Cohort Analysis and LTV for Indian D2C

Most Indian D2C dashboards lead with blended numbers. Total revenue, overall repeat rate, average order value across everyone. These averages feel reassuring and tell you almost nothing about whether your business is getting healthier. A blended repeat rate can fall in a great month simply because you acquired a lot of new customers who have not had time to come back. The average is not lying, it is just answering the wrong question.

Cohort analysis answers the right one. It asks how a specific group of customers behaves over time, and it lets you compare like with like.

What a cohort actually is

A cohort is a group of customers who share a starting point. The most useful one is the month of first purchase. Everyone who bought for the first time in a given month becomes one cohort, and you follow that group forward. In month one, what share ordered again? In month two, month three, and so on?

Because each cohort starts together, you can line them up and compare. The January cohort at month three sits next to the February cohort at month three. If later cohorts retain better than earlier ones, your product, onboarding, or customer experience is improving. If they retain worse, something is quietly breaking, and the blended number would never have told you.

Read the retention curve honestly

Plot each cohort as a line showing the share still purchasing in each following month. Two things matter.

• The shape. Most curves fall fast, then flatten. That flat tail is your loyal base, the customers who keep coming back. A curve that never flattens means you are renting customers, not keeping them.
• The drift between cohorts. Are newer cohorts landing above or below older ones at the same age? That drift is your real trend, cleaned of the noise that acquisition volume adds to blended metrics.

In India, purchase cycles vary enormously by category. A snack brand may see monthly reorders, a supplement brand a monthly refill, and an apparel brand a gap of several months. Read your curve against your category’s natural rhythm, not a generic benchmark.

Build LTV on margin, not revenue

Lifetime value is where many brands fool themselves. The tempting version multiplies average order value by expected number of orders and calls it LTV. That is revenue LTV, and for a discount-driven business it is close to fiction.

Use contribution margin LTV instead. Start with what the customer paid, then subtract the cost of goods, the shipping you covered, the payment fees, the returns, and the discounts you handed out to win the sale. What remains is the money that customer actually leaves in the business. Sum that across the expected life of the relationship and you have an LTV worth trusting.

This matters most in India because free shipping and aggressive first-order discounts are common. A customer who looks valuable on revenue LTV can be unprofitable on margin LTV once you count the incentives that acquired and retained them. You cannot see that until you do the subtraction.

Compare LTV to CAC by channel

LTV means little on its own. Its job is to be compared to what you paid to acquire the customer, the customer acquisition cost. The ratio of the two tells you whether a channel is building the business or draining it.

Do this by channel, not blended. A performance channel might deliver customers cheaply who never return, while a slower channel delivers fewer customers who reorder for years. On a blended view they average into a comfortable, misleading number. Split by acquisition channel and the picture sharpens. You will often find one or two channels quietly carrying the profitability of the whole business, and one or two you should stop feeding.

Compare margin LTV to CAC within each channel over a defined window. If it takes eight months of a customer’s life to earn back their acquisition cost, you need the cash and the retention to survive those eight months. That is a planning constraint, not a footnote.

Turn the numbers into decisions

Analysis is only useful if it changes what you do. A few decisions cohorts and LTV should drive.

• Where to spend. Shift budget toward channels with strong margin LTV to CAC, even if their upfront CAC looks higher.
• What to fix. If a recent cohort’s month-one retention drops, investigate the first purchase experience for that period. Something changed.
• Which discounts to keep. If a promotion lifts first orders but the resulting cohort never reorders, you bought revenue, not customers.

None of this requires a large data team. A clean orders table with first-purchase dates, a margin calculation per order, and a simple cohort view will take you most of the way. The discipline is not technical. It is the willingness to stop looking at flattering averages and start looking at how each group of customers really behaves over time.