchase.

By looking at the “two-quarters-later ARPU” of the Q4 cohort vs. other cohorts, you can get an idea of how well you are keeping those customers engaged and making repeat purchases (and how holiday customers differ from others).

This information can be crucial for marketers in their holiday strategy. Comparing the way that ARPU drops off quarter by quarter among different cohorts can help ascertain how effective your holiday retention strategy has been.

Additionally, you can study holiday cohorts across multiple years to understand which marketing mix has historically worked best.

Final Thoughts

Cohort analysis is a crucial tool for marketers, but it’s important to be aware of its limitations.

The major drawback of cohort analysis is that it’s purely historical. If your company’s competitive landscape, business model, product mix or strategy has changed dramatically over time, past results may not be indicative of future performance.

That being said, it’s still an excellent tool for understanding long-term trends in customer engagement and retention.