Contribution margin is what is left from an order once you subtract every cost that moves with that order. Cost of goods, shipping, packaging, payment processing, the discount code, the share that comes back as a refund. Not the rent, not the salaries, not the software you would pay for anyway. That figure, as a percentage of revenue, is the only honest source of a bid target, because one divided by it is the return on ad spend at which you break even.
This is for you if somebody handed you a target and nobody could explain where it came from.
What contribution margin is, and what gross margin misses
Gross margin stops at cost of goods. That is a manufacturing view of the business, and it is far too generous for anyone buying media.
The version you need also removes the costs that arrive with each additional order: pick and pack, shipping, packaging, payment processing at two or three percent, and the orders that come back. What is left is the profit per order before a single fixed cost is paid.
Rent, salaries and your store subscription stay out. None of them changes when you sell one more unit, so they belong in a different calculation entirely.
That distinction between contribution margin and the gross figure on the accounts is wider than most people expect. A store quoting 45% gross can easily sit near 30% once fulfilment and processing come out, and lower again after refunds.
Bid against the first number and you will spend right up to a break-even that does not exist.
Why your bid target should come from contribution margin
One divided by your contribution rate is your break-even return on ad spend. That is the entire mechanism, and it fits on a napkin.
At 50%, break-even sits at 2.0x. At 40% it sits at 2.5x. Drop to 20% and you need 5.0x before the advertising has paid for itself.
Across the ecommerce clients I run, the rate lands somewhere between 20% and 50%. So break-even can be anywhere from 2.0x to 5.0x depending on whose account is open, which is why a borrowed industry target is worth nothing.
Smart bidding cannot see any of this. Target ROAS is the only place the arithmetic enters the auction, so the number you type is where your margin gets respected or gets ignored.
A worked example on a 20% to 50% margin range
Take an average order value of $114.70. Shopify published that figure on 2 December 2025 as the average across its full merchant base over Black Friday Cyber Monday weekend, which makes it first-party and dated, unlike most benchmarks in this space.
At a 40% contribution margin, each order leaves $45.88. Break-even cost per purchase is therefore $45.88, and break-even return on ad spend is 2.5x.
Move to 20% and the same order leaves $22.94, so you need 5.0x. Move to 50% and it leaves $57.35, where 2.0x will do.
Three lines, one business, three different margins. Nothing about the advertising changed. What you can afford to pay for a sale moved by a factor of two and a half.
The same account, three very different purchase costs
An average cost per purchase hides more than it reveals.
In a US flag and banner ecommerce account I run on Meta, June to August 2026, six sales ad sets averaged $67.24 across 64 purchases. Remove one failed location test and the average becomes $53.72 across 61 purchases.
Underneath those averages the spread is severe. The best ad set bought purchases at $29.37. The worst active one paid $93.96, for the same offer, from the same creative library, in the same three months.
Set that against a generic $45.88 break-even and the account would split in two: part comfortably profitable, part burning money, with the average looking merely disappointing.
I could not run that sum on this client. They have never disclosed a contribution margin, which is far more common than the textbooks admit.
What contribution margin does to a discount code
A discount comes straight out of contribution, because cost of goods does not fall when the price does.
Run the same $114.70 order at a 40% rate, then take 20% off. Revenue drops to $91.76. Cost of goods stays at $68.82. Contribution falls from $45.88 to $22.94, which is exactly half.
The rate on that discounted order is now 25%, so break-even return on ad spend jumps from 2.5x to 4.0x.
Almost nobody moves the target when the promotion launches. The campaign carries on buying at 2.6x, the dashboard stays green, and the month loses money quietly.
This is the single most common way I see a profitable account turn unprofitable without anyone noticing at the time.
Returns are the line that never reaches the media buyer
On every ecommerce account I have run, returns are handled by the client’s own in-house team.
They get processed weeks after the sale, in a different system, by people who never open an ad platform. So the revenue a media buyer optimises towards has not been reduced by anything that came back.
That makes most bid targets pre-returns targets. Returns quietly reduce the contribution margin you thought you were bidding to. If a tenth of orders are refunded, a 2.5x break-even is really nearer 2.8x, and nobody in the account has been told.
Say the awkward part plainly: no credible primary source exists for return rate by category, apparel included. You cannot look yours up. Ask the client, then write the date beside the answer.
How to turn contribution margin into a tROAS you can enter
Four steps, and none of them needs a template.
Pull last quarter’s average order value from the store rather than the ad platform. Subtract cost of goods, fulfilment, packaging and payment processing for a typical order. Reduce the result by your refund rate. Divide what remains by the order value to get a percentage.
One divided by that percentage is break-even. Then decide what profit you want above it. A target set at 1.3 times break-even keeps roughly 23% of the contribution once the media is paid for.
Enter that figure and write the date on it. Recalculate when a promotion launches, when shipping rates move, and when a supplier puts prices up.
When nobody will give you the margin at all
Plenty of owners will not share cost of goods, and a few genuinely do not know it.
Two approaches still work. Ask once, in writing, for one blended percentage rather than a product-level breakdown. People refuse a spreadsheet and answer a sentence.
Failing that, work backwards from a figure they will confirm. IRP Commerce reports UK and EU ecommerce running at a cost of acquisition of 9.34% of revenue in June 2026, alongside a £127.06 average order value. IRP calls its own numbers informed estimates rather than audited values, built from last-click data across SME and mid-market stores.
Use that to open the conversation rather than to settle it. An owner who will not name a contribution margin will usually correct one that looks wrong.
What to change this week
Three things, and the first is the only slow one.
Work out one contribution margin for your best-selling product, using last quarter’s actual costs rather than the ones on the price list. It takes about an hour and the answer is usually lower than expected.
Then divide one by it, compare the result against the target sitting in your campaigns right now, and write the difference down where somebody else can see it.
Last, set a reminder for the first day of your next promotion to run the sum again. A discount changes the arithmetic on the day it launches, not when the month closes.
Working out what a lead is worth applies the same logic on the lead generation side. The two purchase costs I can date and evidence show what these targets look like at real volume. Total revenue over total spend is the account level check on all of it, and the difference between a media ratio and a profit number covers why those two keep getting confused. Checkout fees and shipping logic live in the site itself, which is where my web development work usually starts. To have your own numbers worked out properly, book a teardown.