Break even ROAS is the ratio where campaign revenue exactly covers the cost of goods plus the ad spend. One divided by your gross margin. That is the entire calculation, and most accounts I audit have never done it.
Everything else in ecommerce reporting sits on top of that one number. Without it, a 3.2x is just a number on a screen. You cannot say whether an ad set should be cut, or what figure belongs in a bid strategy.
This is for anyone running ecommerce ads against a client target that nobody can explain.
What break even ROAS actually is
It marks the point where a sale stops losing money, and it claims nothing more than that.
Take your gross margin as a decimal, then divide one by it. A 40% margin gives 2.5. At 2.5x, a hundred pounds of spend returns two hundred and fifty pounds of revenue. Of that, one hundred and fifty is cost of goods and one hundred went to the platform. Nothing is left.
Gross margin here means what remains from a sale after every cost that scales with that sale. Product cost, inbound freight, pick and pack, delivery you absorb, payment processing, and any discount code applied at checkout.
Overheads stay out of it. Rent, salaries and software do not move when you sell one more unit. They belong in the target, not the floor.
The margin range I see, and why the spread matters
Across the ecommerce clients I have run ads for, gross margin lands somewhere between 20% and 50%.
That range puts the floor between 5.0x and 2.0x, and the spread is the whole point. A 20% margin retailer and a 50% margin brand can post an identical 3.0x in the same month. For one it is a loss, for the other it is comfortable.
So a return figure with no margin attached carries almost no information. When somebody tells me their account runs at 2.8x, my first question is the gross margin. That 2.8x is either well clear of the floor or well underneath it.
None of this is available in the ad platform. Break even ROAS gets worked out in a spreadsheet, using numbers that live in the client’s finance system.
What belongs in the margin, and what gets left out
Four costs get forgotten more or less every time.
Delivery you absorb is the biggest. A free shipping threshold is a discount you pay on most orders. It rarely appears in the margin a client quotes from memory.
Discount codes come second, and they bite harder than people expect. A 20% off code on a 40% margin product does not remove a fifth of the margin. It removes half.
Payment processing and marketplace fees are small individually and relentless in aggregate. Returns are the awkward one, because the refunded order takes its revenue back while the postage stays spent.
If precise numbers are not available, treat the client’s stated margin as optimistic. An approximate break even ROAS that people actually use beats an exact one nobody ever calculated.
Break even ROAS is a floor, not a target
Hitting break even ROAS exactly means the business made nothing at all that month.
Your target sits above the floor by whatever the company needs to cover fixed costs and produce a profit. That gap is a business decision. It should come from the client, in writing, before launch.
Profit margin after overheads is a different question from gross margin. Confusing the two is how targets end up quietly guaranteeing losses.
There is one honest exception. A business buying first orders below the floor to win repeat customers is choosing a strategy. It only works when somebody measures the repeat rate, and that measurement is far rarer than the strategy.
Why average order value does not move break even ROAS
Raising basket size does not shift the ratio at all. It changes what the ratio costs you to reach.
Break even ROAS is a percentage relationship, so it depends only on margin. Double your average order value and the floor stays exactly where you left it.
What moves is the allowable cost per purchase. At a 2.5x floor, an eighty pound order allows thirty two pounds of spend. A one hundred and sixty pound order allows sixty four, which is a much easier auction to win.
For scale, IRP Commerce put UK and EU ecommerce average order value at £127.06 in June 2026, with conversion rate at 2.03%. Read their own caveat with it: they describe their figures as informed estimates rather than audited values, drawn from last click data across SME and mid market stores.
Two accounts where the same ratio meant different things
Two of my own numbers show why the floor decides the verdict.
An EU fashion ecommerce client ran at 4.7x across the full year 2022, on Meta and Google combined, with the majority of sales coming through Google Search and Shopping. That was 1,715 orders at €14.18 each. On a 30% margin the floor would sit at 3.33x, so 4.7x was genuinely profitable rather than merely impressive.
A US electric mobility client produced more than 63,000 purchases at $5.05 each across the full year 2022, running Google Search, Performance Max and Display together. Five dollars per purchase sounds unarguable.
It still tells you nothing about profit until you know the order value and the margin behind it. Cost per purchase is the operational number, so it drives daily decisions. Break even ROAS is the judgement that decides whether those decisions were worth making.
What break even ROAS changes inside the account
Once the floor exists, four recurring arguments simply end.
First, cutting becomes mechanical. Any ad set sitting under the floor for long enough to be believable is losing money, so the argument about giving it one more week is over.
Target ROAS bidding gains a real input instead of a guess. Google’s own documentation says Search and Shopping campaigns need at least 15 conversions in the past 30 days before that strategy is available, with Demand Gen needing 50 in 35 days, and Google advises waiting four weeks or three conversion cycles after a reporting change before judging anything.
Budget allocation gets easier, because ad sets are ranked against a fixed line rather than against each other. Finally, scaling gets a stopping rule instead of a feeling.
The client who cannot tell you their gross margin
This happens far more often than the textbooks suggest.
One of the ecommerce accounts I run in 2026 has never disclosed its gross margin to me, and I have asked. That is a reasonable commercial choice on their side, and it means I will not publish a profit verdict on that account, here or anywhere else.
So I cannot publish a break even ROAS for them, and I optimise against a proxy instead. Cost per purchase measured against the ceiling they are willing to state, plus the direction of travel week on week.
If you are in the same position, ask for a range rather than a number. Then ask for the allowable cost per order, which is the same information dressed differently. Should both requests fail, build the margin up from the parts you can see: retail price, visible shipping cost, and the discount rate on their own site.
Where the floor stops being the right number
Break even ROAS answers one question well and several others badly.
First, it assumes the channel you are measuring is the only one contributing. In an account running Meta, Google and email together, a single channel figure gets compared against a whole business floor. That understates the channel every time.
Second, it assumes each order is a standalone event, which breaks the moment repeat purchase matters. It also assumes the revenue you see is the revenue that stays, which returns quietly undo.
None of that makes the calculation wrong. Instead it makes the calculation a starting point. Work out the floor first, then decide which of those corrections your business actually needs, because most shops need one and almost none need all three.
What to change this week
Three concrete steps.
Get the gross margin figure out of the client this week, as a range if that is all they will give, and write it at the top of the reporting sheet where everyone can see it.
Then divide one by it and put the result next to every campaign in your dashboard, so the floor is visible at the moment anyone looks at performance.
Then find the ad sets that have sat below that line for a full month and decide about each of them, rather than leaving them running because nobody wants the conversation.
Putting a defensible number on a lead runs the same arithmetic on the lead generation side, the purchase costs I can actually show gives two disclosed volumes with their date ranges attached, revenue over spend is not profit covers how the ratio gets misread in reporting, and what an acceptable return depends on explains why no industry answer exists. For worked examples on live accounts, the case studies have the numbers. To have your floor calculated and your targets rebuilt on top of it, book a teardown.