What is Cohort Analysis?
Cohort Analysis is an analytical technique that groups users who share a common characteristic or experience during a defined time period and tracks their behaviour over subsequent periods. It reveals patterns in retention, engagement, and revenue that aggregate metrics obscure.
What is Cohort Analysis?
Cohort Analysis is a method of breaking users into groups (cohorts) based on a shared characteristic — most commonly the time they first signed up, made a purchase, or started using a product — and then tracking how each group behaves over time. Unlike aggregate metrics that blend all users together, Cohort Analysis reveals how specific groups evolve, making it one of the most powerful tools for understanding user behaviour and business health.
Consider a simple example. Your monthly active user count might be growing steadily, which looks positive. But Cohort Analysis might reveal that users who signed up three months ago have almost entirely stopped using your product, and your growth is entirely driven by new users replacing churned ones. Without Cohort Analysis, this critical retention problem would be invisible behind the growing headline number.
Types of Cohorts
Acquisition cohorts (time-based)
The most common type. Users are grouped by when they first interacted with your product — the week or month they signed up, made their first purchase, or installed your app. This reveals how behaviour changes over the user lifecycle.
Behavioural cohorts
Users are grouped by an action they took (or did not take), such as users who completed onboarding, users who used a specific feature, or users who purchased a premium plan. This reveals how specific behaviours correlate with long-term outcomes.
Demographic cohorts
Users are grouped by characteristics like location, industry, company size, or acquisition channel. This reveals how different user segments perform relative to each other.
How to Read a Cohort Analysis Table
A typical cohort analysis table (also called a cohort retention table) has:
- Rows: Each row represents a cohort, typically defined by the month or week of acquisition.
- Columns: Each column represents a subsequent time period (Month 1, Month 2, Month 3, etc.) after the cohort's starting event.
- Cells: Each cell shows the percentage of the cohort that was active (or performed a desired action) during that period.
For example:
| Cohort | Month 0 | Month 1 | Month 2 | Month 3 |
|---|---|---|---|---|
| Jan | 100% | 40% | 25% | 18% |
| Feb | 100% | 45% | 30% | 22% |
| Mar | 100% | 50% | 35% | - |
This table shows that retention is improving: March's cohort retained 50 percent of users after one month compared to January's 40 percent. This improvement might be due to a product change, better onboarding, or a shift in user acquisition quality.
Key Metrics Tracked in Cohort Analysis
Retention rate: The percentage of users who continue to be active over time. This is the most fundamental cohort metric and reveals whether your product is delivering sustained value.
Revenue per cohort: The total or average revenue generated by each cohort over time. This shows whether users increase their spending (expansion) or decrease it (contraction) as they mature.
Lifetime value (LTV): The total revenue a cohort is expected to generate over its entire relationship with your product, typically estimated by extrapolating observed revenue trends.
Payback period: How long it takes for the revenue from a cohort to exceed the cost of acquiring those users. This is critical for managing cash flow and marketing budgets.
Feature adoption: The percentage of each cohort that adopts specific features over time, revealing which features drive engagement and retention.
Cohort Analysis in Southeast Asian Business Context
Cohort Analysis is particularly valuable for companies operating in Southeast Asia:
- Multi-market comparison: By creating country-level cohorts, companies can compare user behaviour across ASEAN markets. Users acquired in Singapore may behave very differently from users acquired in Vietnam or the Philippines, and cohort analysis quantifies these differences.
- Promotional impact assessment: ASEAN markets are characterised by heavy promotional activity (flash sales, voucher campaigns, cashback offers). Cohort analysis reveals whether users acquired during promotions retain and spend at the same rate as organically acquired users — often they do not.
- Seasonal pattern identification: Cohorts created around major shopping events (11.11, 12.12, Ramadan) help quantify how seasonal acquisition affects long-term user quality and revenue.
- Product-market fit validation: For startups and companies launching in new ASEAN markets, improving cohort retention over time is one of the strongest indicators that the product is achieving market fit.
Performing Cohort Analysis
Step 1: Define the cohort. Choose the grouping criteria (acquisition date, behaviour, demographics) and the time granularity (daily, weekly, monthly).
Step 2: Define the metric. Choose what you are measuring for each cohort over time (active users, revenue, feature usage, etc.).
Step 3: Collect and organise data. Query your database to calculate the metric for each cohort at each time period.
Step 4: Visualise the results. Create a cohort table or chart. Heatmaps are particularly effective for spotting patterns.
Step 5: Analyse and act. Look for trends across cohorts (is retention improving?), anomalies (did a specific cohort behave differently?), and correlations (do users from a specific channel retain better?).
Tools for Cohort Analysis
- Amplitude: A product analytics platform with built-in cohort analysis and retention reporting.
- Mixpanel: Another product analytics tool with strong cohort capabilities.
- Google Analytics 4: Provides basic cohort analysis for web and app analytics.
- SQL: Cohort analysis can be performed directly in SQL against any data warehouse, giving maximum flexibility.
- Excel/Google Sheets: Suitable for small datasets and manual analysis.
- Looker/Tableau/Power BI: Business intelligence tools that can visualise cohort data with custom queries.
Cohort Analysis is one of the most important analytical tools for understanding whether your business is genuinely healthy or growing unsustainably. For CEOs, cohort metrics like retention rate and lifetime value provide a far more accurate picture of business health than aggregate metrics like total users or total revenue, which can mask underlying problems.
The most critical insight from Cohort Analysis is retention. If your product retains a higher percentage of users from each successive cohort, it means you are improving your product, your targeting, or both. If retention is declining, no amount of marketing spending will produce sustainable growth — you will always be running faster just to stay in place.
For CTOs, Cohort Analysis provides the data needed to prioritise product development. Features that measurably improve retention and engagement for specific cohorts should receive more investment. Features that do not move cohort metrics may need to be reconsidered or removed.
In Southeast Asia, where customer acquisition costs are rising and competition for users is intensifying, the shift from growth-at-all-costs to sustainable, retention-driven growth makes Cohort Analysis essential. Investors and board members increasingly ask for cohort-level metrics, not just top-line growth numbers, because cohort data reveals the quality and sustainability of that growth.
- Always analyse retention at the cohort level, not in aggregate. Aggregate retention metrics blend old and new users, masking whether the product is actually improving for newly acquired users.
- Compare acquisition cohorts from promotions versus organic acquisition to understand the true quality of promotional users. Heavy discounting often drives short-term growth but poor long-term retention.
- Create market-specific cohorts when operating across multiple ASEAN countries. User behaviour varies significantly across markets, and blended cohorts can hide important regional differences.
- Track revenue per cohort alongside retention. A cohort with moderate retention but increasing per-user revenue may be more valuable than a cohort with high retention but flat spending.
- Use behavioural cohorts to measure the impact of specific product changes. Group users by whether they experienced a new feature and compare their subsequent behaviour to users who did not.
- Automate cohort reporting so it updates regularly. Stale cohort data loses its value quickly, and manual analysis does not scale.
Frequently Asked Questions
What is the difference between Cohort Analysis and segmentation?
Segmentation divides users into groups based on current characteristics (e.g., premium vs. free users, users in Singapore vs. Indonesia) at a point in time. Cohort Analysis groups users based on a shared experience or characteristic and tracks them over time. The key difference is the temporal dimension: segmentation is a snapshot, while Cohort Analysis is a movie. A segment tells you who your users are today. A cohort tells you how a specific group of users has evolved since a defining moment. Both are valuable and complementary.
How many cohorts should I track?
Start with monthly acquisition cohorts for the past 6 to 12 months. This gives you enough data to identify trends without overwhelming your analysis. As you become more sophisticated, add behavioural cohorts (e.g., users who completed onboarding vs. those who did not) and channel-based cohorts (e.g., users acquired through paid ads vs. organic search). The goal is not to track as many cohorts as possible but to track the ones that answer your most important business questions.
More Questions
Good retention varies dramatically by industry and product type. For consumer mobile apps, 25 to 30 percent Day-30 retention is considered good, while many apps see less than 10 percent. For SaaS products, monthly retention above 95 percent (less than 5 percent monthly churn) is considered healthy. For e-commerce, repeat purchase rates of 20 to 40 percent within 90 days are typical for strong performers. The most important metric is not the absolute number but whether your retention is improving over successive cohorts, which indicates your product is getting better for new users.
Need help implementing Cohort Analysis?
Pertama Partners helps businesses across Southeast Asia adopt AI strategically. Let's discuss how cohort analysis fits into your AI roadmap.