Revenue is not profit
Marketing reports are full of revenue. Platforms report conversion valueThe monetary value an advertiser assigns to a conversion, such as the order's revenue or a lead's expected worth.; analytics reports transaction revenue; campaign summaries quote return on ad spend (ROASRevenue generated per unit of ad spend.) as revenue divided by media cost. All of these are useful, and none of them says whether the business made money.
Between the price a customer pays and the profit the business keeps sit several layers of cost. Some vary with each sale: the product itself, packing and delivery, card and payment fees, marketplace commission, the cost of handling returns, the discount the customer used. Others are fixed for the period regardless of how many sales happen: rent, salaries, software, the website. Marketing spend sits between the two, variable in principle but decided in advance.
Treating revenue as the measure of marketing success hides the first layer entirely. A campaign that drives a large volume of heavily discounted, frequently returned, expensive-to-ship orders can look excellent on revenue and return on ad spend and still lose money on every order. The finance team sees this in the profit and loss account months later; marketing should see it before the campaign starts.
The fix is one number, contribution margin, agreed with finance and used as the base for every target. It is not complicated to calculate, but it needs inputs only finance and operations hold, which is why it is often missing from marketing's planning.
From gross margin to contribution margin
Gross margin is the familiar first step: revenue less the cost of goods sold. Finance reports it at company and often at product level. It is the right starting point and the wrong stopping point, because several costs that rise with every order sit below it in the accounts.
| Layer | What it subtracts | Where the figure usually comes from |
|---|---|---|
| Revenue | Starting point: price paid, excluding sales tax | Order system or ledger |
| Less discounts | Promotional codes, bundle discounts, loyalty redemptions | Order system |
| Net revenue | What the customer actually paid for goods | Ledger |
| Less cost of goods | Product cost, including landed import and duty where relevant | Finance, by product |
| Gross margin | The figure most finance reports show | Ledger |
| Less fulfilment | Picking, packing, packaging, outbound delivery | Operations or logistics invoices |
| Less payment and platform fees | Card processing, payment providers, marketplace commission | Finance |
| Less returns cost | Refunded revenue on returned items, return shipping, write-offs | Operations and finance |
| Less other variable costs | Customer service per order, affiliate commission, anything that scales with each sale | Finance |
| Contribution margin | What each sale contributes to marketing, fixed costs and profit | Agreed between marketing and finance |
Express the result as both a currency amount per order and a share of net revenue. The amount per order is what break-even cost per acquisitionHighest cost per acquisition at which a sale still covers its variable costs. needs; the share is what break-even return on ad spend needs. Work at the level of average order value (AOVAverage revenue per order.) for the segment, not for the whole business, because averages across very different products hide the cases that matter.
The comparison of gross margin and contribution margin sets out which of the two margins each pricing, finance and marketing decision needs.
Deriving break-even ROAS and CPA
Once contribution margin is agreed, both break-even targets follow directly. The break-even return on ad spend (Break-even ROASMinimum ROAS to avoid a loss.) is the return at which marketing spend exactly uses up the contribution from the revenue it generated:
Break-even ROAS = Revenue divided by contribution margin before marketing(ratio)
A business keeping a large share of each sale as contribution can tolerate a low return; a business keeping only a thin share needs a high one. The break-even ROAS glossary entry gives the definition in short form, and the break-even calculator does the arithmetic.
The break-even cost per acquisition (CPACost of each conversion.) works in currency rather than ratios. It is the most marketing can pay to win one order without losing money on it:
Break-even CPA = Average order value multiplied by contribution margin share(currency)
Equivalently, it is the contribution per order in currency. For lead generation, the same logic runs backwards through the funnel: the contribution of a customer, multiplied by the share of leads that become customers, gives the most that can be paid per lead. The CPA calculator and the ROAS calculator check current performance against both lines.
Both figures assume attributed revenue is real. If the attribution model over-credits marketing, which most platform models do to some extent, a campaign apparently at break-even is in reality below it. That uncertainty is one reason targets sit above break-even rather than on it.
Returns, discounts and fulfilment in the margin
The three costs most often left out of marketing's margin are also the ones most likely to vary by campaign.
Returns. In categories such as fashion and footwear, a large share of orders may come backShare of orders or units sold that customers sent back.. A returned order generates revenue in the platform report and none in the bank, and it also costs return shipping and handling, and sometimes a write-off. Use the return rate for the specific product or category, and if possible for the channel, since some channels attract more speculative buyers. Platforms report conversion value at the moment of purchase, so returns never appear in their figures.
Discounts. A promotion that lowers the price lowers the contribution margin by the full discount, because the product, delivery and fees cost the same. A campaign built around a discount code needs its own, lower margin and therefore a higher break-even return. Running discount campaigns against the full-price target is a classic way to lose money while reporting growth. The ROAS calculator for discount promotions works through the post-discount margin and the higher break-even it produces.
Fulfilment. Delivery cost often varies with order size, location and speed. Free delivery thresholdsThe order value above which an online store delivers for free, set to encourage larger baskets without giving away too much margin., heavy or bulky items, and international orders change the contribution of each order considerably. Campaigns that skew towards small orders or distant regions carry a lower margin than the average suggests.
Shopping and catalogue campaigns make these differences visible because spend runs per product, and the shopping and catalogue ads guide covers how to reflect margin in product grouping and bidding there.
Targets above break-even: the profit you want
Break-even is the line where marketing earns nothing. A target needs to sit above it by an amount that reflects two things: the profit the business wants from each order, and the uncertainty in the numbers.
The profit buffer is a business decision. Finance and leadership should state how much of each order's contribution they want retained as profit after marketing, and the target is set so that marketing can only spend the rest. A business chasing growth may accept a thin buffer; one protecting cash will want a thick one.
The uncertainty buffer reflects how confident the team is in the margin inputs and the attribution. If the return rate is volatile, or platform-reported revenue is known to overstate, the target should be stricter to compensate. That buffer can be reduced as measurement improves, for example after an incrementality test shows how much of the attributed revenue was truly caused by the campaign.
Some businesses deliberately set first-order targets below break-even because customers come back. That can be right, but it is a lifetime value decision, not a contribution margin one, and it needs evidence about repeat purchase and a payback period finance can fund. The growth root's guide to lifetime value and acquisition cost covers that judgement; customer acquisition cost (CACCost to acquire one new customer.) is defined and analysed there.
Margin by product, segment and channel
A single contribution margin for the whole business is better than none, but most marketing decisions happen at a finer level. Products differ in cost and return rate, customer segments differ in order size and discount use, and channels differ in the kind of order they attract.
Start by calculating contribution margin for the product lines or categories that account for most revenue, and group the rest. Then check whether segments, such as new versus returning customers or domestic versus international, differ enough to need separate targets. Finally, look at whether particular channels systematically produce orders with different discount use, return rates or delivery costs. These margins come before any marketing cost; once agency fees, tools and team time are shared out across channels or products for reporting, the guide to allocating shared marketing costs explains why that allocated view should sit beside the direct cost and never set the targets.
The output is a short table of break-even and target returns by product group or segment, agreed with finance and refreshed when prices, costs or return rates change. That table is the bridge between the finance team's margin data and the bidding targets set in each platform. When spend grows, the next question is whether each additional unit of spend still clears the target, which the marginal returns and forecasting guide takes up. The unit economics hub shows how the two guides fit together. Applied to a whole period rather than a single order, the same margin line sits in the middle of the marketing P&L template, between attributed revenue and contribution after marketing.
Break-even targets for a footwear retailer
Treadfleck Shoes sells trainers at an average order value of £110 after discounts. Cost of goods is £44, giving a gross margin of 60%. Fulfilment costs £7 per order, payment fees £2 and the expected cost of returns, with a 25% return rate, averages £16 per order including refunded margin and return shipping. Contribution before marketing is therefore £41 per order, 37% of revenue. Break-even return on ad spend is 1 ÷ 0.37, about 2.7, and break-even cost per acquisition is £41. Finance wants £15 of profit per order retained, so the target cost per acquisition is £26 and the target return is £110 ÷ £26, about 4.2. The previous target, set on gross margin at 2.0, had been approving orders that lost money.
Applied to search ads, the paid search channel page for online sales turns these break-even figures into a value target per margin band and a budget ceiling at marginal break-even.
If campaigns already meet their return targets and the business still loses money, the page on a strong ROAS that is not profitable checks the margin behind the target alongside refunds, overlapping attribution and costs outside media.
The same margin arithmetic decides the discount depth experiment, which compares deeper, shallower and no discounts on gross profit per recipient after returns.
Margin by product group is also the first gate in the guide to choosing which products to put marketing behind, which weighs it with demand, stock, repeat purchase and returns to decide which lines get paid support at all.
Common mistakes
- Setting targets on gross margin. It ignores delivery, fees and returns, so break-even looks lower than it is and marketing overspends.
- Using platform conversion value as revenue. It is recorded at purchase, before returns and sometimes including tax and shipping, and it overstates what the business keeps.
- One margin for every product. Account-wide targets subsidise low-margin products with high-margin ones and hide which campaigns lose money.
- Running discount campaigns at full-price targets. Every discount comes straight out of contribution, so the break-even return rises.
- Bidding to break-even. It leaves no profit and no room for attribution error; the target needs a buffer.
- Never refreshing the inputs. Price changes, supplier costs and return rates move, and targets set once drift out of line within months.
