Your ROAS is a vanity number.
This is the one that pays you.
ROAS Genius models your whole store the way a trading desk models a book: revenue in, every leak accounted for, one contribution figure at the bottom. Change an input and the waterfall redraws instantly.
What if you moved spend?
holdEfficiency decays as spend rises. This model applies a 0.78 elasticity exponent, so doubling spend returns roughly 1.72× the revenue rather than 2×. Calibrate against your own historical curve before committing budget.
Channel desks
The engine above answers the portfolio question. When you need to price a single channel, these desks isolate the maths that channel actually runs on.
Facebook Ads ROAS calculator
Break-even ROAS, target ROAS, and the gap between what Meta reports and what your bank account records after post-iOS attribution loss.
TIKTOKTikTok break-even calculator
AOV, COGS, tiered shipping and TikTok Shop commission resolved into a hard break-even CPA for impulse-led catalogues.
FUNNELCPM to ROAS funnel calculator
Trace CPM, click-through rate and conversion rate down to ROAS, and find out which stage of the funnel is leaking the most money.
How to read the numbers
Blended MER beats channel ROAS
Marketing efficiency ratio is total revenue divided by total advertising spend. Every dollar you spent, every dollar that arrived. No attribution window, no deduplication argument, no platform incentive to flatter itself. If three channels each claim a 3× return and your blended MER is 2.1, the difference is double counting.
Blended MER = total revenue ÷ total ad spendMER is a ratio, not a verdict. A 2.1 MER is excellent for a 65% margin skincare brand and fatal for a 28% margin electronics reseller. That is why the engine always prints your break-even MER beside it.
Break-even MER is set by your margin, not your ambition
Contribution margin is what remains from a dollar of revenue after the costs that move with every order: goods, freight, payment fees and returns. Invert it and you have the MER at which advertising exactly pays for itself.
Break-even MER = 1 ÷ contribution marginA 40% contribution margin breaks even at 2.50. At 3.10 you are earning 60 cents of contribution on every advertising dollar before overheads. At 2.20 you are buying revenue with equity.
Net contribution margin is the number to defend
Revenue minus goods, fulfilment, fees, returns and advertising, divided by revenue. It is the share of every sale left to cover the team, the rent and the profit. Watch it weekly. It moves before net profit does, and it moves for reasons you can act on.
- Falling AOV against flat CAC is usually discount depth or a shift in product mix, not a media problem.
- Rising freight per order compounds quietly. Ninety cents across 2,000 orders is $1,800 of margin a month.
- Returns above 8% on apparel are typically a sizing or photography failure, and no amount of creative testing fixes them.
- Payment fees drifting up often means a rise in international cards or instalment providers, which carry higher take rates.
Scale to marginal contribution, not to a round number
Efficiency falls as spend rises because you exhaust the cheapest demand first. The right question is never whether ROAS sits above 3. It is whether the next $10,000 returns more than $10,000 of contribution. Drag the sensitivity slider until simulated net profit stops climbing. That inflection, not a threshold, is where you stop.
A working weekly cadence
- Pull revenue, spend and orders for the trailing seven days and compute blended MER.
- Compare it against break-even MER. Record the spread, not the raw figure.
- If the spread widened, add 10 to 15% spend and re-measure after one full purchase cycle.
- If it narrowed two weeks running, fix margin before touching the media plan.
- Re-derive break-even MER whenever COGS, freight or discount policy changes. It is not a constant.
What this model deliberately ignores
Cash timing, inventory carrying cost, tax and payment settlement delays are all excluded. A store can be contribution-positive and still run out of cash if it pays suppliers in 30 days and collects from a marketplace in 60. Treat this as a margin model, not a cash flow forecast.
Questions operators actually ask
What is a good blended MER?
There is no universal figure. A good MER is any MER comfortably above your break-even MER, which is one divided by your contribution margin. A store at 35% contribution breaks even at 2.86, so 3.5 is healthy. A store at 70% contribution breaks even at 1.43, and 3.5 would mean it is badly underspending.
Why does platform ROAS look better than my bank account?
Platforms attribute on their own click and view windows, and they cannot see each other, so the same order gets claimed more than once. They also report gross revenue before discounts, returns, payment fees, freight and goods. Blended MER measured against total store revenue removes the attribution argument entirely, because it only uses two numbers you already trust.
Should cost of goods include freight?
Keep them separate here. Goods scale with revenue and are entered as a percentage. Fulfilment and freight scale with order count and are entered per order. Splitting them means the model reacts correctly when average order value changes, which is exactly the moment blended figures mislead you.
How do I handle returning customers?
Blended MER already includes repeat revenue, which flatters a mature store and punishes a young one. To isolate acquisition, run the model twice: once on first-order revenue only, once on the full book. The spread between them is what your retention programme is actually contributing.
Why does the simulator show revenue growing slower than spend?
Because it almost always does. Each additional dollar reaches a less interested audience than the one before it. The elasticity exponent of 0.78 is a conservative middle estimate drawn from typical direct-to-consumer scaling curves. If your own history shows a steeper or flatter response, trust your history.
Is my data stored anywhere?
No. Every calculation runs in your browser. Nothing you type is transmitted to a server, logged or saved. Reload the page and the inputs reset to their defaults.