Like our new look? Read on.

Cohort analysis explained

Published Jun 4 2024 8 minute read Last updated Jun 5 2024
cohort analysis
Contributors Dropdown icon
  • Sean Dougherty
    Written by Sean Dougherty

    A copywriter at Funnel, Sean has more than 15 years of experience working in branding and advertising (both agency and client side). He's also a professional voice actor.

What does cohort mean? In most contexts, it’s one or more supporters, companions or people with shared characteristics. For you, your cohorts are your tribe, your besties — you know, the people who really get you. For your e-commerce business, cohorts are groups of people who would all get along just fine if they met. In other words, they're all very similar, maybe buying the same kind of products, hanging out on the same social platforms, or perhaps they prefer to buy via an app rather than a computer.

Because these shoppers don't actually know each other, they don't realize they're in a cohort. In fact, it's up to you as an e-commerce marketer to use cohort analysis to decide who is going to hang out together — virtually speaking, of course. Analyzing and leveraging your cohort data could be the key to understanding your customer base a little more deeply and using that knowledge to expand it.

What is a cohort in marketing?

The term cohort is used in loads of different types of people analysis and usually carries the same meaning: Groups of people with similar or shared characteristics. In sociology, people may be split into cohorts to study how they react in different situations. In epidemiology, medical professionals may decide what cohort of people is most likely to respond to certain types of treatment.

In e-commerce marketing, cohort analysis starts by splitting your audience up into groups of buyers or leads that make sense together. One cohort may include:

  • Customers with a matching acquisition date
  • Customers who bought the same items
  • Customers of a similar age or other demographic
  • Customers who only browse your social media channels
  • Customers who frequently browse but rarely buy
  • High-spending customers
  • Online visitors who don't buy but share your branded material, e.g., social media posts

Cohort analysis is the process of looking at how these groups of people behave over a period of time — in other words, it's a niche of behavioral analytics. Understanding user behavior helps marketers focus campaigns more accurately and deliver content and offers that are relevant to these potential buyers.

Beyond improving marketing campaigns and driving personalization, customer cohort analysis helps marketers understand customer lifetime value (CLV). Going back to your besties analogy, is there one friend who always borrows money and demands lifts but never helps you out? Maybe it’s time to stop investing so much time in that friendship.  Similarly, are you pumping your marketing budget into attracting customers who aren't likely to buy your goods or share your content? If so, it could be time to make a change.

Understanding the behavioral patterns of cohorts helps your company fine-tune its marketing or advertising strategy and ensure every dollar is pushing your business forward.

Types of cohort analysis

For marketers, there are two primary categories of cohort analysis. From here, you can create various cohorts within these two main umbrellas. Your marketing cohort analysis is just the same: data-driven decisions come faster and easier when you group similar types of data together.

Acquisition cohort analysis

You acquire a customer when they first buy from you or interact with you in a significant way. For example, if you have an app that drives user engagement with the hope of leading to sales later down the line, you could count your point of acquisition from the moment someone downloads the app.

Whether you define your acquisition cohort by a customer's first purchase or some other data point,you end up with groups of customers at the same stage on their journey. You can monitor how these customers behave over a specific period, and record how they respond to different marketing efforts throughout varying lifecycle stages.

Monitoring customers based on how long they've been with you and how long they stay helps you understand retention rates and what drives a customer to head to your competitors instead.

Behavioral cohorts analysis

A behavioral cohort is a group of customers or leads with similar behaviors or actions. You could group customers into a single behavioral cohort if they:

  • Only visit your online store or app when offers are on
  • Buy every month on payday
  • Constantly abandon items in their cart but still make the occasional purchase
  • Download your shopping app but don't make purchases
  • Make regular purchases and use your referral program to recommend your services to friends

As you can see, these behaviors differ wildly. Yet by grouping these people into related groups, you can gain a much deeper understanding of how to maintain the desired behaviors and how to change the ones you're not so keen on.

How cohort analysis impacts customer retention

Marketers can use cohort analysis to understand what impacts the customer experience, both negatively and positively. They can use these findings to boost user engagement and drive up user retention. Why is customer experience so important? Because research shows that 80% of customers think the experience they receive from a company is just as important as the goods they sell.

So, as an e-commerce business, you could look at how different cohorts respond to:

  • Your returns processes
  • The online customer journey from browsing to checkout
  • Blogs and articles
  • Social media campaigns
  • Customer service interactions

All these factors drive a customer's impression of your brand. But not all customers respond in the same way, which is why data from cohort analysis is so vital. Your new customers, for example, may react differently to your branded social media posts than loyal customers who have been with you for years. That's why they need to be in a different cohort, so you can understand their needs and analyze their data separately.

It's like heading into the kitchen and introducing your new work friend to some of your besties, then not talking to your new friend for the rest of the evening. That new work buddy will have a worse party experience than your long-term friends, even though the environment is the same.

In a similar way, different types of customers can respond in seemingly opposite ways to the same campaigns, because they need different things. Gaining a deeper understanding of those requirements and gaining actionable insights from cohort analysis helps reduce churn and provide better experiences to existing and new users alike. Plus, with the right customer data, you can monitor your retention rates and stay in control of your business expansion.

Steps to perform cohort analysis

Performing cohort analysis relies on having accurate data from as many sources as possible. Identifying trends is simpler when you've got all the information to hand. Your marketing data stack should include data integration tools, data transformation solutions, and data storage facilities. A simple spreadsheet software solution won't usually cut it, although existing spreadsheets with historical data can certainly feed into cohort analysis. E-commerce businesses also need business intelligence (BI) solutions with data analytics tools that help highlight cohort performance and behaviors.

Your cohort analysis process could utilize the following steps.

Decide on a marketing goal

All marketing strategies must have a goal, whether that's increased sales or improved customer retention. By defining what you want to achieve from cohort analysis, you can create specific cohorts that fall in line with those goals. Of course, your goals can have multiple dimensions: you can analyze cohorts with the goal of increasing sales over a defined time span and boosting your retention rates. Just be sure to document where you are now, where you want to be, and when you expect to achieve that by.

Define your cohorts

The next step once you know what you want to find out is to consider how to split your customers. Decide whether you'll be looking at behavioral or acquisition cohorts and what you'll use as cohort identifiers. These may include when a customer first interacts with you or makes a purchase, or when they remove their contact details from your site/delete their account. However, you may also define cohorts by specific behaviors such as buying low-demand products, shopping in sales, or reacting to social media content.

Your cohort identifiers should relate to your data analysis goals.

Extract the right data

To ensure you're getting all the information on your different groups of customers, you need to connect to the right data sources. These may include:

  • Your customer relationship management system (CRM)
  • Google Analytics or other data analysis tools
  • Your e-commerce platform, e.g., Shopify, WooCommerce, BigCommerce
  • Social media platforms
  • Your blog
  • Other content channels, e.g., YouTube

Again, the channels you'll connect to depend on your goals. However, the more data you have, the better equipped you are to understand your different cohorts based on their behaviors. You may want to assess customer retention data, sales volumes, comments on social media, blog shares, or completely bespoke data points based on your particular e-commerce niche. Ensure you have ways to transform and store this data and feed it through your BI tools for analysis.

Calculate life cycle stages

One of the defining differences between different cohorts is that they will hit different parts of the customer lifecycle at different times. Just like those folks in the kitchen: they've now got through six bags of chips and are looking for the pizza, while the guests in the garden are still on their appetizer. They're at different stages and need different things, just like your cohorts.

You can use charts and tables to show when each cohort takes certain actions, from joining your loyalty program to making a purchase. A particular cohort chart might show purchases since the acquisition date for a period of so many months. Another may show a retention curve: how many customers stop purchasing within a specific period of time.

Once you know when significant actions occur, you can determine why they occur, and make the changes you need in order to retain customers.

Create a cohort analysis report

Great data needs to be presented in an accessible and accurate way. Your cohort analysis report should highlight the common characteristics you've explored and how those impact the goals you set at the start of the journey.

Effective cohort analyses should quickly highlight what types of customers shop with you for the longest and, hopefully, why. When you have access to this data, you can replicate effective marketing moves while removing funding from those that have no impact. In this way, cohort analysis:

  • Saves time and money
  • Helps you allocate resources more effectively
  • Drives customer retention up
  • Boosts sales
  • Increases customer satisfaction scores
  • Enhances brand awareness
  • Helps you provide a more personalized experience

Personalization is so important in modern marketing. Customers want to feel like they're being personally catered to, and with so many tools available to help you achieve this, it's expected. Nine out of ten marketers agree that investing in personalization techniques for their marketing campaigns improves ROI.

It's knowing that some of your party guests are ready for after-dinner drinks, some just got started on the buffet, while others are simply hanging around by the door, waiting to leave. You know you've got to treat them all differently, so make a note of who needs what and find the best ways to keep each group happy.

A cohort analysis example

Let's see a cohort analysis example in action. One aspect marketers are often interested in is the effectiveness of their content marketing strategy. It takes time to create content, so if it's not working as expected, something needs to change.

To gain actionable insights, it's critical to phrase the question right. You might start off with the query, "Is our company blog effective?" But, to make this investigation measurable, you can change this to, "How many people go on to make a purchase within 24 hours of reading our company blog?"

Your cohort size for this should match the number of people reading your blog over specific time periods. Marketers could narrow these user groups down further, for example, splitting up new users and existing customers. However, for our example, we're going to stick with just people reading the blog.

Month

Qty blog readers

Buy within 24 hrs

Buy within 1 week

Buy within 2 weeks

Buy within 3 weeks

Click through, no buy

No click

Jan

1500

150

75

75

30

225

945

Feb

1200

360

120

120

60

300

240

Mar

300

15

15

15

0

75

180

Apr

2700

270

54

54

54

1080

118

May

1400

25

140

70

0

350

490

This table shows how easy it is to get started with an understanding of cohort behavior. These monthly cohorts are all differing sizes but they all have one defining characteristic: they all read the e-commerce company's blog. To gather this information, marketers must integrate metrics from their blogs into their data stack. Encouraging readers to log in when they visit your website is a way to understand exactly who is clicking CTAs and who is purchasing.

When you have access to this type of data, you can quickly see that there are some patterns. By far the most popular time to purchase is almost immediately after reading a blog. While more people read the blog in April, the best conversion rates were in January and May. E-commerce marketers can assess what topic or SEO techniques were used in those months and try to replicate them. This is a great reminder to avoid "vanity metrics": those figures that look impressive but ultimately don't actually mean that much. 2,700 blog readers as a figure looks impressive at first glance. However, because the conversion rate is lower, a high volume of readers doesn't mean that April's blog was the most successful.

Marketers can also see a distinct drop-off in sales after Week 2, suggesting blogs should be published at least fortnightly but probably weekly to provide the best impact.

An added benefit of utilizing cohort analysis in this way is that marketers can quickly spot anomalies and ensure they understand them better. For example, why did blog users drop off so sharply in May? A quick exploration could uncover that the blog administrator had not shared the links to social channels as they usually do. One quick conversation could help boost ongoing sales and brand awareness.

Cohort analysis for e-commerce growth

The core feature of cohort analysis is remembering that, when it comes to marketing, there's no such thing as the average customer. That's why personalization is such an effective marketing strategy. Customers want to be treated as individuals.

Cohort analysis is a fantastic way to reduce churn by helping you understand what different types of customers really want. You can improve aspects of the customer lifecycle by drilling down into quite granular data on customer behavior and tweaking your marketing efforts accordingly. With the right marketing data stack to enable effective cohort analyses, your e-commerce business strategy will be the most informed it's possible to be.

Want to work smarter with your marketing data?
Discover Funnel