About
The Lifetime value page in Margins groups your customers by the month they first bought from you, then follows each group forward to show what they went on to spend, what they cost to win, and what they were really worth.
Most reports tell you what happened last month. This one tells you what a customer is worth over their whole relationship with you, which is a different question and a more useful one. It's the number that decides how much you can afford to spend on acquisition.
You'll find it in the Lifetime value tab on your Dashboards page.
This article covers what lifetime value is, the questions the page answers, how to read the grid, and every metric on it: what it shows, why it's useful, and exactly how we calculate it.
In this article
What lifetime value is
Why it matters
Questions this page answers
How to read the grid
The metrics, at a glance
Revenue metrics
Profitability metrics
Repeat behaviour metrics
Acquisition economics metrics
Sales quality metrics
How we calculate the numbers
FAQs
1. What lifetime value is
Here's why the whole relationship matters, in numbers:
A customer spends $60 on their first order.
Over the next year, they come back and spend another $110.
Their lifetime value is $170, not $60.
If you only look at that first order, you'll decide you can spend about $20 to win a customer. Look at the whole relationship and you can comfortably spend $70, and outspend every competitor who's still only counting the first order.
To answer that question properly, we group your customers into cohorts: everyone whose first-ever purchase happened in the same month. Your January cohort, your February cohort, and so on. We then follow each cohort forward, month by month.
Grouping by first-purchase month is what makes the comparison fair. If you simply looked at revenue per customer across your whole customer base, a strong month of new customers would drag the average down, not because anything got worse, but because you've just added a lot of people who have only bought once so far. Cohorts remove that distortion.
Key takeaway: A cohort is everyone who first bought from you in the same month, followed forward over time.
2. Why it matters
It tells you what you can afford to pay for a customer. This is the most important number in ecommerce growth. If a customer is worth $85 in profit over their life and costs you $40 to win, you have room to scale. If they're worth $30, you're buying revenue at a loss, and the faster you grow, the faster you run out of cash.
It separates a growth problem from a retention problem. Flat revenue can mean you're winning fewer customers, or that the customers you win are worth less than they used to be. Those two problems have opposite solutions. Cohorts show you which one you have.
It shows you a trend long before your Profit & Loss (P&L) does. You can see the quality of a cohort within a month or two of winning it. Your P&L won't show the effect for a year. If your March cohort is coming back at half the rate your January cohort did, you want to know in April.
It tells you when your marketing spend comes back as cash. Two channels with the same lifetime value are not equally attractive if one pays you back in 2 months and the other in 11. The fast one lets you spend the same cash several times a year.
3. Questions this page answers
How much can we afford to pay for a new customer?
How much can we afford to pay for a new customer?
Look at True LTV and CAC payback together. True LTV, your lifetime value after costs, is your ceiling; payback tells you how long your cash is tied up getting there.
True LTV is $72, so $72 is the absolute most a customer can be worth to you.
CAC payback is 8 months.
If you only have 3 months of cash runway, your real limit is whatever those first 3 months deliver, not the $72.
Are our newest customers as good as our older ones?
Are our newest customers as good as our older ones?
Read down a column, not across a row. Every cohort has an M1, so comparing the M1 column compares every cohort at the same age.
January cohort at M1: $80 of lifetime value.
June cohort at M1: $52.
Same age, lower value. The customers you're winning now are worth less than the ones you won in January. That usually means a channel has been scaled past its best audience.
Important: Comparing a new cohort's Overall figure against an older cohort's is the most common mistake on this page. The newer one hasn't had as long to earn, so it will always look worse.
Comparing at the same age is enough for most businesses. But if seasonality is a big part of your year, a Q4 that's half your revenue or a genuinely seasonal product, then also compare the same points in the year: this January's cohort against last January's, or Q2 against Q2.
A cohort won in the Black Friday rush is often full of gift buyers and deal hunters.
A cohort won in June is more likely to be buying for themselves at full price.
At the same age those two can still look very different, and neither one is wrong.
① The November cohort: the biggest of the year at 26,300 customers, the lowest first order at $79, and the palest row across every age.
Is the problem fewer repeat buyers, or smaller orders?
Is the problem fewer repeat buyers, or smaller orders?
Compare Retention rate with Average order value. These two problems look identical in your revenue and have nothing in common as fixes.
Retention holding at 12%, average order value down from $60 to $48 → a pricing, discounting or product-mix problem.
Average order value holding at $60, retention down from 12% to 7% → a product or lifecycle problem.
Can we spend more this quarter?
Can we spend more this quarter?
Look at True LTV : CAC. Below 1x, your cohorts never earn back what you paid for them. Around 1x you're breaking even and funding growth out of working capital. Comfortably above, you have room, and the thing to check next is CAC payback because that's what decides how quickly the cash comes back.
Are our discounts buying loyalty, or just cheaper orders?
Are our discounts buying loyalty, or just cheaper orders?
Compare your Discount rate on the first order with the Repurchase rate later on.
A cohort won with a 25% welcome offer that repurchases at 34% → the discount bought you a customer.
A cohort won with the same 25% offer that repurchases at 9%, when your usual rate is 30% → the discount bought you one cheap order and nothing else.
4. How to read the grid
Rows are cohorts, columns are ages
① Each row is a cohort, everyone whose first order was in that month. The row label is that month.
② Each column is an age, not a date. M0 is a cohort's own first month, M1 is the month after, and so on. So the M0 column holds January's January, February's February and March's March: three different calendar months, lined up so you can compare cohorts at the same stage of life.
The reference columns
Column | What it shows |
First order at | The month the cohort was won, meaning the month its customers made their first-ever purchase. This is the row label. |
New Customers | How many customers joined that cohort. |
First order | The acquisition order on its own. This isn't the same as M0, because month 0 also includes any repeat orders placed within that first month. |
The Overall column
Overall closes each row with the cohort's whole life so far. How it reads depends on the metric: it either adds the months up, takes where a running total finished, or pools every month together. Hover any Overall cell and the tooltip tells you which.
The Average and Total rows
Average is your typical cohort. It's a weighted average, not a simple one: bigger cohorts count for more, so the figure reflects your business rather than treating a 12-customer month and a 900-customer month as equals.
On the running-total metrics, the Average row builds up month by month. It takes the average gain in each month and adds it to what came before. This matters because your newest cohorts haven't reached the later months yet. Without it, the far right of the row would describe only your oldest cohort.
Note: Because the Average row credits a typical cohort with a full run of months, it can read higher than any single cohort on screen. It describes the path a cohort follows, not a figure any one cohort has banked yet.
Total appears only where adding a column up means something: whole-cohort money and order counts. It's the straight sum of the cohorts in each column, so it naturally gets smaller towards the right, where fewer cohorts have lived long enough to appear.
① The Average row builds up month by month, which is why it ends at $162 when no cohort above it has banked that yet.
The shading
Cells are shaded from pale to dark by where they rank within the grid you're looking at. The scale always uses its full range, so there's always a lightest and a darkest cell.
This means shading is relative to the current view, not an absolute scale. The same $8,296 can be pale in one period and dark in another, so compare shades within a grid, never between two different views.
A month that's 100% for every cohort by definition, like retention's month 0, is left unshaded. It would otherwise use up part of the scale saying nothing.
Why the bottom right is empty
Your newest cohort has only lived one month, so it has no M6 yet. We leave those cells blank rather than showing a zero, because "hasn't happened yet" and "nothing happened" are two different things.
① Cells stay blank where a cohort hasn't lived that month yet. ② Overall closes each row at whatever the running total had reached.
Key takeaway: Read down a column to compare cohorts fairly. Reading across a row tells you about one cohort's life, not about whether your business is improving.
5. The metrics, at a glance
There are 19 metrics, in five groups.
Group | Metrics | What the group tells you |
Revenue |
| What your customers spend, and how the spending is shaped |
Profitability |
| What you keep out of that spending |
Repeat behavior |
| Whether customers come back, and how often |
Acquisition economics |
| Whether a cohort earned back what you paid for it, and when |
Sales quality |
| What returns and discounting are costing you |
Every metric below uses the same worked example so you can follow one cohort all the way through:
A January cohort of 100 customers, who cost $4,000 to acquire.
They placed 100 orders in total worth $6,000 in January.
They spent $2,000 across 25 orders in February, and $1,500 in March.
Their gross profit was $3,000 in January, $1,000 in February and $700 in March.
6. Revenue metrics
Sales (GMV) per customer
Sales (GMV) per customer
What it shows: What a customer spent in each individual month. GMV is gross merchandise value, the total your customers paid before any costs come out.
Why it matters: The raw spending pattern, and how much a typical customer spends in their first month, their second, and so on. It shows the shape of the relationship before any costs.
How we calculate it: The cohort's sales that month, divided by the number of customers in it.
M0: $6,000 ÷ 100 = $60
M1: $2,000 ÷ 100 = $20
M2: $1,500 ÷ 100 = $15
Overall: $60 + $20 + $15 = $95
Note: Every customer in the cohort counts every month, including the ones who never came back. That's deliberate. This is value per customer you won, not per customer who is still active.
Gross LTV
Gross LTV
What it shows: The running total of what a customer has spent.
Why it matters: This is the headline lifetime value figure, before costs. It answers "what has a customer been worth in sales so far", and its shape tells you how long your customers keep contributing.
How we calculate it: The same per-customer figure, added up as the months go on.
M0: $60 → M1: $80 → M2: $95
Overall: $95, where the running total finished, not the sum of the three cells.
Note: Because it adds up, it never falls. A flat stretch means no further spending, not a loss.
Total sales
Total sales
What it shows: The cohort's sales each month, without dividing by customers.
Why it matters: The size of a cohort. A cohort can have excellent per-customer value and still be too small to move your business. This is where you see that.
How we calculate it: The cohort's sales that month. Overall adds the months together.
M0 $6,000 + M1 $2,000 + M2 $1,500 = Overall $9,500
Cumulative orders
Cumulative orders
What it shows: The running total of orders the cohort has placed.
Why it matters: Orders drive your picking, packing and support workload, which per-customer revenue figures hide completely. Two cohorts worth the same in sales can cost very different amounts to serve.
How we calculate it: Orders each month, added up as the months go on.
Average order value
Average order value
What it shows: The average value of the orders placed in each month.
Why it matters: It separates "customers bought more often" from "customers spent more each time". Rising sales with a flat order value is a frequency story; rising order value is a basket story.
How we calculate it: The cohort's sales that month, divided by its orders that month. Overall pools every order across every month, total sales divided by total orders, rather than averaging the monthly figures.
M0: 100 orders at $60 = $6,000
M1: 25 orders at $80 = $2,000
Overall: $8,000 ÷ 125 orders = $64
Averaging the two monthly figures would give $70, letting a 25-order month speak as loudly as a 100-order one.
Note: The First order column shows the acquisition order's own average value. It's usually lower than M0, because M0 also includes any repeat orders placed in that first month.
7. Profitability metrics
Gross profit per customer
Gross profit per customer
What it shows: Profit per customer each month, after product cost and all delivery costs.
Why it matters: Sales flatter. A cohort buying heavily discounted, frequently returned, expensive-to-ship products can look strong on revenue and produce almost no profit.
How we calculate it: The cohort's gross profit that month, divided by the number of customers. See How we calculate gross profit below.
M0: $6,000 of sales, $3,000 of gross profit = $30 per customer
M1: $2,000 of sales, $1,000 of gross profit = $10 per customer
True LTV
True LTV
What it shows: The running total of gross profit per customer.
Why it matters: This is the number that sets your acquisition ceiling. Gross LTV tells you what a customer spends; True LTV tells you what you keep.
How we calculate it: Per-customer gross profit, added up as the months go on.
M0: $30 → M1: $40 → M2: $47
Overall: $47, so $47 is the most you could pay to win a customer and still break even.
Total gross profit
Total gross profit
What it shows: The cohort's gross profit each month, without dividing by customers.
Why it matters: The absolute profit a cohort has produced.
Gross margin
Gross margin
What it shows: Gross profit as a share of that month's sales.
Why it matters: It strips out volume, so you can see whether each sale is actually worth less than it used to be. A falling margin here is a real change in your unit economics.
How we calculate it: Gross profit divided by sales, both taken from the same underlying orders so the two can't drift apart. Overall pools every month together, total profit over total sales, so a big month counts for more than a small one.
M0: $3,000 of profit on $6,000 of sales = 50%
M1: $1,000 of profit on $2,000 of sales = 50%
Overall: $4,000 on $8,000 = 50%
8. Repeat behavior metrics
Retention rate
Retention rate
What it shows: The share of the cohort that placed an order in that particular month.
Why it matters: The clearest read on habit. It tells you whether customers keep coming back regularly, rather than just once.
How we calculate it: The number of customers from the cohort who ordered that month, divided by the size of the cohort.
100 customers in the cohort, 12 of them ordered in M1 = 12%
Note: Month 0 is always 100%, because every customer bought in their first month, so that cell isn't really measuring anything. That's why we leave it out of the shading and out of the Overall figure.
The Overall figure reads "the share who were active in a typical month after their first purchase". If we included month 0, a cohort that's only a month old would score a perfect 100% and your youngest cohorts would always look like your best ones.
Net revenue retention
Net revenue retention
What it shows: Each month's sales measured against what the cohort spent on its first purchase.
Why it matters: It tells you whether a cohort's spending is growing or shrinking compared with how it started. Above 100% in a later month means the cohort is spending more than it did on day one.
How we calculate it: That month's sales divided by the cohort's first-purchase sales.
First purchase: $4,800. M1 sales: $2,000 = 42%
Note: Month 0 is 100% by definition, and we leave it out of the Overall figure for the same reason as retention. We also don't add the months together. That would give you a retention rate of several hundred percent, which is really Gross LTV measured a different way.
Repurchase rate
Repurchase rate
What it shows: The share of the cohort that has ever ordered again.
Why it matters: The cleanest measure of whether your product earns a second purchase. Repeat buyers cost you nothing to acquire, so this is the strongest lever you have on lifetime value.
How we calculate it: We count each customer once, in the month of their first repeat order, then add them up as the months go on, so the figure only ever climbs.
Of 100 customers: 12 came back in M1, 8 more in M2.
M1: 12% → M2: 20%
Overall: 20% have come back at least once.
Note: Read this alongside Retention rate.
They answer different questions. Repurchase rate asks how many ever came back; retention rate asks how many were active in a given month. Retention can never be higher than repurchase, and the gap between them tells you how often your repeat buyers actually return.
Repurchase 20%, retention 18%: your repeat buyers come back nearly every month.
Repurchase 20%, retention 5%: they came back once and disappeared.
9. Acquisition economics metrics
These five metrics measure everything against what you paid to win the cohort. If a cohort has no marketing recorded against it, the cells are left blank rather than showing an enormous return.
Gross LTV : CAC
Gross LTV : CAC
What it shows: Cumulative sales as a multiple of what the cohort cost to win.
Why it matters: A quick read on whether a cohort is paying for itself in sales terms.
How we calculate it: The cohort's cumulative sales divided by its acquisition cost.
$9,500 of sales against $4,000 of marketing = 2.4x
Note: This is sales, not profit, so it always looks healthier than reality. Use True LTV : CAC to make a decision.
True LTV : CAC
True LTV : CAC
What it shows: Cumulative gross profit as a multiple of acquisition cost.
Why it matters: This is the ratio that decides whether you can scale. Below 1x a cohort never earns back its marketing. Comfortably above it, you can afford to spend more.
How we calculate it: Cumulative gross profit divided by acquisition cost.
$4,700 of gross profit against $4,000 of marketing = 1.2x
The same cohort read 2.4x on sales, which is why the sales version misleads.
Return on investment
Return on investment
What it shows: The same relationship as a percentage return, where 0% is break-even.
Why it matters: 0% means you got your marketing back exactly. 20% means you got it back and 20% on top. It's easier to compare against other things you could do with the money.
How we calculate it: Cumulative gross profit less acquisition cost, divided by acquisition cost.
$4,700 of profit on $4,000 of marketing = 18%
Cumulative contribution profit
Cumulative contribution profit
What it shows: The cohort's cumulative gross profit less what it cost to win, in dollars.
Why it matters: The actual cash a cohort has left behind. A cohort can have a healthy ratio and still be too small to fund anything.
How we calculate it: Cumulative gross profit minus acquisition cost.
M0: $3,000 of profit − $4,000 of marketing = −$1,000
M1: $4,000 − $4,000 = $0, the cohort has just paid for itself
M2: $4,700 − $4,000 = +$700
Note: It's negative early in a cohort's life by design. You pay the marketing up front and the profit arrives over time. The month it crosses zero is your payback point.
CAC payback
CAC payback
What it shows: The same figure per customer.
Why it matters: It turns positive the moment a cohort has paid for itself. Reading across the row tells you when that happened, which decides how quickly you can put the same marketing cash to work again.
How we calculate it: Cumulative gross profit per customer minus CAC per customer.
M0: $30 of profit − $40 of CAC = −$10 per customer
M1: $40 − $40 = $0, paid back in one month
M2: +$7 per customer
10. Sales quality metrics
Refund rate
Refund rate
What it shows: Returns as a share of that month's sales.
Why it matters: Returns hit you twice. You lose the sale and you don't get the shipping back. A cohort with heavy returns is worth far less than its sales suggest.
How we calculate it: Returns divided by sales for that month. Overall pools every month together rather than averaging the monthly rates.
M0: $600 returned on $6,000 = 10%
M1: $400 returned on $2,000 = 20%
Overall: $1,000 on $8,000 = 13%, not the 15% that averaging the two would give.
Discount rate
Discount rate
What it shows: Discounts as a share of that month's sales.
Why it matters: Read against Repurchase rate, this tells you whether discounting is buying loyalty or just a cheaper single order. Compare the First order column with the later months to see how heavily you lean on a welcome offer.
How we calculate it: Discounts divided by sales for that month.
First order: 25%. M1: 6%. M2: 5%.
You're buying the first order at a quarter off, which is worth it only if the cohort comes back.
11. How we calculate the numbers
Which customers join a cohort
A customer joins the cohort of the month their first-ever purchase landed.
Only customers we can follow over time can join a cohort. Amazon tells us who the buyer is on Fulfilled by Amazon (FBA) orders but not on Fulfilled by Merchant (FBM) ones, so a customer who has only ever bought through FBM is left out. If we included them, they'd count in every per-customer figure while never being able to appear in a later month, which would drag every one of your numbers down.
Membership is decided once. After a customer is in a cohort, everything they buy afterwards counts: on any store, any product, however it's shipped.
When a sale counts
We count everything by when your customer actually ordered, not when the transaction reached your books. A cohort column is a statement about behavior, so an order belongs to the month it was placed.
How we calculate gross profit
We use the same method as your Channel Performance report:
Net sales − product cost − selling fees − fulfillment fees − payment processing fees
Returns are your actual returns, not an estimate. Because we count everything by order date, a return lands against the order it reverses, which makes your real return data both available and the more accurate figure. That matters more here than anywhere else: using an average return rate would erase exactly the differences between cohorts this page exists to show you.
We measure your delivery rates per store, per month, using a three-month window centred on the month in question, so a cohort's later months are costed at the rates that applied then, not the rates you're paying today.
If you've adjusted your selling, affiliate, FBA or Fulfilled by TikTok (FBT) rates in Channel Performance, we use those adjusted rates here too.
Important: Gross profit is not net profit. It's after product cost and all delivery costs, but before your overheads, salaries and fixed costs. A cohort that looks profitable here isn't necessarily profitable once everything else is paid.
How we handle the most recent months
Returns keep arriving after a month ends. About 60% of a month's returns are in by the time the month closes, and about 94% a month later. So if we costed a just-finished month on what we can see today, we'd be counting all of its sales against only some of its returns, and your newest month would look like your most profitable one, every single month.
Here's what that would look like:
It's 10 March. Your February cohort sold $10,000.
So far we can see $400 of returns against those sales.
If we used that, February's gross profit would look like $5,600.
But your settled months show this store ends up at a 10% return rate, about $1,000 on $10,000 of sales.
So we charge $1,000, and February's gross profit reads $5,000.
That $600 difference is the trap we're avoiding.
So for months that haven't finished settling, we use what your settled months measured for that store, rather than the incomplete picture. Once the month settles, the real figures take over.
How we calculate CAC
Your acquisition cost is made up of three things, charged to the month you won the cohort:
Ad spend, split across your stores using the same rules as your Dashboard.
Affiliate commission, based on that store's sales at its recent commission rate.
Giveaway cost you've actually booked.
Two things are worth understanding here.
All of a month's marketing is charged to that month's new customers. We treat your spend as going entirely towards winning customers, with none of it set aside for keeping existing ones. It's the standard approach, and it makes your CAC slightly cautious, a little higher than reality, if you run retention campaigns.
Marketplace ad spend, and any spend we can't match to one of your stores, is charged to nobody rather than spread across cohorts that didn't earn it. This is why the figure lines up with your Channel Performance report, which leaves out the same spend.
How the Stores and SKUs filters work
These filters work on how a customer was won, not what they bought later. Pick a store and you get the customers whose first order was on that store, and then we count everything they went on to buy, anywhere.
This is the only way to answer "what kind of customer does this channel bring me?". If we filtered their later orders too, your retention would look worse than it really is.
① Stores and ② SKUs filter on how a customer was won, not on what they bought later.
The period you choose
The period sets both which cohorts appear and how long we track each one, and it always ends at the last completed month. A longer period gives your oldest cohorts more months to accumulate, so extending it doesn't just add rows, it makes the existing ones longer.
Key takeaway: Everything on this page is counted by order date and costed at the rates that applied at the time, so a cohort's later months are never re-priced at today's rates.
FAQs
Why is my newest cohort's Overall figure so much lower than my oldest?
Why is my newest cohort's Overall figure so much lower than my oldest?
Because it's had less time to earn. Overall covers a cohort's whole life so far, and a cohort that's one month old has one month in it. Compare cohorts by reading down a single age column instead, where every cohort is the same age.
Why is the Average row higher than any cohort in the grid?
Why is the Average row higher than any cohort in the grid?
The Average row credits a typical cohort with a full run of months, building up the average gain in each month. Your newest cohorts haven't reached the later months yet, so no single cohort on screen has banked that full path. The row describes the path a cohort follows, not a figure anyone has reached.
Why are some cells blank instead of showing 0?
Why are some cells blank instead of showing 0?
The cohort hasn't lived that long yet. A blank cell means "hasn't happened yet"; a zero would mean "nothing happened", and those are different. Acquisition metrics are also left blank when a cohort has no marketing recorded against it.
Why is my retention rate lower than my repurchase rate?
Why is my retention rate lower than my repurchase rate?
That's expected. Repurchase rate counts everyone who has ever come back; retention rate counts only who ordered in that specific month. Retention can never be higher than repurchase, and a wide gap means your repeat buyers come back rarely rather than regularly.
Why don't my Amazon FBM customers appear?
Why don't my Amazon FBM customers appear?
Amazon doesn't tell us who the buyer is on Fulfilled by Merchant orders, so we can't follow those customers over time. A customer who has only ever bought through FBM is left out of cohorts entirely. If they bought through FBA at any point, they're included from that first FBA purchase.
Does filtering by store change which orders count?
Does filtering by store change which orders count?
It changes which customers count, not which of their orders. Picking a store gives you the customers whose first order was on that store, and then counts everything they bought afterwards, on any store.
Why doesn't this match my sales channel's own numbers?
Why doesn't this match my sales channel's own numbers?
Everything here is counted by order date and after fees, returns, product cost and delivery costs. Your sales channel reports gross sales on its own timing and doesn't deduct your product cost. The gap between the two is the fees and costs the channel doesn't show you.
Need help?
If a cohort's figures don't look like what you'd expect, or you're not sure whether a difference between two cohorts is real, reach out to our Support team through chat or at [email protected], and include the cohort month and the metric you're looking at.