- Calculate customer lifetime value on contribution margin over a fixed horizon (90, 180 or 365 days), measured from real cohorts, not from an average order value multiplied by a guessed lifespan.
- Most LTV numbers are inflated by the same few traps: revenue instead of margin, blended averages, open-ended lifespans, ignored refunds and discounts, and judging young cohorts too early.
- Use LTV to set a maximum CAC for the payback window your cash can afford, then check it by channel and first product before you scale spend.
What is customer lifetime value, and which version should you use?
Customer lifetime value (LTV or CLV) is the value a customer brings your business over their relationship with you. The definition sounds simple. The problem is that “value” can mean three different things, and each gives a very different number.
| Version | What it counts | Typical use | Risk |
|---|---|---|---|
| Revenue LTV | Everything the customer paid you | Reported by most store and email tools | Overstates value; can’t be compared to ad spend |
| Gross margin LTV | Revenue minus product cost | Finance and merchandising | Ignores shipping, fees, discounts and returns |
| Contribution margin LTV | Revenue minus all variable costs, before ad spend | Setting CAC targets and payback | Needs cost data per order |
Only the last one answers the question a growth team actually asks: how much can we afford to pay for a new customer? Gross margin removes product cost but leaves out shipping, payment fees, packaging, discounts and returns, which on many DTC orders are larger than people expect. Contribution margin removes them all, so it can sit directly against customer acquisition cost.
There’s also a time question. Historic LTV is what customers have generated so far; predicted LTV is a model’s estimate. Each has its own job.
What is the customer lifetime value formula?
You will see this formula everywhere:
LTV = average order value × purchase frequency × margin × customer lifespan
It’s fine for a back-of-envelope check. Its weak spot is “customer lifespan.” For a non-subscription store, nobody cancels, so lifespan is usually estimated as 1 ÷ churn rate. If annual churn is guessed at 20%, the formula assumes the average customer stays five years. Five years of repeat orders from a brand that has existed for three is not a number you should base ad spend on.
The more reliable approach is cohort-based and has a fixed horizon:
LTV (horizon) = first-order contribution margin + (repeat orders per customer within the horizon × contribution margin per repeat order)
Or, if you can pull it directly: LTV (horizon) = total contribution margin from a cohort within the horizon ÷ number of customers in the cohort.
A cohort is a group of customers who placed their first order in the same month. Cohort analysis lets you watch how much margin each group has produced at 30, 90, 180 and 365 days. You stop guessing lifespan and start reading it.
Worked example: one cohort, two answers
The numbers below are illustrative. A store acquires 1,000 new customers in January.
- First orders average $70 in revenue and leave $25 in contribution margin after a 15% welcome discount, shipping, fees and a returns allowance.
- Repeat orders average $75 and leave $33 each, because they rarely carry a discount.
| Days since first order | Cumulative repeat orders | Revenue LTV per customer | Contribution margin LTV per customer |
|---|---|---|---|
| 0 | 0 | $70.00 | $25.00 |
| 90 | 200 | $85.00 | $31.60 |
| 180 | 340 | $95.50 | $36.22 |
| 365 | 520 | $109.00 | $42.16 |
Now say CAC is $35. On revenue, 12-month LTV:CAC is about 3.1, the textbook “healthy” ratio. On contribution margin, it’s about 1.2. Same customers, same year, and one view says scale while the other says be careful. The cohort crosses its $35 CAC somewhere between day 90 and day 180, so payback takes roughly five months.
What are the traps that inflate LTV?
Most bad LTV numbers come from reasonable-looking shortcuts, not calculation errors.
| Trap | What it does to LTV | Fix |
|---|---|---|
| Revenue instead of margin | Inflates it by the full variable cost of every order | Calculate per-order contribution margin first |
| Open-ended lifespan | Assumes years of orders you haven’t observed | Use a fixed 90, 180 or 365-day horizon |
| One blended average | Hides weak channels behind strong ones | Split cohorts by channel, first product and first-order discount |
| Refunds and returns ignored | Counts revenue you gave back | Use net revenue and a returns allowance per category |
| Discount cost left out | Treats a 30%-off first order like a full-price one | Record the discount on each order and subtract it |
| Judging young cohorts | Last month’s customers look worthless | Only compare cohorts that have reached the horizon |
| “Average customer value” across all customers | Overweights your oldest, most loyal buyers | Measure from first-order cohorts, not the whole customer base |
| Duplicate profiles and guest checkouts | Splits one repeat buyer into several one-time buyers | Merge on email before counting repeat orders |
Two of these deserve more attention.
Blended LTV hides your best and worst customers
A store-wide LTV mixes customers who came from branded search, cold Meta prospecting, influencer codes and marketplace promotions. Their behavior often differs a lot. If customers who first bought a starter kit come back far more than those who first bought on a sitewide sale, a blended number tells you neither. Split your cohorts by acquisition channel, first product and whether the first order used a discount. That’s where LTV starts changing decisions, such as which products to feature in prospecting ads, and how deep a welcome offer should be.
Predicted LTV is a ranking tool, not a budget
Klaviyo shows a predicted CLV on each profile once an account meets its data requirements, which include at least 500 customers with orders, 180 days of order history and recent orders. Its default prediction period is one year, and it’s built on order value. That makes it useful for ranking customers, for example to choose who gets an early-access text or a VIP offer. Before you compare it to CAC, convert it to margin. Our Value × Intent (RFMB) model uses the same idea: classic RFM scores rank value, and a Buying Stage layer (B1 visit, B2 product view, B3 add to cart, B4 checkout, B5 purchase, last 30 days) adds intent. Neither replaces a margin-based cohort LTV for budgeting.
How do you use LTV to set CAC targets?
LTV on its own is a score. It becomes useful when you turn it into a spending limit.
Start with your CAC payback period: how many months until a new customer’s cumulative contribution margin covers what you paid to acquire them. Then pick a payback window your cash position can handle.
Maximum CAC = cumulative contribution margin per customer at your payback horizon
Using the illustrative cohort above:
| Payback window you can afford | Maximum CAC |
|---|---|
| First order only | $25.00 |
| 90 days | $31.60 |
| 180 days | $36.22 |
| 365 days | $42.16 |
A brand with thin cash reserves might cap CAC at the 90-day value. A well-funded brand with stable cohorts might accept 365 days. The LTV:CAC ratio is the same idea expressed as a multiple: divide LTV at your chosen horizon by CAC. A common rule of thumb is 3:1, but as the example shows, that target only means something if both sides are measured on margin.
Two practical rules:
- Set targets per channel. If cohorts from one channel repurchase more, that channel can afford a higher CAC. One store-wide limit underfunds your best sources and overfunds your worst.
- Use new-customer CAC. Divide acquisition spend by first-time buyers only. Platform “cost per purchase” includes returning customers and will make CAC look lower than it is.
We cover the per-order math in contribution margin for eCommerce. This article’s focus is the customer-level view built on top of it.
What raises LTV without giving margin away?
The fastest way to raise LTV on paper is a bigger discount on the second order. It often raises revenue LTV and lowers contribution margin LTV. Levers that tend to raise both:
- The second order. In most stores the largest drop-off is between order one and order two. We explain why in the first-to-second purchase rate.
- A lighter first-order discount. A smaller welcome offer raises first-order margin, and the cohort split above tells you whether it changes repeat behavior.
- Product sequencing. Lead acquisition with the products whose buyers come back most.
- Repeat-order AOV. Bundles and replenishment sizes on repeat orders lift margin per order without discounting.
- Retention flows matched to value. Reserve incentives for customers who need them, and keep loyal buyers on full price.
How do you calculate your LTV this week?
- [ ] Export orders with customer ID or email, order date, net revenue, discounts and refunds from Shopify, WooCommerce or Magento.
- [ ] Merge duplicate customer records on email address.
- [ ] Calculate contribution margin per order: net revenue minus product cost, shipping, fulfillment, payment fees and a returns allowance.
- [ ] Group customers by first-order month and calculate cumulative margin per customer at 90, 180 and 365 days.
- [ ] Split the cohorts by acquisition channel, first product and first-order discount.
- [ ] Compare each cohort’s margin LTV to new-customer CAC for that channel.
- [ ] Choose a payback window, set a maximum CAC per channel and review it each quarter.
FAQ
What is a good customer lifetime value?
There isn’t a universal number, because LTV only means something next to what you pay to acquire a customer. A useful test is whether margin-based LTV at your chosen horizon covers CAC with room to spare, and whether newer cohorts are holding steady or improving compared with older ones.
How often should I recalculate LTV?
Quarterly works for most stores, with a monthly check on new cohorts. Recalculate sooner after a big change in pricing, discount policy, shipping costs or acquisition channel mix, because each of those changes the margin your customers produce.
Should LTV include ad spend on repeat orders?
For setting CAC targets, calculate LTV before ad spend and compare it to CAC. If you pay to bring existing customers back, for example through retargeting, subtract that spend from repeat-order margin so the LTV reflects what those orders really contribute.