What your cohort LTV curves are actually telling you
Two companies show the same lifetime value on the summary slide: $4,200 per customer. One built it from cohorts that are still climbing eighteen months after acquisition. The other built it from cohorts that reached $4,000 by month four and have barely moved since. Same number, two different businesses, and only one of them is worth more next year than it is today.
The lifetime value most decks report is a single figure, usually a period endpoint or a projection. It tells you where a cohort landed. It says nothing about how it got there, and the path is where the information lives.
The number is an endpoint. The curve is the business.
A cohort LTV curve plots cumulative revenue per customer against months since acquisition, one line per acquisition cohort. Because it is cumulative, every line only rises. The question is never whether it goes up. The question is the slope, and whether the slope holds.
A curve still climbing at month fifteen is a cohort whose surviving customers are spending more than the churned ones took away. A curve that flattened at month five is a cohort that collected its revenue early and then stopped. Both can print the same lifetime value if you wait long enough. Only the first one compounds.
Read cohorts at equal age, not equal date
The most common mistake is comparing a two-year-old cohort to a two-month-old one and concluding the older one is better. Of course it is higher. It has had twenty-two more months to accumulate revenue.
The comparison that means something is at equal cohort age. Where is each cohort's per-customer revenue at month six? At month twelve? Line the vintages up at the same age and the divergence becomes legible. The 2023 cohorts sitting well above the 2024 cohorts at month six is a real signal that acquisition quality slipped, and it is one the P&L will not surface for another year, because the newer cohorts are still small.
A flattening curve is usually a retention story
When a cohort's LTV curve stalls, the cause is almost always underneath it in the retention data. Customers left, and the ones who stayed did not expand enough to cover the gap. The cumulative curve keeps the revenue those customers already paid, so it never falls. It just stops rising.
This is why the LTV curve and the logo retention curve have to be read together. If per-customer LTV flattens at the same age the cohort's retention drops a step, the two are telling one story: the cohort is shedding customers, and the survivors are not expanding fast enough to matter. If LTV keeps climbing while logo retention erodes slowly, you have the healthier pattern, a smaller base of customers each worth progressively more.
The margin question sits on top of all of this
Revenue-based LTV curves answer whether a cohort keeps paying. They do not answer whether that revenue was worth acquiring. A cohort can compound beautifully on a 40% gross margin and still be a poor use of capital if it cost too much to land.
That is a different curve: cumulative revenue per customer weighted by gross margin, divided by what it cost to acquire the cohort, tracked over the same cohort age. Above 1.0, the cohort has returned more gross profit than it cost. The month it first crosses 1.0 is the payback point. A cohort that compounds but takes twenty months to cross is a slower machine than one that flattens early but crosses by month seven.
What to take into the room
Before the meeting, know the shape of your own curves. Which vintages are still sloping up, which have flattened, and at what age. Whether the flattening is churn or a lack of expansion. And whether the cohorts that compound are the ones that were cheap to acquire or the ones that were expensive. An investor will reconstruct all of this from your transaction file in the first hour. Better to have read it first.
Levian plots cumulative revenue per customer for each acquisition cohort as its own line over cohort age, so the vintages that keep compounding and the ones that flatten land on the same axes at equal age. Logo retention by cohort sits in the same section, so when a curve stalls you can see whether it is customers leaving or survivors that stopped expanding.