Every example and amount on this page is illustrative. It shows the method and is not a benchmark, a quote or advice.
What Cohort Profitability Actually Measures
Cohort profitability measures how much profit a defined group of customers generates over a common period of time. A cohort is formed around a shared starting event, usually the date of each customer’s first order. For example, the January cohort contains customers whose first order occurred in January; the February cohort contains customers whose first order occurred in February.
The measurement then follows each group for the same observation window: 30, 90, or 365 days from its starting event. Within that window, it includes the cohort’s eligible orders, repeat purchases, product costs, returns, refunds, and other costs included in the selected profit definition. The result shows how profit develops as the cohort moves from its first purchase to later purchases, rather than showing only what customers spent on day one.
The fixed window is essential. Comparing a January cohort after 90 days with a February cohort after 20 days makes the newer group appear less profitable simply because it has had less time to purchase. Cohorts should be compared at the same age, and the currency should remain consistent within the comparison.
This is narrower than general customer profitability analysis. Customer profitability asks what an individual customer or the whole customer base is worth. Cohort profitability asks how a group with the same starting point performs as a unit. It is also not a retention report. Retention can show whether customers order again; cohort profitability shows whether those orders produce enough profit after the relevant costs to improve the group’s economics.
The profit layer must be named clearly. A cohort can be assessed on gross profit, profit after advertising, or net profit, but those are different measurements. A report that labels revenue as profitability, or silently changes which costs are included, cannot support a reliable comparison. ORVX keeps the cohort’s realized revenue tied to delivered orders and can show the resulting gross, net, and advertising-related economics at the chosen level.
The Formula in Plain English
Start with three linked calculations. Realized revenue = the value of delivered orders − discounts − refunds. Include customer-paid shipping only if it is treated consistently as revenue; exclude taxes, cancelled orders and orders later returned. The time window must be explicit: for example, revenue generated by customers during their first 90 days after the cohort’s first purchase. Use customer age, not just calendar months, so a customer acquired on January 28 and one acquired on February 2 are measured across comparable periods.
Variable costs are the costs that change with those orders or customers: product cost (COGS), fulfilment and shipping paid by the merchant, payment processing fees, marketplace or transaction fees, and the marketing spend assigned to acquiring or serving the cohort. Contribution profit = realized revenue − variable costs. This is profit before fixed costs such as salaries, rent and software. For marketing, state whether you use attributed spend or a blended allocation, then apply the same method to every cohort; do not compare one cohort using attributed ads with another using total channel spend.
Cohort size is the number of unique customers who entered during the defined acquisition window. Keep that original count fixed when calculating contribution profit per customer: total cohort contribution profit ÷ original cohort size. Customers who have not reordered remain in the denominator. Include repeat orders within the observation window, and record refunds in the period or cohort they relate to rather than silently treating refunded sales as retained revenue. A profitability dashboard should therefore separate revenue, refunds, shipping, fees and advertising before presenting the contribution result. ORVX’s product profitability reporting follows the same need to evaluate sales after COGS, returns and ad spend.
| Component | Calculation and treatment |
|---|---|
| Realized revenue | Delivered order value − discounts − refunds; exclude cancelled and returned orders from retained revenue |
| Variable costs | COGS + merchant-paid fulfilment and shipping + payment fees + other transaction fees + allocated marketing spend |
| Contribution profit | Realized revenue − variable costs |
| Per-customer result | Total contribution profit ÷ the original number of customers in the cohort |
A Worked Example from First Order to Payback
Assume 100 customers acquired in January through paid media. Their first delivered orders show $8,400 in sales. After $600 of discounts and $300 of refunds, realized first-order revenue is $7,500. The cohort’s paid acquisition cost is $4,500, or $45 per customer.
The table separates revenue, variable costs and acquisition cost so the payback point is visible rather than hidden inside a single margin.
To complete the example, assume the product cost on those delivered orders is $3,000 and other order-variable costs, such as payment and fulfilment charges, are $1,200. First-order contribution before advertising is therefore $7,500 − $3,000 − $1,200 = $3,300. After the $4,500 acquisition cost, the cohort is at a $1,200 loss, or negative $12 per acquired customer. The first order has not paid back acquisition.
Payback requires later contribution, not merely another order. Suppose the same customers generate a later delivered order period with $2,400 of realized revenue, $900 of product cost and $300 of other variable costs. That period adds $1,200 of contribution, reducing the cumulative loss from $1,200 to zero. In this simplified example, payback occurs at the end of that period. If only 80 customers return, the required contribution per returning customer is $1,200 ÷ 80, or $15; the cohort does not need every customer to buy again, but it does need enough profitable repeat contribution.
The useful decision is whether to preserve, reduce or test paid acquisition under the stated cost assumptions. A negative first-order result is not automatically a reason to stop if repeat contribution has a credible path to recover it; a positive first-order contribution does not prove the cohort will remain profitable after later returns, service costs or additional acquisition. The calculation also does not explain why customers repeat, which products create the margin, or whether the acquisition channel is being credited consistently. Those questions require product, customer and advertising views alongside the cohort total.
| Period | Realized revenue | Product cost | Other variable costs | Contribution before acquisition | Acquisition cost | Cumulative cohort profit |
|---|---|---|---|---|---|---|
| First delivered order | $7,500 | $3,000 | $1,200 | $3,300 | $4,500 | −$1,200 |
| Later delivered orders | $2,400 | $900 | $300 | $1,200 | — | $0 |
The Accounting Choices That Change the Result
The cohort start date must be operationally consistent. Grouping customers by first order date in one report and first delivered order date in another can move customers between months, especially when fulfilment takes time. Payment date, order date and delivery date each answer a different question. Choose one, document it, and use the same rule for every cohort.
Revenue basis changes the result before any cost is considered. Gross order value includes amounts later removed by discounts, refunds, cancellations or returns. Net revenue should reflect what was actually retained from delivered orders. Taxes and customer-paid shipping also need an explicit treatment: exclude pass-through taxes from revenue, for example, and include shipping only when it is part of the merchant’s economic revenue. Mixing gross sales in one cohort with net sales in another makes growth look like profitability.
Cost coverage is equally important. Product cost alone produces gross profit, not the profit available after fulfilment, payment fees, packaging, discounts, returns and advertising. If one cohort includes ad spend and another omits it, the comparison is not valid. The same applies to attribution: assigning all advertising to the first order, spreading it across repeat orders, or using platform-attributed revenue will produce different answers. Keep blended MER separate from attributed ROAS, and state which spend and revenue each uses. Advertising profitability analysis covers that distinction.
Returns create a timing problem as well as a revenue adjustment. A customer acquired in March may generate an order in March and a refund in April. Decide whether the cohort is measured by order activity date or by the date revenue becomes final, then apply that choice consistently. A recent cohort is also only partially observed: its early profit may look poor because later repeat orders and returns have not arrived. Compare cohorts at the same elapsed age, or label newer cohorts as incomplete rather than ranking them against mature ones.
- Before comparing cohorts, record the cohort date, revenue basis, included costs, attribution rule, tax treatment, return window and observation age.
| Choice | If handled inconsistently | Control |
|---|---|---|
| Gross versus net revenue | Discounts, refunds or cancellations inflate apparent profit | Use one retained-revenue definition |
| Missing costs | Gross profit is mistaken for net profit | Include the same cost categories in every cohort |
| Returns and partial periods | Older cohorts appear worse or newer cohorts appear incomplete | Compare cohorts at the same observation age |
What the Number Can and Cannot Tell You
Cohort profitability is useful when the decision concerns the economics of a group acquired during the same period. It can show whether customers from a campaign, channel, month or product entry point are generating enough contribution to justify the acquisition cost. That makes it relevant to channel budget reviews: compare cohorts on the same elapsed-age basis before shifting spend, rather than ranking channels by revenue or attributed ROAS alone. Advertising profitability reporting helps keep MER and attributed ROAS separate while examining the profit left after advertising.
It also supports retention investment. If a cohort is unprofitable after its first order but becomes profitable after repeat purchases, the result is evidence for testing retention work against that specific customer group. If repeat orders add revenue but little gross profit after product costs, discounts and returns, more retention activity is not automatically justified. Customer-level profitability and repeat behaviour provide the detail needed to see whether the cohort result is broad or driven by a small number of customers. This is different from treating every future order as equally valuable.
The metric can support payback targets as well. A business might define a target such as recovering acquisition cost within 90 days, then compare each cohort's cumulative realized profit at day 30, 60 and 90. This measures progress against a target; it does not establish whether a cohort will reach it. A product profitability view can add product and collection economics when a cohort's result needs to be explained.
Cohort profitability is not cash profit. It may recognize delivered-order revenue and associated costs while cash is still tied up in inventory, payment settlement, supplier invoices or refunds that have not yet cleared. It is not company-wide profit either: corporate overhead, payroll, financing, tax and other costs may sit outside the cohort allocation. Those figures belong in financial reporting, not in a customer cohort comparison.
Nor is it the same as customer lifetime value. Cohort profitability describes realized performance over a stated observation window. LTV includes expected future purchases and therefore depends on assumptions about retention, order frequency, margin and time horizon. A customer profitability report can show realized customer economics; it should not be read as a promise of future value. ORVX can report the store's own historical numbers and keep currencies separate, but the metric cannot establish causation, predict demand response or prove that increasing a budget or retention spend will produce the same result.
Turning the Result into an Operating Decision
Review cohorts on the same maturity schedule: compare each group after 30, 60 or 90 days rather than comparing a three-month-old cohort with one acquired last week. Set thresholds before reviewing the result. For example, define a minimum acceptable contribution margin, a maximum payback period and a repeat-purchase rate worth pursuing. These thresholds should reflect your cash constraints and historical performance, not a generic benchmark.
When a cohort crosses a threshold, investigate the components before changing spend. Separate the effect of acquisition cost from changes in realized revenue, product mix, discounts, refunds, COGS and repeat orders. A lower margin may come from higher returns rather than weaker demand; a stronger revenue figure may be offset by expensive acquisition. Advertising profitability analysis helps keep MER, attributed ROAS and contribution distinct, while customer profitability analytics helps identify whether repeat behaviour is improving the cohort result.
Treat incomplete or unusually small cohorts as provisional. Label them by elapsed time, show the customer count beside the result and avoid making a scale-or-stop decision from one or two customers. If a cohort is too small for a stable comparison, combine adjacent cohorts only when they share the same acquisition and offer conditions, or wait for more delivered orders. Investigate material changes against the last comparable cohort, document the suspected driver, then choose a measured response: scale, reduce, fix the economics, test or stop. Recheck the next comparable cohort rather than treating the first response as proof.
Cohort profitability is useful when the cohort definition, time window and cost treatment are consistent. Calculate it from net revenue through variable costs, include acquisition cost when assessing payback, and use the result to change budgets or retention efforts only after checking that the cohort is mature enough to compare.