What ROAS is profitable? Calculating break-even ROAS for marketplace sellers
There is no universal good ROAS. How to compute break-even ROAS from contribution margin after marketplace fees, why platform ROAS runs high, and a three-channel example.
By The DataxLand team· · 4 min read
"A ROAS of 4 is good" is the most common answer to "what is a good ROAS?". The problem: for one shop a ROAS of 4 is very profitable; for another, the same ROAS loses money on every order. The right threshold is not in your ad account — it is in the contribution margin of what you sell, after marketplace fees and returns. This post shows how to compute that threshold and why the ROAS in your ad dashboard is usually higher than reality.
Numbers come from the DataxLand sample data (a simulated scented-candle shop, 1 Aug – 29 Oct 2026), not a customer's data.
What ROAS divides by what
ROAS = Revenue ÷ Ad spendSimple — but "revenue" can mean three very different things:
| ROAS type | Numerator | Source | Usually biased |
|---|---|---|---|
| Platform ROAS | Conversion value attributed by the ad platform | Shopee Ads, TikTok Ads, Meta | High: counts orders later cancelled or returned; several platforms claim the same order |
| ROAS on GMV | Gross revenue of the channel | Marketplace dashboard | High: before discounts, vouchers, refunds |
| True (blended) ROAS | Net revenue of the channel | Orders + settlements | Closest to reality |
DataxLand keeps them separate: roas_platform (platform conversion value ÷ spend), roas_blended (net revenue ÷ spend), and a
reported-vs-actual revenue gap metric.
Break-even ROAS is one division
Ads break even when each unit of ad spend brings back exactly one unit of contribution margin — what is left after cost of goods and all order-level fees:
Contribution margin % = (Net revenue − COGS − platform fees − payment fees − shipping − return shipping) ÷ Net revenue
Break-even ROAS = 1 ÷ Contribution margin %Comparing actual ROAS with break-even ROAS is also how large ad platforms teach it — see Amazon Ads, "Ads math": a campaign at or above break-even ROAS does not lose money. The twist for South-East Asian marketplaces is that margin must be computed after marketplace fees, which in Vietnam rose several times in 2025–2026.
| Contribution margin | Break-even ROAS | ROAS for a 10% net margin |
|---|---|---|
| 50% | 2.0 | 2.5 |
| 40% | 2.5 | 3.33 |
| 30% | 3.33 | 5.0 |
| 25% | 4.0 | 6.67 |
| 20% | 5.0 | 10.0 |
The last column is 1 ÷ (contribution margin − target net margin). At a 25% contribution margin, "ROAS 4 is good" means you only
break even.
Example: three channels, three thresholds
| Channel | Net revenue | Contribution margin % | Break-even ROAS | Ad spend | True ROAS | Result |
|---|---|---|---|---|---|---|
| Shopee | 397,961,000 ₫ | 32.2% | 3.11 | 38,163,000 ₫ | 10.43 | Profit 89,927,600 ₫ |
| TikTok Shop | 228,021,000 ₫ | 18.5% | 5.42 | 55,896,000 ₫ | 4.08 | Loss 13,802,100 ₫ |
| Website | 129,572,000 ₫ | 41.8% | 2.39 | 41,011,000 ₫ | 3.16 | Profit 13,175,700 ₫ |
TikTok Shop's true ROAS (4.08) sounds better than the website's (3.16). But the website breaks even at 2.39 while TikTok Shop needs 5.42, because its contribution margin is only 18.5% — largely due to a 12.4% return rate. The same ROAS is profitable in one channel and a loss in another. Comparing ROAS across channels without comparing break-even points is comparing the wrong thing.
When platform ROAS and cash disagree
In the sample data, ads for the website report conversion value 92,828,000 ₫ above the website's real net revenue. Recomputed, platform ROAS is about 5.42 while true ROAS is 3.16. Reading only the ad dashboard, you would call the campaign excellent and raise the budget.
Typical reasons:
- Wide attribution windows — the platform claims orders for days after a click or view, even if the buyer came from elsewhere.
- Several platforms claim one order — Meta, Google and TikTok can all attribute the same website order.
- Cancellations and returns still count — conversion value is captured at order time and not updated after a return.
- Pre-discount values — some setups send list price, not what the buyer paid.
Nobody is wrong; each platform measures its own way. Budget decisions still belong on net revenue.
Falling ROAS: read it early
The LIVE campaign TTA-201 on TikTok in the sample data has a platform ROAS of 3.04 in September and 2.14 in October. Both
are below the channel's break-even of 5.42 (and platform ROAS overstates), and the trend is down. That is a signal to stop or
restructure, not to add budget. An alert such as "weekly campaign ROAS down more than 20% week over week" surfaces it after a week
instead of a month.
Checklist before raising ad budget
- Am I reading platform ROAS or ROAS on net revenue?
- What is the contribution margin of this channel or product, after current fees and returns?
- Break-even ROAS = 1 ÷ contribution margin. How far above it is true ROAS?
- How big is the gap between reported and actual revenue?
- Is ROAS trending up or down over the last four weeks?
- Does the advertised product have a positive contribution margin? (Advertising a losing SKU is paying to lose faster.)
Measuring true ROAS automatically
True ROAS requires joining ad spend with net revenue by day and channel. DataxLand does that join when you connect order sources and ad sources, and shows platform ROAS, true ROAS and the reported-vs-actual gap side by side; you can ask Si "which campaigns are below break-even this month?" and get an answer with metric sources. See Si, the DataxLand docs, or the Shopee + Facebook Ads combination page.
Read next: How to calculate true profit on Shopee, TikTok Shop and Lazada