What Is Cohort Analysis?
Cohort analysis is a form of behavioural analytics that divides users into groups, called cohorts, based on a shared event within a defined time window. The most common cohort is users who first signed up or made a purchase in the same week or month. Once the cohort is defined, you track how members behave over subsequent periods, revealing patterns that aggregate web analytics data cannot show.
The key insight cohort analysis provides is how behaviour changes over time. Aggregate metrics like average session duration or monthly revenue tell you what happened in a period, but they hide whether things are getting better or worse for new users. Cohort analysis strips that ambiguity away. If the January cohort shows 40% of users returning in month two and the March cohort shows only 25%, you know something changed in the product or acquisition channel during that window.
There are two main types of cohort. Acquisition cohorts group users by the date they first encountered your brand or product, making them ideal for measuring how retention trends improve or decline over time. Behavioural cohorts group users by a specific action taken, such as users who watched a product video or users who used a discount code, allowing you to compare the downstream behaviour of different activity groups. Both types are available in tools like Google Analytics 4, Mixpanel, and Amplitude.
For South African subscription businesses, cohort analysis is particularly valuable because it reveals whether churn is improving month over month. A SaaS company in Johannesburg might discover that cohorts acquired through LinkedIn Ads retain at 60% after three months, while cohorts from Google Ads retain at only 35%. That single insight could redirect tens of thousands of rand in media spend.
Cohort Analysis In Practice
The two scenarios below are illustrative examples, not Juicy Designs client results. The figures indicate the scale of effect that cohort analysis work typically surfaces, so treat them as indicative rather than measured.
Picture a Cape Town-based e-commerce retailer selling homewares that wants to understand whether its loyalty programme is working. Using cohort analysis in GA4, the team could segment customers who joined the loyalty programme in January against those who did not. Three months later, the loyalty cohort might plausibly show a repeat purchase rate around 2.4 times higher and an average order value around 20% higher than the non-loyalty group. Evidence of that kind would give management what it needs to invest further in the programme rather than cutting it in a budget review.
Another common application is campaign evaluation. Rather than looking only at the immediate conversion rate of a campaign, imagine a Pretoria digital marketing team tracking what the cohort of users acquired during a specific campaign does in the 90 days after acquisition. This long-view approach would often reveal that lower-cost acquisition channels produce higher-quality customers, even if their initial conversion metrics look similar to premium placements. Cohort analysis shifts the business from optimising for cost per click to optimising for customer lifetime value, which is a far more profitable perspective.
How cohort analysis works
Cohort analysis is an analytical method that groups users into cohorts, sets of people who share a common characteristic within a defined time period, usually when they first engaged, and then tracks how each cohort behaves over time. For example, you might group all users who first visited or made a purchase in a given month into a cohort, then follow how that cohort's behaviour, such as repeat visits, retention or spending, changes in the weeks and months after. By comparing cohorts, those who joined in different periods, or under different conditions, you can see how behaviour evolves over the customer lifecycle and whether changes over time reflect genuine shifts or just differences in who joined when. This time-based grouping is what distinguishes cohort analysis: rather than a static snapshot of all users at once, it follows defined groups through time, revealing patterns like how retention decays after acquisition that aggregate figures conceal.
What cohort analysis reveals
Cohort analysis is especially powerful for understanding retention and lifecycle behaviour, which aggregate metrics hide. It reveals how long customers keep engaging or purchasing after they first arrive, whether retention is improving or worsening for newer cohorts, and how the value of a cohort accumulates over time, insight central to subscription, ecommerce and app businesses. Because it isolates groups by when they started, it can show whether a change, a new product, a marketing shift, an improvement to onboarding, actually improved outcomes, by comparing cohorts from before and after. This makes it a tool for genuine cause-and-effect insight rather than surface trends: a stable overall figure might hide that newer cohorts retain far worse than older ones, a warning aggregate data would miss. The distinction from segmentation is that a segment isolates a group by shared traits at a point in time, while a cohort follows a time-defined group through its lifecycle, so cohort analysis answers questions about behaviour over time, retention, repeat purchase, lifecycle value, that a static segment cannot.
FAQ
What is the difference between cohort analysis and segmentation?
Segmentation splits your audience into groups based on static attributes like age or location. Cohort analysis tracks a specific group over time, measuring how their behaviour changes from a shared starting event such as sign-up date or first purchase. Cohort analysis adds a time dimension that static segmentation lacks.
How do South African e-commerce businesses use cohort analysis?
South African e-commerce stores use cohort analysis to compare retention rates across acquisition channels, promotional periods, and product categories. Customers acquired during a Black Friday sale are tracked monthly to see if they return without discounts, helping businesses judge the true quality of promotional traffic versus organic or email-driven customers.
How do ecommerce businesses use cohort analysis?
To understand retention and repeat-purchase behaviour: grouping customers by when they first bought and tracking how much and how often each cohort buys over the following months. This reveals whether newer customers retain and spend better or worse than older ones, and whether changes to the product, onboarding or marketing genuinely improved customer value over time.