How to use Cohort LTV Curves
A practical guide to plotting cumulative LTV by acquisition cohort — how to build the view, what healthy curves look like, and what to do when newer cohorts decay.
Cohort LTV Curves
A chart of cumulative customer LTV plotted by acquisition cohort against months since first order.
Cohort LTV curves group customers by the month (or week) they first purchased, then plot the cumulative revenue or gross profit each cohort generates as it ages. Each line on the chart is one cohort; the x-axis is months since acquisition; the y-axis is cumulative LTV per customer.
The view answers a question a blended LTV number can't: are the customers you acquired this quarter going to be worth more or less than the ones you acquired last year? Because every cohort is observed at the same age, shifts in pricing, product, discounting, or channel mix show up as cohorts that sit visibly higher or lower than their predecessors.
A single blended LTV number hides the most important signal in your business: whether the customers you're acquiring today behave like the ones from two years ago. Cohort curves expose that drift directly.
For an online store paying €25-€40 in CAC, a 10% decay in 12-month LTV between cohorts isn't a rounding error — it's the difference between a payback period of 6 months and one of 9. The curve view is how you catch it before the P&L does.
Reading the curves
Three things to look for on every cohort chart: the slope of each line, the gap between lines, and the point at which lines plateau. Together they tell you whether retention is healthy, whether newer cohorts are stronger or weaker, and how long it takes a customer to reach their lifetime value.
Healthy curves are concave: steep in the first 60-90 days as repeat orders land, then flattening as the long tail of loyalists carries the cohort forward. A curve that goes flat by month three is a one-and-done cohort — the customers came, bought, and never returned.
The gap between consecutive cohort lines is the real headline. If your January cohort sits below your previous January at the same age, something changed — new channel mix, a price increase that broke value perception, a discount-heavy launch that brought in worse repeat behaviour.
Don't compare immature cohorts to mature ones
A cohort that's only three months old will always look 'worse' than one that's 24 months old — because it hasn't had time to accumulate revenue. Only compare cohorts at the same age (e.g. M6 vs M6, M12 vs M12). Truncate the view so every cohort you display has reached the comparison age.
Building the view in practice
You need three columns from your orders table: customer_id, order_date, and order_value (gross profit is better than revenue if you can get it). Define each customer's cohort as the month of their first order. Then for every (cohort, age_in_months) pair, sum order value and divide by the number of customers in that cohort.
Most apparel and beauty stores use monthly cohorts and plot 18-24 months of history. Stores with shorter purchase cycles (consumables, supplements) get cleaner signal from weekly cohorts; high-AOV categories (furniture, electronics) usually need quarterly cohorts to smooth the noise.
Cumulative LTV per customer by acquisition cohort
2022 cohort
2023 cohort
2024 cohort
In the chart above, the 2024 cohort starts strong — higher M1 spend, possibly from a better launch offer — but its curve flattens earlier than 2023's. By M12 it sits €28 below 2023 at the same age. That's a 15% LTV regression hiding inside a blended number that probably still looks fine.
Benchmarks for healthy cohort progression
There's no universal LTV benchmark — a beauty subscription and a furniture store live in different worlds. But the ratio of M12 LTV to M1 LTV (the 'expansion multiple') is comparable across categories, and gives you a quick read on whether your cohorts are building value or stalling.
The table below shows typical ranges by category for stores in the €1M-€15M revenue band. Use it to sanity-check your own curves: if your M12/M1 multiple is below the lower bound, you have a retention problem dressed up as a CAC problem.
Typical cohort LTV progression by DTC category
| Category | M1 LTV (€) | M6 LTV (€) | M12 LTV (€) | M12/M1 multiple |
|---|---|---|---|---|
| Beauty & skincare | 45-65 | 110-150 | 160-220 | 3.0x-3.8x |
| Apparel & accessories | 70-110 | 130-190 | 180-260 | 2.2x-2.8x |
| Supplements & consumables | 40-60 | 140-200 | 230-340 | 4.5x-6.0x |
| Home & furniture | 180-280 | 240-340 | 290-420 | 1.4x-1.8x |
| Electronics & accessories | 90-140 | 130-200 | 170-260 | 1.6x-2.1x |
Pair the multiple with your LTV to CAC ratio for a fuller picture. A 3.5x M12/M1 multiple on supplements is healthy in isolation, but if CAC has crept from €18 to €32 in the same window, your payback period has lengthened even as cohorts look fine.
Acting on what the curves tell you
When a newer cohort under-performs an older one at the same age, the diagnostic path is: channel mix first, offer second, product third. Cut the chart by acquisition source — if a single channel (often a new paid social campaign or a discount-heavy affiliate) is dragging the cohort down, you've found it in 20 minutes.
If channel mix is stable, look at the entry product. Cohorts acquired through a discounted hero SKU often have weaker LTV than cohorts acquired through full-price bundles, even after controlling for first-order value. That's a merchandising decision, not a marketing one.
Use cohort curves to set LTV-to-CAC targets by channel
Once you have stable cohort curves, you can forecast M12 LTV from M3 behaviour with reasonable accuracy. Combine that with channel-level CAC and you can set a target LTV-to-CAC ratio per channel — and pause channels that consistently bring in cohorts that fail to clear it.
Frequently asked questions
It's a line chart where each line represents customers acquired in the same month (or week), and the y-axis is cumulative LTV per customer plotted against months since first order. It lets you compare how cohorts age against each other at the same point in their lifecycle.
Blended LTV averages across all customers regardless of when they were acquired, which masks improvement or decay in newer cohorts. Cohort LTV isolates each acquisition group so you can see the trend over time and react before the blended number moves.
Gross profit is the better choice because it's what you can actually spend on acquisition. Revenue curves can look healthy while margin curves decay — common when newer cohorts skew toward discounted SKUs or higher-return categories.
Twelve months is the practical minimum to see meaningful curve shape; 18-24 months is better because it shows the long tail where loyal customers contribute disproportionately. Stores under 12 months old can still use the view — they just won't have mature cohorts to anchor against yet.
Use weekly cohorts for high-frequency categories like supplements or consumables, monthly for most apparel and beauty stores, and quarterly for high-AOV low-frequency categories like furniture. The right grain is whatever makes individual cohorts statistically meaningful — at least a few hundred customers each.
Cohort LTV is the numerator of a meaningful LTV-to-CAC ratio. A blended LTV-to-CAC of 3.0x can hide the fact that recent cohorts are tracking to 2.0x — cohort curves let you compute the ratio for each cohort against the CAC you paid to acquire it.
It usually means you're acquiring one-and-done buyers — customers who make a single purchase and don't return. Common causes are discount-heavy acquisition, a hero product that doesn't lead naturally into repeat purchases, or a poor post-purchase experience.
Yes — once you have a few mature cohorts to anchor the shape of the curve, you can extrapolate from M3 or M6 behaviour to estimate M12 LTV with reasonable accuracy. Most DTC stores find M3 cumulative revenue is a strong predictor of M12 LTV within a stable category.
Often it's discount-led acquisition: the higher first-order value reflects a stronger promotion, which both inflates M1 and selects for price-sensitive customers who don't repeat. The curve flattens earlier and ends lower than full-price cohorts.
Monthly is the right cadence for most stores — frequent enough to catch a regressing cohort within one or two months of acquisition, infrequent enough to avoid reacting to weekly noise. Pair it with a channel-level cut whenever the curve shape changes meaningfully.
Track CAC, channels, and funnel conversion in one place
Metricuno connects ad spend, funnel events, and revenue so you can see CAC by channel, cohort, and campaign — without stitching together five tools.