When we think about the growth of a digital business, we often focus on surface-level metrics such as the number of visits or the number of sales. Even though these matter, they only give us a snapshot of the present. They tell us what is happening, but not why it is happening.
To understand how users really behave over time, we need to go further and analyse their journey across time. That is where cohort analysis becomes especially useful, because it allows us to group users who share the same starting point and then follow their evolution to uncover patterns and optimisation opportunities.
Contents
What is cohort analysis?
Cohort analysis is a technique used to study a group of users who share a common characteristic or experience within a specific period of time. Instead of looking at all users as one large undifferentiated group, cohort analysis breaks them into smaller subgroups, or cohorts. This makes it possible to see how the behaviour of each group changes over time.
For example, in an e-commerce business, instead of analysing the sales rate for a full month, cohort analysis makes it possible to track how purchases evolve for customers who signed up during a specific week or month. That way, you can identify whether new customers are coming back to buy again or leaving after their first purchase.
In practice, cohort analysis helps answer questions that global metrics cannot. It can show whether recent marketing campaigns are bringing in high-quality customers who stay and buy again, or whether they are only generating one-off sales. It can also help assess the impact of product or UX changes by comparing the behaviour of a cohort exposed to the new version with that of an earlier cohort.
What is a cohort and what types are there?
A cohort is simply a group of users who share a common characteristic or have performed a specific action during a given period of time. That characteristic or action becomes the starting point for the analysis. Once the cohort has been defined, you can track that group over time to understand how it behaves.
There are different types of cohorts, depending on what you want to analyse.
- Acquisition cohorts: These are the most common. They group users by the moment they were acquired, whether that means their signup date, first visit, or first purchase. For example, “all users who signed up in March.”
- Behavioural cohorts: These group users who performed a specific action. For example, “all users who watched a product demo video” or “all users who added an item to the cart but did not buy it.” This kind of cohort is especially useful for understanding retention and engagement around specific website features.
- Segment cohorts: These are based on audience segmentation. You can build cohorts based on acquisition channel, geography, or even device type, such as mobile versus desktop.
Key metrics for cohort analysis
Cohort analysis is not only about grouping users. It is about tracking the right metrics to understand how each group behaves over time. One of the most important is the retention rate, which shows the percentage of users in a cohort who come back to the website or app in later periods. It is the main metric for measuring loyalty and long-term customer value.
On the other hand, the churn rate is the opposite of retention. It measures the percentage of users in a cohort who stop interacting with the business. A high churn rate in recent cohorts can be a warning sign that something in the onboarding process or in the product itself is not working properly.
Another key metric is the conversion rate per cohort. This makes it possible to see how the conversion rate, for example from visitor to buyer, evolves for a specific cohort. It is useful for evaluating whether users from a particular marketing campaign are more likely to convert than others.
Finally, cohort analysis is one of the best ways to calculate customer lifetime value (LTV) properly. It lets you see the accumulated value that a cohort generates over time, which is crucial when making decisions about customer acquisition investment.
How to do cohort analysis
Running a cohort analysis is a structured process that goes beyond simply looking at numbers in a report. It is a way of breaking down user behaviour and extracting actionable insights. The key is to follow a clear sequence, from the initial definition of the groups to the interpretation of the results.
Step 1: Define the cohort
The first step is to decide which event or characteristic will act as the starting point for the analysis. What exactly do you want to measure? If the goal is to evaluate new customer retention, the cohort could be “all users who made their first purchase in a given month.” If the goal is to analyse the impact of a marketing campaign, you could instead group users by the channel through which they arrived, such as Google Ads or social media.
Step 2: Select the metrics
Once the cohorts are defined, the next step is to choose the key metrics you want to track over time. In e-commerce and digital marketing, the most common ones are retention rate, conversion rate, average order value, and LTV. The choice of metric should be directly linked to the question you are trying to answer. If you want to know whether customers from a new channel are more valuable, then LTV should probably be your main focus.
Step 3: Analyse the data
Once the data has been collected, the real work starts with interpretation. At this stage, it is crucial to look for patterns and trends in the cohort table. Check whether there are sharp drops in retention or conversion at a specific point, or whether one cohort behaves significantly differently from the others. These kinds of anomalies often point to either a problem or an opportunity.
Step 4: Form hypotheses and experiment
Once the trends have been identified, the next step is to form hypotheses. For example, if you notice that retention is lower for one specific monthly cohort, the hypothesis might be that the onboarding process during that month was confusing for new users. From there, you can design and run experiments to improve that part of the experience.
Step 5: Iterate and compare cohorts
Cohort analysis is not a one-off exercise. After implementing changes based on those experiments, you need to create new cohorts and measure the impact. Comparing the retention or LTV of the cohort that experienced the change with that of an earlier cohort will help you validate whether the optimisation was successful and whether customer behaviour has genuinely improved over time.
How to interpret the results
Once the cohort table has been created, the challenge is understanding what the numbers are actually telling you. The key is to look for patterns, anomalies, and trends over time.
The first step is usually to analyse retention. If the retention rate of a cohort drops sharply after the first month, that is a sign that something in the post-purchase or post-signup experience is not working. Users may not be finding enough value in the product, onboarding may be weak, or there may be no communication strategy encouraging them to come back.
It is also important to compare cohorts against each other. If one monthly cohort performs significantly better than the others, it is worth investigating what happened during that period. Was a new and more effective marketing campaign launched? Were changes made to the homepage or checkout flow? The answer may reveal a strategy worth repeating.
Finally, longer-term trends matter too. It is useful to check whether the retention of more recent cohorts is improving over time. If it is, that is a good sign that optimisation efforts and product improvements are working. If it is not, that may point to deeper structural issues that still need to be addressed.
Best practices for cohort analysis
To get the most out of cohort analysis and avoid common mistakes, it helps to follow a few best practices:
- Do not limit yourself to acquisition data: Acquisition cohorts are the most common, but they are not the only useful ones. It is also valuable to create cohorts based on user behaviour, such as users who watched a video or used a new feature, to understand how those actions affect retention and long-term value.
- Choose the right time interval: The cohort frequency, whether daily, weekly, or monthly, should match the nature of the business. If the product is used daily, such as a messaging app, a daily cohort may be more revealing. For e-commerce, weekly or monthly cohorts are often more useful because they better reflect the purchase cycle.
- Combine cohort analysis with other metrics: Decisions should not be based on cohort analysis alone. It is important to review cohort performance alongside other business metrics such as acquisition costs by channel, average order value, and qualitative user feedback. That gives you a fuller picture and helps avoid misleading conclusions.
Recommended tools
You do not need to be a data science expert to run effective cohort analysis. Many digital analytics tools already include this functionality natively.
Google Analytics 4 is one of the most accessible and widely used options. It includes a cohort exploration report that allows you to analyse retention and other key metrics over time. In GA4, you can create cohorts based on acquisition date, triggered events, transactions, or conversions, define return criteria, and choose the metric you want to analyse.
For businesses that need deeper or more specialised analysis, platforms such as Mixpanel and Amplitude are strong options. Both are designed for detailed user behaviour analysis and allow much more granular segmentation based on very specific actions, which makes them especially useful for businesses that need a product-led view of retention and engagement.
If you work with large volumes of data and need full customisation, business intelligence tools such as Tableau and Power BI can also be used. These make it possible to connect multiple data sources and build custom cohort visualisations tailored to the exact needs of the business.
In short, cohort analysis is a fundamental tool for understanding growth in digital marketing and e-commerce. By grouping users around a shared starting point, it makes it possible to move beyond surface-level metrics and understand behaviour over time. That is what makes it so valuable for identifying the real value of customers and the areas where the user experience can still be improved.
