# The LTV curve

> Cumulative gross profit per cohort user by month, how to read it, and when a young cohort can be extrapolated.

- Course: [Product Analytics: From Data to Decisions](https://productovi.com/courses/product-analytics.md)
- Module: Are we making money on it?
- Lesson 26/32, ~30 min
- Source: https://productovi.com/courses/product-analytics/ltv-curve

## What is a Halves user worth?

**Alex Tutor:** Theo is planning ad spend for Halves and needs a number from you, the person who answers from the data.

**Theo:** I'm planning the summer budget for paid social. What is a new Halves user worth? I need a number to bid against.

**Alex Tutor:** The honest question is "worth how much by when". We answer it with an **LTV curve**: for each monthly cohort, the cumulative gross profit per user, month by month since sign-up.

![A row of glass jars, each holding a few more coins than the one before, like one jar filling month by month.](https://productovi.com/courses/product-analytics/inline/ltv-curve-illustration-2.webp)

*Value builds up over time, so ask what a user is worth by when.*

**Sofia:** Why gross profit and not revenue?

**Alex Tutor:** The store commission, payments and receipt scanning take 20% before Halves sees a cent, and Theo's bids are paid from what's left. Also divide by **every** user in the cohort, not just payers: Theo pays for installs, and most installs never subscribe.

**Dataset (CSV):** [Halves cohorts by month since sign-up, January 2025 – June 2026](https://productovi.com/datasets/product-analytics/ltv-curve/halves-ltv-cohorts.csv)

Months 0–11 of every monthly cohort; young cohorts have fewer rows. Data to 30 June 2026.

- `cohort`: Sign-up month
- `cohort_users`: Users who signed up that month (staff excluded)
- `month_since_signup`: Age of the cohort: 0 is the sign-up month
- `paying_users`: Cohort users with an active paid plan that month
- `revenue_usd`: Revenue from the cohort that month, USD
- `gross_profit_usd`: Revenue minus store commission, payment and receipt-scanning costs (80% of revenue)

**Alex Tutor:** Start with the January 2026 cohort: add its gross profit for months 0 to 5 and divide by its users.

> **Exercise:** What is the LTV of the **January 2026** cohort at month 5: cumulative gross profit for months 0–5 per cohort user? Round to cents.

**Alex Tutor:** Theo wants a ceiling for his bids. January 2025, the oldest cohort, shows two numbers at month 11. Which one is his ceiling?

> **Exercise:** By month 11, the January 2025 cohort brought in \$2.24 of revenue and \$1.79 of gross profit per user. Which number should cap the bid per install of Theo, the Halves growth marketer?

## A curve, not a number

**LTV curves: cumulative gross profit per cohort user**

| Months since sign-up | January 2025 | July 2025 | January 2026 | April 2026 |
| --- | --- | --- | --- | --- |
| 0 | 0.37 | 0.25 | 0.37 | 0.48 |
| 1 | 0.57 | 0.39 | 0.57 | 0.73 |
| 2 | 0.74 | 0.49 | 0.73 | 0.93 |
| 3 | 0.89 | 0.57 | 0.88 |  |
| 4 | 1.03 | 0.64 | 1.02 |  |
| 5 | 1.16 | 0.7 | 1.15 |  |
| 6 | 1.27 | 0.75 |  |  |
| 7 | 1.38 | 0.8 |  |  |
| 8 | 1.48 | 0.83 |  |  |
| 9 | 1.59 | 0.87 |  |  |
| 10 | 1.69 | 0.92 |  |  |
| 11 | 1.79 | 0.96 |  |  |

*Four cohorts from the table, USD per user by month since sign-up.*

**Alex Tutor:** Here are four cohorts. January 2025 and January 2026 lie almost on top of each other. July 2025 runs lower and flattens: trip groups subscribe for one holiday and cancel.

**Alex Tutor:** Read a curve by its steps as well as its height. How much did July 2025 add in its second half-year?

> **Exercise:** Use the cohort table. How much gross profit per cohort user did the **July 2025** cohort add from month 6 to month 11 (LTV at month 11 minus LTV at month 5)? Round to cents.

**Sofia:** April 2026 starts higher than everyone: \$0.48 in its first month against \$0.37.

**Maya:** Is that the new onboarding? It shipped in April.

**Alex Tutor:** Check people before money. In month 0, 4.34% of the April cohort paid and 4.37% of March's. Gross profit per payer was \$11.02 against \$8.48.

> **Exercise:** Halves shipped a new onboarding in April 2026. In month 0 the April 2026 cohort earned \$0.48 per user against \$0.37 for March, with the same share paying. What best explains the gap?

## Twelve months from six

**Maya:** For the board I need a 12-month LTV for the January 2026 cohort. It's only six months old.

**Alex Tutor:** Don't draw a straight line through six points and keep going. Borrow the **shape** from an older cohort that has lived twelve months under the same conditions.

**Alex Tutor:** January 2025 reached \$1.1564 at month 5 and \$1.7945 at month 11. The ratio, 1.5518, says how much a winter cohort still grows between those ages.

**Sofia:** And February to May 2025 give ratios from 1.543 to 1.556, so it's stable across winter cohorts.

**Alex Tutor:** That's what makes it usable. Multiply January 2026's month-5 value by it.

> **Exercise:** Project the **12-month LTV** (month 11) of the January 2026 cohort: its month-5 LTV times the January 2025 ratio of month-11 to month-5 LTV. Round to cents.

## The wrong template

**Theo:** Why not use July 2025? It's the biggest cohort, so it's the most reliable.

**Alex Tutor:** Size doesn't fix a different kind of user. July's ratio is 1.37, and it would put January 2026 at \$1.59, twenty cents lower. Its users are mostly trip groups who cancel when the trip ends.

![A tailor holds a large coat pattern against a small customer, while a pattern of matching shape hangs nearby.](https://productovi.com/courses/product-analytics/inline/ltv-curve-illustration-3.webp)

*Borrow the shape from a cohort like yours, not just from the biggest one.*

**Sofia:** So the template has to match the cohort, not just be big.

**Alex Tutor:** Yes. And April 2026 can't be projected with any 2025 cohort as it is: prices changed, so its dollars grow differently even if its people behave the same.

> **Exercise:** To extrapolate a young cohort, the biggest older cohort makes the best template, whatever month it signed up in.

**Alex Tutor:** So what would you do with April 2026, the first cohort at the new prices?

> **Exercise:** Since 1 April 2026 new Halves Plus subscribers pay \$4.99 a month instead of \$3.99, and yearly plans went up too. Maya, the product lead, asks for a 12-month LTV of the **April 2026** cohort, the first one that pays the new prices. Which approach is the most defensible?

## When is it safe?

**Alex Tutor:** Four checks before you extrapolate. The template has actually reached the target age. It came through the same season, prices and plans. Several comparable cohorts agree. And you stop at the oldest age you've observed.

**Sofia:** So no "lifetime" number. The table stops at month 11, and winter curves still add about ten cents a month there.

![A walker stands where a paper map ends, while the path keeps going into blank land beyond its edge.](https://productovi.com/courses/product-analytics/inline/ltv-curve-illustration-4.webp)

*Name the number's age: a 12-month value, not a lifetime.*

**Alex Tutor:** Right. Call it a 12-month LTV and give the range the comparable cohorts allow.

> **Exercise:** The January 2026 cohort is six months old. Its 12-month LTV was projected as its month-5 LTV times the January 2025 cohort's ratio of month-11 to month-5 LTV. Which facts make this projection trustworthy? Select all that apply.

## The line for the board

**Maya:** Give me something I can put on a slide.

**Alex Tutor:** Write it with its age, its method and its caveats. Theo will read it too, before he bids.

> **Exercise:** Write Maya, the Halves product lead, 3–4 sentences for the board slide: the 12-month LTV of the January 2026 cohort, how you got it, and what could make it wrong.

**Alex Tutor:** An LTV without an age is a guess dressed as a fact. Say how old the cohort is, what you multiplied it by, and why that template fits.

- Previous lesson: [How to calculate ARPU and ARPPU](https://productovi.com/courses/product-analytics/revenue-metrics.md)
- Next lesson: [CAC, payback and ROI](https://productovi.com/courses/product-analytics/cac-payback-roi.md)

---
Productovi Learning. Free, no sign-up needed to read.