ROAS
What it means
ROAS is the revenue Amazon credits to your ads divided by what those ads cost, written as a multiple. Spend €500 and earn €2,000 of ad-attributed sales and it reads 4.0 — every euro of spend brought back four euros of revenue. Organic sales are excluded entirely, so this judges the campaign, not the business.
This page carries four ROAS numbers and they are not the same measurement. Each tab divides a different revenue by a different spend, so a 4.0 here and a 4.0 on the Meta tab are unrelated facts. Never compare them to each other, and never add them up.
Show the math
Formula and a worked example
Ads Attr. SalesAds Attr. SalesWhat advertising directly drove; weigh against spend. is revenue Amazon credits to your ads, for orders placed inside Amazon's attribution window after a click.
Ad SpendAd SpendRead it next to ROAS and sales; spend only helps if the return holds. is what Amazon charged you for clicks in the period.
Worked example. A campaign spends €500 in March and Amazon credits it with €2,000 of sales. ROAS = 2,000 ÷ 500 = 4.0.
That is the same statement as an ACOSACOSThe share of ad revenue eaten by the ads themselves.Under 20% is healthy of 25%, because 500 ÷ 2,000 = 25%. The two cards on this page are one number written upside down — if you read both, you've read one.
Revenue is not profit. At a 35% gross margin those €2,000 of sales carry €700 of margin, and after the €500 of spend you keep €200. A ROAS of 4.0 was a thin win, not a quadrupling.
It answers the question
For every euro these ads spent, how many came back as revenue Amazon credits to them? It's the scaling question — whether to put more money into a campaign, or take it out.
Why it matters
It's the number that decides whether a campaign stays on. Because it ignores the organic business running alongside, it reads the campaign cleanly, without a good month on the rest of the catalogue flattering a bad one on the ads.
The threshold that matters isn't the band, it's your margin. At a 35% gross margin you need a ROAS above roughly 2.9 for an advertised sale to clear its own cost, because 1 ÷ 0.35 = 2.86. A campaign at 2.5 is losing money on every order while still looking respectable.
What good looks like
These bands grade Amazon's ad-attributed ROAS. The Meta tab is graded lower — 3 excellent, 1.2 poor — because it divides a different numerator, so a Meta campaign at 3.2 and an Amazon campaign at 3.2 are not equally healthy. Read this alongside TACOSTACOSHow dependent your whole business is on ads.Under 8% is healthy: ROAS judges the campaign, TACOS judges what the whole business pays for advertising.
How to improve it
| Lever | What you do | Expect | How long | Watch out for |
|---|---|---|---|---|
| Fast Cut bids where it sits under margin | Halve bids on keywords that have run below your margin threshold for 14 days straight | ROAS up, spend down | 1–2 weeks | You improved the ratio by shrinking. Total units usually fall with it, and the organic rank those sales were feeding falls too. |
| Fast Add negative keywords | Block search terms taking clicks with no attributed sales | ROAS up within a fortnight | 1 week | Impressions and discovery fall. Terms that convert slowly can look like waste in a 14-day window and get blocked by mistake. |
| Slow Fix the listing the ads point at | Better main image, clearer bullets and more reviews on the advertised product | Same spend converts more, so ROAS rises without touching a bid | 3–6 weeks | Nothing moves for weeks, and the gain is easy to credit to whatever else you changed that month. |
| Slow Widen the margin underneath it | Raise price or cut unit cost so the ROAS you already have clears its threshold | ROAS roughly unchanged, profit up | 1 quarter | A price rise costs conversion, so ROAS gets worse before margin catches up. Cutting unit cost usually means a bigger order and cash tied up in stock. |
Every lever costs something somewhere. The last column is the one to read twice.
Read it with
ROAS on its own gives a campaign's return. Next to spend it tells you whether that return survives being scaled.
Scaling that held
More money in and a better return out. Rare, and it usually means you found headroom on terms that were already working.
Buying volume
Extra spend is producing sales at a worse rate. A decision rather than a problem while it clears your margin — a problem the moment it doesn't.
Trimmed the waste
You cut the losing keywords and the return improved. Watch that you didn't cut the terms feeding organic rank.
The real alarm
Spending less and earning less per euro. Something changed outside the campaign — price, stock, the Buy Box, or a new competitor on your terms.
They're reciprocals, so they can never disagree — 4.0 is 25%, 2.0 is 50%. The reason both exist is habit: ACOS reads naturally against a margin percentage, ROAS reads naturally against a budget decision. Pick the one that matches the question and ignore the other.
ROAS sees only advertised sales; TACOS sets ad spend against the whole business. A strong ROAS with a climbing TACOS means the campaigns look efficient while the business leans harder on them every month. Neither number shows that alone.
Common misreads
You made four times your money in revenue. Cost of goods, Amazon's fees and shipping all come out of that before anything is yours. Compare it to your margin, not to 1.0.
Amazon divides sales it attributes inside its own window; Meta divides purchase value it attributes inside its. Different numerators, different windows, different bands. The two are not a like-for-like ranking.
Zero means no attributed sales yet — clicks with no conversions, or a campaign too new for the attribution window to have closed. Read it as no data rather than as a fault.
Also called
Return on ad spend · ad return · return on advertising spend
See yoursYour ROAS by campaign and by product, beside ad spend, ACOS and attributed sales.
Open Amazon Ads →ROAS
What it means
On the Marketing Summary, ROAS is your Total Sales divided by your Ad Spend — your entire Shopify top line over what you paid Meta and Google. It is not an attribution number. Orders from email, organic search, direct traffic and returning customers all sit in the numerator, whether or not an ad touched them.
That makes it a whole-business efficiency ratio wearing an advertising name. A store with a large repeat customer base can post a high figure here while its ads perform poorly, because most of the numerator never came from advertising at all.
Show the math
Formula and a worked example
The formula. ROAS = Total Sales ÷ Ad Spend. Total Sales is your Shopify top line — gross sales plus shipping and tax, less discounts and refunds. Ad Spend is Meta plus Google for the period. Amazon is not in either side of this one.
Worked example. A store takes €200,000 of Total Sales in March and spends €50,000 across Meta and Google. ROAS = 200,000 ÷ 50,000 = 4.0.
That is the same fact as an MERMERWhole-business marketing efficiency across every channel.Under 30% is healthy of 25%, because 50,000 ÷ 200,000 = 25%. MER and this number are reciprocals — one card is the other turned over.
Now suppose €120,000 of those sales came from returning customers who never saw an ad. The ads are being credited with revenue they didn't produce, and the figure would barely move if you turned half of them off for a week.
It answers the question
How much total revenue is the business producing per euro of advertising? It's a pacing number for the whole marketing budget, not a verdict on any campaign.
Why it matters
Platform-reported returns always add up to more revenue than you actually banked, because Meta and Google each claim the same order. This number can't double-count: it divides one top line by one spend total, so it's the honest check on whether the whole budget is affordable.
It's most useful as a trend, not a level. Watching it move while ad spend moves tells you whether extra budget is still buying growth — which is a question no single-platform figure can answer.
What good looks like
Judge this one against its own history first, because the band assumes a business where advertising drives most of the revenue. Note that this card and the MERMERWhole-business marketing efficiency across every channel.Under 30% is healthy card beside it don't line up: 4.0 here equals an MER of 25%, but MER's own excellent mark is 30%, which is a ROAS of 3.33. A store sitting at 3.6 will see one card call it average and the other call it excellent. Both are right; they're graded on different scales.
How to improve it
| Lever | What you do | Expect | How long | Watch out for |
|---|---|---|---|---|
| Fast Move budget to the better channel | Shift spend from the platform with the weaker return into the stronger one, in 10% steps | Blended return up within a fortnight | 1–2 weeks | Channels aren't interchangeable. Cutting the top of the funnel lifts the ratio for a month and starves the retargeting that was converting it. |
| Fast Grow the revenue that costs nothing | Run the email and SMS campaigns to segments you haven't mailed this month | Numerator up with the denominator flat | Same week | It flatters the ratio without improving a single ad. Every extra send costs you list size, so the revenue is borrowed from next quarter. |
| Slow Raise AOV instead of cutting spend | Bundle, add a threshold for free shipping, or push the higher-margin variant | Same ad spend returns more revenue | 1 quarter | Thresholds and bundles are discounts by another name — Contribution Margin can fall while this number rises. |
| Slow Cut the spend that isn't buying anything | Turn off the campaigns that don't move total sales when you pause them for a week | Denominator down, revenue roughly flat | 3–6 weeks | A genuine test costs you a week of that channel's real sales, and brand campaigns often show their loss a quarter later rather than that week. |
Every lever costs something somewhere. The last column is the one to read twice.
Read it with
This number moves when either sales or spend moves. Reading it beside the top line tells you which one it was.
The top line tells you which half of the division moved. Sales up with the ratio up is growth that paid for itself — find the channel that did it and fund it. Sales up with the ratio down means you bought the growth, which is a decision for a launch and expensive as a standing setting. Sales down with the ratio up is a shrinking business posting an improving return, so confirm profit held before calling it a win.
They're reciprocals, so they carry the same information — but their grading bands differ, which is why the two cards can disagree on the same store in the same month. Use MER when you're talking about what marketing costs as a share of revenue, and this one when you're talking about what a budget returns.
This ratio counts revenue from customers you already had; NCPA prices the ones you didn't. A steady return with NCPA climbing means your existing base is holding the number up while new customers get more expensive — the shape that precedes a hard year.
Common misreads
The numerator is every sale the store made, including organic, email, direct and repeat. The ads are credited with revenue they never touched. It's an affordability ratio, not an attribution one.
It never will, and the platform figures won't match each other either — each claims orders the other also claims. This divides one real top line by one real spend, which is exactly why it reads lower.
Both are right. This tab uses Shopify sales over Meta and Google spend. The Summary page adds Amazon sales to the top and Amazon spend to the bottom, which lands on a different number for the same store and month.
Also called
Return on ad spend · blended return · all-channel return
See yoursYour all-channel return beside MER, ad spend and the Meta and Google split.
Open Marketing Summary →ROAS
What it means
On the Meta tab, ROAS is Purchase Value divided by Spend — the purchase value Meta attributes to your ads over what you paid Meta. Both halves are Meta's own measurements, taken inside Meta's attribution window, which is why this figure can report more revenue than your shop actually banked.
It's the platform's own scorecard. That makes it the right number for ranking campaigns, ad sets and creatives against each other, and the wrong number for deciding what the business earned.
Show the math
Formula and a worked example
The formula. ROAS = Purchase Value ÷ Spend.
Purchase ValuePurchase ValueTotal purchase value attributed to this format. Pair it with spend to judge which formats return the most revenue is the value of purchases Meta credits to an ad after someone saw or clicked it, within Meta's window. Two platforms can both claim the same order.
SpendSpendWatch it to control budget and back the winners. is what Meta charged you over the period.
Worked example. An account spends €4,000 in March and Meta reports €12,000 of purchase value. ROAS = 12,000 ÷ 4,000 = 3.0 — exactly the app's excellent mark for this tab.
At the poor mark of 1.2, that same €4,000 would return €4,800 of reported revenue. After cost of goods, that ad set is losing money on every order it claims.
It answers the question
Which of these campaigns, ad sets and creatives is returning most per euro, by Meta's own reckoning? It ranks things inside Meta reliably, even where its absolute level is generous.
Why it matters
It's the number the platform optimises toward, so it's the one that explains Meta's own behaviour — why budget drifted to one ad set, why a creative stopped being served. Reading it lets you follow the auction's logic rather than argue with it.
It also moves fastest of the return figures on the site, because Meta reports attribution continuously. That makes it the earliest warning that a creative has stopped working, days before the change is visible in your own books.
What good looks like
The app grades Meta's ROAS on its own scale: 3.0 or more is paying back well, and under 1.2 is losing money. Between the two, the creative and the audience are the things to work on.
Those marks sit below the 4.0 used on the Amazon and all-channel tabs, and the reason is the denominator underneath each. This one divides only Meta's spend into only Meta's attributed purchase value; the all-channel tabs divide a whole top line, including revenue no ad produced, by a whole budget. Different inputs, different scales — a 3.2 here is not a worse result than a 3.6 on another tab.
Read it by ad set rather than as one account number. Retargeting sits far above cold prospecting, so a healthy account average can hide a prospecting ad set running under 1.0.
How to improve it
| Lever | What you do | Expect | How long | Watch out for |
|---|---|---|---|---|
| Fast Turn off the ad sets under the poor mark | Pause anything still below 1.2 once it has spent enough to be judged | Account figure up immediately | 1 week | The average rises because you removed the weak rows, not because anything improved. Reach narrows and Frequency climbs on what's left. |
| Fast Refresh the creative on the top spender | Put three new concepts against your highest-spending ad, changing the idea rather than the colour | Return recovers where fatigue was the cause | 1–2 weeks | New ads restart learning, so cost per purchase is unstable for several days and the account figure dips before it rises. |
| Slow Widen the audience before it tires | Broaden targeting on the ad sets whose Frequency has been climbing for a fortnight | Return holds instead of decaying | 2–4 weeks | A wider audience converts at a lower rate at first, so the number usually dips for a week or two before the fatigue effect lifts. |
| Slow Fix what happens after the click | Improve the landing page and checkout the ads send traffic to | Same spend returns more purchase value | 1 quarter | It lifts every channel's numbers at once, so you can't credit the gain to Meta — and it needs site work rather than media work. |
Every lever costs something somewhere. The last column is the one to read twice.
Read it with
This figure tells you an ad returned. Whether it can keep returning depends on how much of the audience is left.
Frequency tells you how much audience is left to sell to. A strong return with Frequency climbing works until it doesn't, and the turn is usually sudden — build the next audience while this one still pays. A falling return with Frequency climbing is plain fatigue, and widening the targeting fixes it faster than new creative does. A falling return with Frequency flat means fresh people are seeing the ad and passing, which points at the offer rather than the audience.
The click rate is the top of the chain and the return is the bottom. Clicks up with the return flat means the creative is winning attention it can't convert — often a hook that widens appeal past the people who actually buy.
One is a ratio, the other a price. The return can hold steady while CPA climbs, if higher-value orders are covering for more expensive customers. That's fine until the high spenders run out, and CPA sees it first.
Common misreads
Expected, not broken. Meta credits purchases inside its own attribution window and at its own end, and other platforms claim some of the same orders. Use this figure to rank ads against each other, and your own top line to judge the business.
They measure different things and are graded differently — 3.0 is this tab's excellent mark. The Summary figure divides your whole top line, most of which no Meta ad touched, by your whole budget.
Small, tightly retargeted ad sets post high returns on tiny audiences and collapse the moment you scale them, because the people who were already going to buy run out. Check its Frequency and its reach before you move the budget.
Also called
Return on ad spend · purchase ROAS · Meta return
See yoursYour Meta return by campaign, ad set and ad, beside spend, CPA and CTR (Link).
Open Meta Overview →ROAS
What it means
On the Summary page, ROAS is Total Sales divided by Total Ad Spend with Amazon included on both sides. Total Sales here is Shopify sales plus Amazon sales; Total Ad Spend is Meta plus Google plus Amazon. It's the widest version of the number the app produces — everything you sold over everything you spent to sell it.
This is the one that differs from the Marketing Summary tab, and the difference catches people out. Same store, same month, two cards, two answers — because only this one has Amazon in it.
Show the math
Formula and a worked example
The formula. ROAS = Total Sales ÷ Total Ad Spend, where Total Sales = Shopify Sales + Amazon Sales and Total Ad Spend = Meta + Google + Amazon.
Worked example. A store takes €200,000 on Shopify and €40,000 on Amazon, so Total Sales is €240,000. It spends €50,000 on Meta and Google and €20,000 on Amazon, so Total Ad Spend is €70,000. ROAS = 240,000 ÷ 70,000 = 3.43.
The Marketing Summary tab, for that same store and month, reads 200,000 ÷ 50,000 = 4.0. Neither is wrong. Amazon here is carrying €40,000 of sales on €20,000 of spend, which is a weaker ratio than the rest of the business, so adding it pulls the blended figure down.
That's the useful reading: the gap between the two cards tells you which way Amazon is pulling your marketing efficiency.
It answers the question
Across every channel you sell on and every channel you advertise on, what does a euro of marketing return? It's the board-level version — one number for the whole operation.
Why it matters
It's the only figure on the site with the entire business on both sides of the division. For a store where Amazon is a serious channel, every other return number on the site is a partial view, and this is the one that can't be flattered by leaving a channel out.
It's also the right number for a budget conversation. Whether to spend the next €10,000 on Meta, Google or Amazon is a question about the blend, and this is the blend.
What good looks like
The same band is used on the Marketing Summary tab, but on different inputs — so a store can sit above it on one card and below it on the other in the same month. Check which card you're reading before you act on the grade. Because the numerator includes revenue no advertising produced, treat this mainly as a trend against your own history, and read Contribution MarginContribution MarginProfit left after variable costs to fund the business.30% or more is healthy beside it to see whether the revenue was worth having.
How to improve it
| Lever | What you do | Expect | How long | Watch out for |
|---|---|---|---|---|
| Fast Rebalance across all three platforms | Move spend toward whichever of Meta, Google or Amazon shows the better return on its own tab | Blended figure up within a fortnight | 1–2 weeks | Each platform's own return is measured differently, so the comparison you're acting on isn't like-for-like. Move in small steps and check the blend after each. |
| Fast Cut the Amazon spend that isn't ranking | Pause Amazon campaigns with no attributed sales and no organic rank to defend | Denominator down with sales roughly flat | 1–2 weeks | Amazon ad spend often supports organic rank as well as direct sales, and that loss shows up a month or two later on units, not immediately here. |
| Slow Grow the channel that needs least advertising | Put effort into email, SMS and repeat purchase so more revenue arrives without spend behind it | Numerator grows with the denominator flat | 1 quarter | It improves the ratio without any campaign improving, which can mask genuinely weak advertising underneath. |
| Slow Fix the weaker channel rather than defunding it | Work on listings, pricing and stock on whichever channel is dragging the blend down | Blended figure rises without cutting reach | 1–2 quarters | Slow, and while you work the blend keeps reading low — the temptation is to cut the channel instead, which is the faster but usually wrong move. |
Every lever costs something somewhere. The last column is the one to read twice.
Read it with
The blend hides which channel moved it. The Amazon share is what tells you.
The Amazon share tells you which channel moved the blend. Amazon's share rising with the return rising means the channel is earning its place. Amazon's share rising while the return falls means you're growing on spend that returns less than the rest of your mix — worth it to build rank, costly as a permanent setting. Amazon's share falling while the return rises needs a check on whether Amazon shrank or Shopify outgrew it.
The reciprocal view of the same division, and the two cards are graded on scales that don't line up, so they can disagree. MER frames marketing as a cost you're carrying; this frames it as a return you're earning. Same arithmetic, different conversation.
A return can rise while profit falls — discounting to grow revenue does exactly that. This ratio can't see cost of goods, fees, shipping or refunds, and all four sit between it and the money you keep.
Common misreads
Both are correct. This card adds Amazon sales to the numerator and Amazon spend to the denominator; the Marketing Summary card leaves Amazon out of both. The gap between them is information about Amazon, not an error.
It also improves when repeat customers come back, when an email campaign lands, or when a channel with high sales and low spend grows. None of those is an advertising improvement. Check each platform's own tab before crediting the ads.
Most of the numerator is revenue that would have arrived anyway. The next euro of spend earns the marginal return, not the blended one, and the marginal return is always the lower of the two.
Also called
Return on ad spend · blended return · combined return
See yoursYour whole-business return beside total sales, total ad spend and the Amazon share.
Open Summary →