The essentials
- Definition: ROAS (Return On Ad Spend) is the revenue attributed to the ads divided by their cost; a ROAS of 4 means €4 of sales for every €1 spent.
- The threshold: break-even ROAS equals 1 divided by the gross margin; with a 30% margin, you need to go above 3.3 to avoid losing money on the margin.
- 2026 benchmarks: Meta e-commerce ROAS of 2 to 5 in acquisition, 5 to 15 in retargeting; TikTok 1.5 to 4; the platform figures are overstated by view-through attribution.
- Levers: reliable measurement, creative, segments, exclusions, conversion values, returns deducted.
The aim of ROAS is to measure the profitability of social advertising for e-commerce and for any business that measures a sales value. It is simple to calculate and easy to misread: a ROAS of 3 can be excellent or catastrophic depending on the margin, and the ROAS shown by Meta is not the one in your income statement. This article defines ROAS, shows how to set the break-even threshold, gives 2026 benchmarks, and explains the attribution traps and the levers. The general definition is in ROAS: definition ; ROAS on the Google side is covered in SEA and e-commerce.
Definition and formula
ROAS = revenue attributed to the ads / ad spend. It is expressed as a multiple (4) or as a percentage (400%). It can be calculated by campaign, ad set, ad, audience, creative or period. It takes into account neither the margin, nor management costs, nor returns, nor the sales the advertising did not cause: it is a metric of advertising efficiency, not of profit. Overall return on investment is covered in how to calculate marketing ROI.
Setting the break-even ROAS
| Gross margin | Break-even ROAS (1 / margin) | Recommended target ROAS (margin + management costs + returns) |
|---|---|---|
| 20 % | 5,0 | 6 to 7 |
| 30 % | 3,3 | 4 to 5 |
| 40 % | 2,5 | 3 to 3.5 |
| 50 % | 2,0 | 2.5 to 3 |
| 70% (services, software) | 1,4 | 1.8 to 2.2 |
At the break-even ROAS, advertising consumes the whole margin of the sale. The target ROAS adds management costs (10 to 20% of the budget), returns and cancellations, and the margin you want to keep. A business can aim for a lower ROAS in acquisition if customer lifetime value (repeat purchases) justifies it; see LTV in digital marketing.
2026 benchmarks in social advertising
| Context | Typical ROAS (platform attribution) | Note |
|---|---|---|
| Meta, e-commerce, acquisition (cold audiences, lookalike audiences) | 2 to 5 | Depends on the basket and the margin; fashion and beauty at the bottom, technical products at the top |
| Meta, retargeting and dynamic catalogue | 5 to 15 | Overstated by view-through; closes sales prepared elsewhere |
| Meta, Advantage+ Shopping | 3 to 8 | Mixes acquisition and existing customers; read the share of new customers |
| TikTok, e-commerce | 1.5 to 4 | Young audience, smaller baskets, strong indirect awareness |
| 2 to 6 | High purchase intent in decoration, fashion and home | |
| Rarely measured as ROAS | Cost per lead and per opportunity more relevant |
The attribution traps
- View-through: Meta attributes by default the purchases made within the day following an impression, even without a click. Some of those purchases would have happened without the advertising. Switch the window to “7-day click” only for a more conservative ROAS, or compare the two.
- Existing customers: a campaign that retargets customers shows a high ROAS on sales that would have happened anyway; exclude them or measure the share of new customers.
- Modelling: Meta and TikTok model the conversions lost to consent; useful, but unverifiable.
- GA4 ROAS: click-based and with data-driven attribution, it is often 30 to 50% lower than the platform's; it is used to arbitrate between channels; see data-driven attribution.
- Returns: a ROAS calculated on gross sales overstates profitability by 10 to 30% in fashion; deduct the returns or use a net value.
- Incrementality: only a test with a control group (Meta Conversion Lift, geographic experiment) measures the sales actually caused by the advertising; available to budgets of several thousand euros per month.
The levers for improving ROAS
- Measurement: pixel and conversions API, exact conversion values, deduplication; a ROAS calculated on incomplete data leads you to cut profitable campaigns.
- Creative: the first lever; 3 to 6 variants, refreshed every 3 to 6 weeks, with the product visible, proof and a clear offer.
- Segmentation by margin: campaigns or product groups separated by margin level, each with its own target ROAS.
- The exclusions: existing customers in acquisition, low-margin or out-of-stock products.
- Targeted retargeting: abandoned baskets and viewed products, with a capped frequency; see retargeting on social media.
- The page and the offer: speed, trust, delivery costs, average basket (bundles, free delivery threshold): a ROAS is also won after the click.
- Seasonality: budgets increased before the peaks, reduced after; target ROAS adjusted to the period.
Reading ROAS over time
Daily ROAS is noisy; read it over a rolling 7 days and over the month, by segment (acquisition, retargeting, customers), alongside volume: a ROAS that rises because the budget is falling is not progress. ROAS paired with the volume of new customers is the right dashboard; it is completed by cost per acquisition and LTV for budget decisions. The full metrics are in the SMA KPIs, and the improvement methods in improving the ROAS of your social ads campaigns.
How GreenRed helps
Rather than juggling several tools, GreenRed's return on investment module brings these metrics together in a single dashboard, compares them over time and tells you which actions come first. You can try it free, with no card, from the Pricing.
Frequently asked questions
What is a good ROAS in social advertising?
The one that exceeds your break-even ROAS (1 divided by the gross margin) with a safety margin: 4 to 5 with a 30% margin, 2.5 to 3 with 50%. In 2026, Meta e-commerce campaigns show 2 to 5 in acquisition and 5 to 15 in retargeting according to platform attribution, to be read with caution.
Why is Meta's ROAS higher than GA4's?
Meta attributes purchases within 7 days of a click and 1 day after a simple view, and models the conversions lost; GA4 attributes on clicks only and splits between channels. A gap of 30 to 50% is common. Use Meta to optimise within Meta, and GA4 or the CRM to arbitrate between channels.
Should returns be deducted from ROAS?
Yes: a ROAS calculated on gross sales overstates profitability, by 10 to 30% in fashion. Import the conversion adjustments or use an estimated net value, otherwise bidding optimises on cancelled sales and your budget decisions are wrong.
Does a high ROAS mean the campaign is profitable?
Not necessarily: a ROAS of 6 on a campaign that only retargets existing customers measures sales that would have happened without advertising. Look at the share of new customers, compare with the click-only window, and for large budgets measure incrementality with a control group.