Growth only helps if each customer is worth more than they cost to get. LTV:CAC puts that in one number:
LTV:CAC = lifetime value of a customer ÷ cost to acquire a customer
If it's below 1, every new customer loses you money, and growing faster makes things worse.
Step 1: Lifetime value (LTV)
The simple formula for a subscription business:
LTV = ARPU × gross margin ÷ monthly churn
| Term | Meaning |
|---|---|
| ARPU | average revenue per user (customer) per month |
| Gross margin | share of revenue left after the direct cost of serving the customer |
| Monthly churn | share of customers lost each month (see How to Calculate Churn Rate) |
1 ÷ churn is the average customer lifetime in months, so this is just monthly gross profit per customer × months they stay.
Use gross margin, not revenue
Many LTV calculations skip the margin. That's fine if serving a customer costs almost nothing. It's very wrong for:
- AI apps, where each request costs model or GPU time;
- app store subscriptions, where the store takes a cut;
- anything with payment fees, support staff or hosting that grows per user.
Direct costs to include: payment and store fees, compute/API costs, hosting that scales with usage, and customer support.
Step 2: Customer acquisition cost (CAC)
CAC = total sales & marketing spend in period ÷ new customers in period
Fully loaded CAC includes everything you spent to win customers: ads, tools, content, sponsorships, and the salaries (or your own time) spent on sales and marketing. Paid-only CAC (ad spend ÷ customers from ads) is useful for tuning campaigns, but it flatters the business.
Worked example
An AI photo app with a 9 per month plan:
| Input | Value |
|---|---|
| ARPU | 9.00 |
| App store fee | 15% → 1.35 |
| AI compute per user per month | 1.50 |
| Other direct costs | 0.30 |
| Gross profit per user / month | 5.85 (65% margin) |
| Monthly churn | 8% |
| Average lifetime | 1 ÷ 0.08 = 12.5 months |
LTV = 5.85 ÷ 0.08 = 73.13
Acquisition, last month:
| Input | Value |
|---|---|
| Ads | 1,800 |
| UGC creators | 600 |
| Tools | 100 |
| New paying customers | 100 |
| CAC | 25.00 |
LTV:CAC = 73.13 ÷ 25 = 2.9
Close to the commonly quoted 3:1. If we had used revenue instead of gross profit, LTV would have been 9 ÷ 0.08 = 112.50 and the ratio 4.5:1. That looks great, but it's wrong.
(Illustrative numbers, not real HairOver data.)
Step 3: CAC payback period
The ratio doesn't tell you how long you wait to get your money back. Payback does:
CAC payback (months) = CAC ÷ monthly gross profit per customer
25 ÷ 5.85 = 4.3 months
For a bootstrapped company, payback often matters more than LTV:CAC. You have to fund those 4.3 months of acquisition spend from cash you already have. If average lifetime is shorter than payback, you lose money on the average customer, whatever the ratio says.
Reading the ratio
| LTV:CAC | Usually means |
|---|---|
| < 1 | Losing money on every customer. Fix retention, margin or pricing before spending on growth. |
| 1–3 | Marginal. Works only if payback is short. |
| ~3 | The common rule-of-thumb target. |
| > 5 | Very efficient, and possibly under-investing in growth. |
These are rules of thumb, not laws. What matters is that the number is honest and improving.
Why early numbers lie
- Churn from three months of data is a guess. Early cohorts are often your most enthusiastic users, so later ones may churn more.
1 ÷ churngrows fast when churn is small. 2% churn implies a 50-month lifetime you haven't observed. Some teams cap LTV at 24 or 36 months to stay honest.- CAC rises as you scale. The first 100 customers come from your network and best channel, and the next 10,000 cost more.
- Organic customers hide paid CAC. Blending free signups into the denominator makes paid channels look cheaper than they are.
How to improve it
- Raise LTV: reduce churn (onboarding, quality, failed-payment recovery), raise prices, add annual plans, cut compute cost per user.
- Lower CAC: referral loops, content and SEO, better conversion on the paywall, and stopping channels that don't pay back.
The pricing side is covered in HairOver Pricing and Unit Economics.