Cohort Analysis: Comprehensive Guide for SaaS
Cohort analysis is one of the most useful kinds of product analytics reports.
In our guide to cohort analysis, we discuss:
- what is cohort analysis
- how it can help you improve your KPIs
- its main types
- how to conduct cohort analysis and interpret cohort tables
- best tools for analysis
It’s quite a bit so without further ado, let’s get into it!
- Cohort analysis involves tracking and analyzing how different groups of users with common characteristics behave inside the product.
- There are two types of cohorts you should focus on analyzing: acquisition and behavioral cohorts.
- User acquisition cohorts consist of customers who signed up for the product or purchased a subscription within the same time period, for example, a week.
- Analyzing acquisition cohorts can help you find out when users churn, how long they stay engaged, or the impact of product changes on your metrics, like the retention rate.
- Behavioral cohorts are made up of users who interact with the product in a similar way. For instance, it could be all users who completed an event or used a feature.
- Behavioral cohort analysis helps you identify patterns in user behavior that affect your KPIs.
- By combining the two types of analysis, you can make the product more sticky and improve conversion rates and user retention. This translates into high customer lifetime value.
- Start cohort analysis by setting clear goals, like ‘Improve customer retention by 10% by January 2024’ and choosing the metrics to track.
- Based on your goals, decide which cohorts to analyze. For example, acquisition cohorts or users who engaged/didn’t engage with a feature.
- Finally, use an analytics tool to generate the cohort analysis reports. More complex use cases normally require multiple reports.
- How to read a cohort report: The vertical axis in a cohort table shows you the cohorts while the horizontal one – the time. The cells of a cohort table contain information on retention for each period relative to the previous one.
- Unlike Heap and Google Analytics, Userpilot offers teams tools to act on the insights from their cohort analyses.
- If you want to learn more about Userpilot cohort retention analytics, book the demo!
What is cohort analysis?
Cohort analysis is a kind of behavioral analytics that involves tracking the behavior of specific user groups sharing similar characteristics.
Types of customer cohort analysis
There are two main types of cohorts: user acquisition cohorts and behavioral cohorts.
Let’s see how they’re different and what they’re best for.
User acquisition cohorts
User acquisition cohorts are based on the user signup date.
For example, an acquisition cohort may consist of users who sign up for the product in a particular week or month. You can then analyze how many of the users return to use the product again, day-on-day, week-on-week, or month-on-month.
This type of cohort is most suitable for tracking user retention and churn over a period of time. In other words, they can tell when users churn.
You could use acquisition cohort analyses to track the success of new products after the launch or the impact of changes to the product.
Behavioral cohorts are groups made up of users sharing specific behaviors or characteristics.
For example, you could create a behavioral cohort of users who have engaged with a particular feature or completed an event, like a purchase. Or you could group them based on the user persona, the acquisition channels, or the subscription plan they’re on.
What’s the use of behavioral cohorts?
Teams use behavioral cohorts to analyze which user behavior patterns lead to higher churn or retention, or the why behind them.
For example, they can help you identify the features that drive engagement or the kinds of users that benefit from the product the most.
How cohort analysis helps and why you should analyze cohort data
There are a number of ways product and marketing teams can leverage cohort analysis to make better-informed decisions and drive their KPIs.
Improve your retention rates
Cohort analysis and retention analysis are often used interchangeably because many teams use them to boost their retention figures.
First, you can use it to understand when users tend to churn and then identify behavior patterns linked to higher churn and retention.
You then double down on promoting the successful behaviors to users behaving as if they might be at risk of churning.
Identify patterns and trends in your customer behavior
Thanks to churn analysis, you can also identify successful and unsuccessful patterns in user behavior.
For example, you may be able to identify the features that churned users don’t engage with or the features that drive increased conversions. For example, how many users who use one feature have returned and used another feature?
When paired with data from time-based cohorts, these insights can help you engage your users to promote the desired behaviors and improve relevant metrics.
Understand drop-off and average engagement time
By analyzing time-based cohorts, you can learn how long your customers stick around and when they’re most likely to churn.
Thanks to such insights, you can engage users at the right time.
For example, if users tend to churn at the beginning of the journey, just after signing up, you will be looking at improving your onboarding. If it’s later, you need to drive user engagement and help them find repeated value.
Improve product stickiness and customer lifetime value
Once users start realizing the value of the product and adopt it as a go-to solution to their problems, they’re more likely to stick around for longer. And that means more cash flowing into your chest.
How to perform an insightful customer cohort analysis
Now that you know the benefits of cohort analysis, you’re probably wondering how to conduct it to gain actionable insights.
Here’s a step-by-step guide.
Start by setting a clear goal
To make your cohort analysis productive, always start by defining the problem you’re trying to solve.
Whenever you’re deciding on the goals, it’s a good idea to check them against a goal-setting framework like SMART and Lock and Latham’s. In this way, you will make sure that they are relevant, clear, and specific.
An example goal could be to improve customer retention by 10% by January 2024.
Define the cohort metrics you need to track
Another advantage of using a framework like SMART is that it forces you to think about how you’re going to measure progress. However, if you haven’t done it yet, this is the time to choose relevant metrics to track for your cohort analysis.
Let’s imagine that you want to improve customer retention.
In this case, the basic metrics to track will be the retention rate and churn rate. If you want to improve conversions at any particular stage of the funnel, you will be looking at the conversion rates.
You get the drift.
Define user groups and cohorts
Now that we have the goal and the metric locked in, select the user cohorts to analyze.
For instance, if the objective is to identify ways to improve trial-to-paid conversion, you can start with time-based cohorts to find the moment when users tend to drop off and who they are.
Next, you can use behavioral cohorts to spot patterns in user behavior and test hypotheses. For example, you can group your users into those who have completed your onboarding flows or engaged with a particular feature.
Use a tool to build your cohort analysis report
Finally, it’s time to generate your cohort analysis report or cohort chart.
You’re not likely to get away with just one. That’s because your metrics are normally affected by multiple behaviors.
For example, there could be a number of variables impacting your retention, and you have to create user cohort reports for each of them.
Apart from the table-style chart, some tools also enable you to visualize the data as a trend.
How to read a cohort table
Once you create your cohort charts, it’s time to interpret the results.
Let’s look at the key parts of the chart, what information you can find there, and what insights you can get from them.
What does the vertical axis of a cohort table tell you
The vertical axis shows you all the cohorts based on the time when they completed the event.
For example, the chart below shows you the cohorts of users who signed up for the product in May 2023, grouped on a monthly basis.
Reading the table vertically can help you identify patterns over your product’s lifetime.
By comparing the data for different cohorts, you can assess the impact of product changes, like updates, specific marketing efforts, or market changes on your metric.
What does the horizontal axis of a cohort table tell you
The horizontal axis represents the time since the event happened. Like cohorts, this could be in months, weeks, or days.
Reading the data horizontally gives you insights into retention changes for each cohort over time.
That’s how you can find out when different users start dropping off and when they drop off the most, how the retention rate changes over the customer lifetime, or when the retention curve flattens.
For example, we can see that retention dropped by 5% after Month 4, which was followed by another 2.6% drop in the following month and almost an extra 3 % drop in month 6.
The intervals depend on the product. For products that are used every day, like messaging apps, you will be looking at daily data. For a weight tracking app, this could be weekly, whereas for accounting software for freelancers – monthly.
What do the cells in the middle of a cohort table represent
The cells in the middle show you the data for the metric you’re evaluating.
For example, this could be the percentage of users retained or churned, relative to the previous month. By hovering over or clicking on the cell, you can often get additional insights like the names of the retained or churned users in that period (depending on the tool you are using to create these reports).
What makes the table easy to interpret is the heatmap-like shading of the cells. The darker the color, the higher the figure.
Cohort analysis examples
Now that we have covered the basics, let’s look at examples of how to perform cohort analysis.
We’re particularly interested in two different analyses: acquisition cohort analysis and customer cohort analysis.
Acquisition cohort analysis example
Acquisition cohort analysis examines how groups of customers who were acquired during the same time period behave over time.
Let’s imagine you’re a marketing manager for a fitness and weight-tracking app and you want to assess how seasonality affects customer behavior to optimize your marketing efforts.
To do so, you decide to conduct the analysis of user behavior in 4 cohorts, Q1-Q4.
Having visualized the data for the 4 cohorts, you can see that users who signed up in Q1 tend to be way more engaged than users from other cohorts. Their retention, however, drops dramatically after 2 months, when they start churning en masse.
Users acquired in the other quarters tend to be less active, with the Q3 cohort having the fewest daily active users. These cohorts tend to be smaller as well.
However, the retention rates for these cohorts are higher than for Q1. Users who complete user onboarding and don’t churn in the first couple of weeks, tend to stick around for longer.
There are 3 main areas for improvement here:
- retention of Q1 users
- acquisitions in Q2-Q4
- user engagement in Q2-Q4
To improve the retention of Q1 cohort and increase engagement of the Q2-Q4, you decide to introduce elements of gamification to keep users motivated.
At the same time, you divert some of your marketing resources to Q2-Q4 and experiment with different acquisition channels.
As you start introducing the changes, you follow up with more cohort analyses to assess their impact.
Customer retention cohort analysis example
Customer retention cohort analysis is often used to improve user retention by identifying the causes of customer churn.
Here’s our hypothetical scenario:
As a product manager of an instant messaging app, you’re trying to boost your retention figures.
You start by conducting an acquisition cohort analysis just like the one below in order to find out when users stop engaging with the core feature of your app. We are looking at repeat engagement of the same feature here.
The chart reveals that most users churn between days 2 and 3 of engaging with the feature for the first time. There could be multiple reasons for that, so you work with your team to create a couple of hypotheses:
- Users who complete the onboarding checklist, have better retention.
- Users who don’t invite at least 3 teammates within the first 2 days, tend to churn.
To test the hypothesis, you create behavioral cohorts:
- users who complete and don’t complete the onboarding checklist.
- users who invite 3 or more teammates in the first two days and those who don’t.
For each, you create the report and compare the trends. This reveals that completing the checklist doesn’t affect retention dramatically. However, inviting 3 teammates does.
Based on that, you decide to optimize the checklist so that it drives the desired behavior (teammate invites).
After updating it, you follow up with another cohort analysis where you compare the before and after retention figures.
The new checklist has improved your retention but there’s still room for improvement, so you keep iterating on it to optimize it even further.
Data analytics tools for an actionable cohort analysis
You can perform cohort analysis with a spreadsheet.
The question is:
Why would you do that if there are brilliant analytics tools that will save you a lot of time and effort?
Let’s look at a few strong options to consider.
Userpilot: for in-depth behavioral analytics and cohort analysis
Userpilot is a product adoption platform with advanced analytics features, including cohort analysis.
This puts it on par with dedicated analytics tools.
What exactly can you do with Userpilot?
Collect user behavior data
For starters, Userpilot allows you to collect accurate data on user behavior inside web apps without any coding.
Thanks to its Chrome extension, you can easily tag features and create custom events to track with a few clicks.
Build cohort retention tables
Once you have the data, you can use it to build the cohort analysis table.
Userpilot allows you to
- create cohorts on the user and company level.
- choose the start and return events defining retention in your product, like ‘Signed in’ and ‘Invited teammate’.
- filter the events by events, user, and company properties.
- choose the interval by which you view the data (days, weeks, months) and date range (e.g. last 2 weeks, or last 180 days).
Clicking on each cell reveals data on specific users that were retained.
You can also visualize the average retention as a line graph, both linear and cumulative.
Cool isn’t it?
It gets even better!
You can enhance the insights from cohort analysis by pairing it up with other analytics features:
- Heatmaps – to analyze user clicks, hovers, and text infills on a page.
- Funnel analysis – to visualize user conversions and drop-offs at key stages in the journey.
- Trends analysis – to track key metrics over time.
- Paths – to visualize all the events leading up to or following a conversion (coming in Q1, 2024).
Heap: for collecting and analyzing product data
Heap is a product analytics platform with powerful functionality.
The process of creating Retention reports is very similar to that in Userpilot. You set the:
- Start and Return event
- Date range and granularity (by hour, day, week, etc)
- Whether to count every time the user completes the return event (default) or only the first time
Just like Userpilot, Heap allows you to present the data in a line graph. However, the data is more granular as you can see the trends for each cohort, not just the average.
Other analytics features that Heap offers include:
- Autocapture – collects data automatically without tagging
- Session replay – for UI optimization
- Illuminate – to discover friction
- Journeys – visual maps of user interactions
- Funnel analysis
Google Analytics: for basic cohort analysis charts
Google Analytics is the go-to solution for marketers. Traditionally, it used to be only a web analytics solution but its latest iteration, GA4, tracks events inside mobile and web apps.
GA allows you to carry out basic types of analysis, like funnel and cohort analysis.
To conduct cohort analysis, choose
- Dimension/cohort type (like acquisition date)
- Size of cohorts (e.g. day, week, or month for acquisition cohorts)
- The metric to evaluate (e.g. retention)
- The date range and number of cohorts
The visualization consists of 2 parts: the graph and the table.
For more granular insights, GA allows you to break down the data by segments. For example, you can view only mobile traffic and events.
Cohort analysis is a powerful tool that allows you to track user retention over time and analyze how the behavior of different user groups affects your key performance metrics.
When combined with other analytics tools, like funnel analysis, it can offer actionable insights that teams can use to improve product performance and customer experience.
If you want to learn more about cohort retention analysis in Userpilot, book the demo!