Key Takeaways
- Cohort analysis groups customers by acquisition date and tracks behavior over time as a group, revealing patterns invisible to aggregate metrics.
- A brand whose Month 3 cohort retention is improving has a compounding business underneath surface noise, even if top-line growth looks flat.
- The most dangerous pattern is cohort decay: newer cohorts retaining worse than older ones at the same point. Aggregate revenue can mask this for 12+ months.
- Discount-acquired customers almost always retain 30-50% worse than full-price customers. Cohort analysis quantifies the true cost.
- You can build a working cohort dashboard in 14 days using a spreadsheet, no analyst required for brands under Rs.30 crore revenue.
In This Article
- The Two Types of Cohorts You Need
- How to Build a Cohort Retention Curve
- Reading the Curves: Healthy and Unhealthy Patterns
- The Metrics to Compute for Each Cohort
- Comparing Cohorts: What to Look For
- Connecting Cohort Analysis to Operating Decisions
- Common Mistakes in Cohort Analysis
- How to Set Up Cohort Tracking in 14 Days
- Frequently Asked Questions
Standard D2C metrics (revenue, CAC, AOV, gross margin) describe the brand at a point in time. They tell you what happened this month. They cannot tell you whether the customers you acquired this month are behaving the way customers you acquired six months ago behaved at the same point in their relationship with the brand.
That second question is the most important one in D2C. A brand whose Month 3 cohort retention is improving has a healthy business compounding underneath surface noise. A brand whose Month 3 cohort retention is declining has a problem getting worse, even if revenue is growing, because new customers are subsidizing the deterioration.
Cohort analysis groups customers by when they were acquired and tracks their behavior over time as a group. The output reveals patterns invisible to aggregate metrics: retention curves (what percentage of a cohort returns at 30, 60, 90 days, and beyond), cumulative lifetime value curves (how contribution margin accumulates across the cohort over time), and channel-by-cohort comparisons.
01The Two Types of Cohorts You Need
Acquisition Cohorts
Customers grouped by when they first purchased. A “January 2026 cohort” is everyone who made their first purchase in January 2026. You track that group’s behavior over following months.
The core question acquisition cohorts answer: are customers we are acquiring now better, worse, or the same as customers we acquired previously? This is the most-used cohort type and the foundation of the whole framework. Build this first.
Behavioral Cohorts
Customers grouped by an action they took. Examples: customers who made a second purchase within 30 days, customers who purchased a specific SKU, customers who used a discount code on first purchase.
The core question behavioral cohorts answer: do certain customer behaviors predict long-term value? The classic D2C behavioral cohort is comparing customers who made a second purchase within 30 days versus those who did not. The gap in lifetime value between these two groups is usually 3-5x. That single finding changes how you think about your Month 1 retention spend.
02How to Build a Cohort Retention Curve
Step 1: Pull the Data
You need order-level data with three fields at minimum: customer ID, order date, and order value. Pull at least 12-18 months of history. Most Shopify stores can export this directly from the Orders section. WooCommerce users can use a plugin or a simple database query. Your order management system will have an equivalent export.
Step 2: Identify the Acquisition Month for Each Customer
For each customer ID, find the earliest order date in your dataset. The calendar month of that date is their acquisition month. A customer who first purchased on March 15, 2025 belongs to the “March 2025 cohort” regardless of how many orders they place afterward. This is fixed at acquisition.
Step 3: Build the Cohort Grid
Rows are cohorts (each acquisition month). Columns are months after acquisition (Month 1, Month 2, Month 3 through Month 12). Each cell shows the percentage of the cohort that placed an order in that month. Month 1 is always 100% by definition, because all customers in the cohort made their first purchase that month.
Here is an example cohort grid using representative data for an Indian beauty brand:
| Cohort | Month 1 | Month 2 | Month 3 | Month 6 | Month 12 |
|---|---|---|---|---|---|
| Jan 2025 | 100% | 28% | 18% | 12% | 8% |
| Feb 2025 | 100% | 30% | 19% | 13% | 9% |
| Mar 2025 | 100% | 32% | 22% | 15% | n/a |
| Apr 2025 | 100% | 31% | 24% | 16% | n/a |
| May 2025 | 100% | 33% | 25% | n/a | n/a |
Color intensity indicates retention strength. Darker green = higher retention. Cells marked n/a are immature cohorts without sufficient history yet.
Step 4: Compute the Cohort Retention Curve
For each cohort, plot the percentage of customers active in each subsequent month. Overlay multiple cohorts on the same chart. The shape of that curve, and whether newer cohorts sit above or below older ones at the same month, tells you almost everything about your brand’s health and trajectory.
03Reading the Curves: What Healthy and Unhealthy Look Like
Retention drops steeply in the first 1-3 months, then flattens at a sustainable level. The flat tail at 8-12% over 6+ months represents the loyal core of the cohort. This is the target shape for most D2C categories. The steep early drop is expected and normal. The flattening is what signals real product-market fit.
Newer cohorts performing better than older cohorts at the same month-since-acquisition mark. This means product quality, marketing targeting, or customer experience is improving over time. Rare, valuable, and almost always indicates a brand that is compounding. Compare Month 3 retention of your most recent 3 months of cohorts to cohorts from 12 months ago.
The curve drops steadily across all months and never finds a stable level. No flat tail forms. This means the brand is not building a loyal core. Every month, more of the original cohort is gone. The business model is essentially “rent new customers continuously,” and CAC must grow to sustain revenue.
Strong Month 1-2 retention, then a sudden sharp drop in Month 3 or 4. Indicates a product or experience issue that surfaces after the first repeat purchase: quality problems becoming apparent, packaging damage on second order, customer service failures, or a product that doesn’t hold up to repeat use. The cliff almost always has a specific operational cause that can be found and fixed.
Recent cohorts performing worse than older cohorts at the same month-since-acquisition point. Compare Month 3 retention of customers acquired 12 months ago versus customers acquired 3 months ago. If newer cohorts retain less, the brand is getting weaker even as it may be growing revenue. This is the most dangerous pattern precisely because aggregate metrics don’t show it. A brand can grow 60% YoY while experiencing cohort decay for 18 months before the math catches up.
“A brand whose Month 3 cohort retention is declining has a problem getting worse, even if revenue is growing. New customers are subsidizing the deterioration. Cohort analysis is the only metric that reveals this before it becomes a crisis.”
Ankit Sarawagi, CFOmatrixCategory Benchmarks: Healthy Month 3 Retention
| Category | Healthy Month 3 Retention |
|---|---|
| Beauty / personal care | 18-28% |
| Subscription food / supplements | 35-55% |
| Non-subscription food | 12-22% |
| Fashion | 8-18% |
| Home / lifestyle | 6-15% |
| Pet care | 25-40% |
Compare your brand to its category benchmark, not to a universal number. A fashion brand at 14% Month 3 retention is healthy. A supplement brand at 14% has a serious retention problem. Context is everything.
04The Metrics to Compute for Each Cohort
Retention percentage is the starting point, not the endpoint. Once the grid is built, layer in these metrics per cohort:
- Cumulative orders per customer: how many total orders has the average customer placed by Month 3, 6, and 12? This is the order-frequency input for an honest lifetime value calculation, grounded in actual cohort behavior rather than assumptions.
- Cumulative revenue per customer: total revenue generated per customer from acquisition to the current date. This is the gross revenue version of cohort LTV.
- Cumulative contribution margin per customer: contribution margin LTV by cohort. This is the number that actually matters for CAC payback and unit economics analysis.
- Cohort payback period: when did the average customer’s cumulative contribution margin exceed the CAC for that cohort? This is the most reliable version of CAC payback, because it uses actual cohort behavior rather than blended averages.
- Channel-level cohort retention: the same cohort grid filtered by acquisition channel. Often reveals that one channel’s customers retain significantly better or worse than others, even when the CAC looks similar.
- Discount-level cohort retention: discount-acquired customers almost always retain worse. Cohort analysis does not just confirm that they do. It quantifies by how much, which enables an accurate calculation of the true cost of a discount-led acquisition strategy.
05Comparing Cohorts: What to Look For
Are recent cohorts performing better, worse, or the same as older cohorts at equivalent month-since-acquisition points? This is the single most important diagnostic for whether your brand is compounding or degrading. Run this comparison every month. If Month 3 retention has improved from 18% to 25% over 12 cohorts, your brand is fundamentally getting stronger.
Customers acquired via Meta vs. Google vs. influencer vs. organic. Channels with worse cohort retention usually deserve less budget, even if the headline CAC is similar. A channel with 20% higher CAC but 40% better Month 6 retention may be materially more valuable once cumulative contribution margin is the comparison basis.
A common finding across Indian D2C brands: customers acquired with 25-40% off their first purchase retain 30-50% worse than full-price customers at Month 3 and beyond. This makes growth on heavy discounts mathematically suspect once cohorts are properly analyzed. The true cost of a discount is the discount amount plus the lost contribution margin from lower retention across the cohort’s lifetime.
Did the SKU of the customer’s first order predict their long-term behavior? Many D2C brands find that customers whose first purchase was a specific “hero” SKU show 1.5-2x the lifetime value of customers acquired via other SKUs. This finding directly shapes which products you prioritize in acquisition marketing and which products you invest in most heavily for inventory and conversion optimization.
06Connecting Cohort Analysis to Operating Decisions
If Channel A customers retain 30% worse than Channel B customers at the same CAC, Channel B is strictly more valuable on a cumulative contribution margin basis. The decision is to shift budget toward Channel B and reduce reliance on Channel A. This decision is invisible without cohort data by channel. With it, the reallocation is mathematically obvious and defensible to any investor or board.
If customers acquired with discounts retain materially worse, the implicit cost of the discount strategy is higher than the discount amount. The full cost = the face value of the discount + the lost contribution margin from lower retention across the cohort’s lifetime. Cohort analysis quantifies this total cost, and the number usually changes discount policy. Brands that run this analysis frequently discover that their aggressive acquisition discount was generating customers who never reached payback.
If customers whose first purchase was a specific SKU show consistently higher cohort LTV, that SKU is your acquisition flagship. Invest in its marketing, inventory depth, and conversion optimization disproportionately. It earns that investment because customers who start there compound more. This is a finding cohort analysis surfaces that no aggregate dashboard ever will.
The cohort retention curve shows precisely where customers drop off. If the steepest part of the curve is Month 1 to Month 3, that window is where retention investment has the highest expected return: post-purchase email flows, second-purchase incentives, lifecycle marketing tailored to days 30-90. Investing in retention without knowing where drop-off occurs is guesswork. With the cohort curve, it is targeted and measurable.
07Common Mistakes in Cohort Analysis
08How to Set Up Cohort Tracking in 14 Days
This is a practical sequence for any D2C brand starting from zero. No BI tool required. One person with Excel or Google Sheets and 4-6 hours per week can complete this.
- 1Days 1-3: Get the data export. Pull 18 months of order data with customer ID, order date, order value, acquisition channel, and SKU. Shopify and WooCommerce both support this with a single export. Your order management system will have an equivalent. Clean the data: remove test orders, remove returns if your system flags them, confirm customer IDs are consistent across orders.
- 2Days 4-7: Build the cohort grid in Excel or Google Sheets. Acquisition month as rows, months-since-acquisition as columns, percentage of cohort active as cell values. Use COUNTIFS or pivot tables. Month 1 is always 100%. Verify by spot-checking three cohorts manually before you trust the formulas.
- 3Days 8-10: Layer in metrics beyond retention percentage. Add cumulative orders per customer, cumulative revenue per customer, and cumulative contribution margin per customer, computed for each cohort at Month 3, 6, and 12.
- 4Days 11-12: Build the channel-segmented views. Duplicate the cohort grid and filter by acquisition channel. Compare your top 3 channels head to head on Month 3 retention and cumulative contribution margin per customer.
- 5Days 13-14: Build the discount vs. full-price view. Duplicate the grid filtered by whether the customer’s first order used a discount code. Compare retention and cumulative contribution margin. This single comparison is often the most commercially important finding in the entire analysis.
After 14 days, you have a working cohort analysis dashboard. The discipline is in refreshing it monthly and using the output to drive actual decisions on channel allocation, discount policy, and retention investment.
Build Cohort Analysis Into Your Monthly Finance Review
CFOmatrix works with D2C founders and finance teams to build cohort dashboards, interpret the data, and connect it to operating decisions. If your brand is past Rs.5 crore in annual revenue and doesn’t have a working cohort view, this is the place to start.
Talk to CFOmatrix