ROAS in e-com Russia 2026
Median ROAS in e-com clothing is 4.2x. In two years it fell from 5.1x - competition in Yandex Direct plus an increase in the cost of a click in VK Ads. This is my working niche range, a reference point for comparison.
what is included in this figure
ROAS = advertising campaign revenue ÷ advertising costs. Metrica without taking into account the cost of goods, operating expenses, returns and logistics costs. Essentially - how many kopecks of revenue comes for every media ruble spent. More details in my ROAS glossaries.
What not included in this figure is 4.2x:
- Margin. ROAS 4.2x with a margin of 30% is advertising return 1.26x per unit of margin. After all the others expenses (salaries, warehouse, acquiring, advertising team) - business can go to zero or minus. Count break-even helps break-even ROAS calculator.
- Returns. In clothing, returns reach up to 30%. If ROAS includes “check revenue” and not “net revenue” after returns” - the figure is overestimated. Correctly calculate ROAS by net revenue.
- LTV. This figure is about one-time revenue. In models with repeat purchases (e-com subscriptions, fashion clubs) ROAS over a 12-month horizon, 2–3 times higher than the first purchase.
When an agency brings “ROAS 8x”, the first thing I ask is: by revenue - gross or net, one-time or with repeat purchases, with or without returns. These four filters change the number 1.5–3 times in any direction.
ROAS table by niche
| Niche | Top 25% (higher) | Median ROAS | Weak (lower) |
|---|---|---|---|
E-com clothing/accessories margin ~35-45% | 6.0x | 3.8x | 2.2x |
FMCG / Retail (low margin) margin ~15-25% | 8.5x | 5.0x | 2.8x |
B2B SaaS (LTV model) We calculate according to LTV, not AOV | 5.0x | 3.0x | 1.5x |
B2B services (high check) margin ~50-70% | 7.5x | 4.2x | 2.0x |
Local services margin ~40-60% | 9.0x | 5.5x | 2.5x |
Online education (courses) long cycle, installments | 5.0x | 2.8x | 1.4x |
Medicine (clinics) LTV is critical: repeat appointments | 6.5x | 3.5x | 1.8x |
Fintech (insurance, brokers) CPA model is more often than ROAS | 4.5x | 2.5x | 1.3x |
The highest ROAS is for local services (median 5.0x) and FMCG (5.5x). For local services, because a margin of 40–60% allows pay per lead, in FMCG - because the volume pulls. The most low ROAS - B2B SaaS (2.8x), but this is not “bad”, it’s different model: there ROAS is calculated based on LTV for 12–24 months, not AOV first subscription.
E-com clothing in the middle - 4.2x median. Explained by margin 35–45% and an average basket size of 4–7 thousand rubles. Supports from below fintech (2.4x) - the CPA model is more common there, ROAS lags behind as metric.
how it changes over time
Trend in e-com clothing:
- 2024: 5.1x median. After the Instagram target is closed advertisers have shifted to VK Ads and Yandex Direct, the auction is still didn't warm up.
- 2025: 4.7x. VK Ads rates have increased, especially in Stories. Yandex Direct began to work more closely with product advertisements.
- 2026: 4.2x. Full growth of competition in both offices. Cost per click +25% year on year. ROAS falls about the same.
The trend is −18% over two years. This is a structural increase in the cost of traffic, not “you’re in a bad mood.” Who hasn't reassembled during this time offer and unit economics - caught in scissors: ROAS is falling, Margins are not growing.
what does your result mean in the context of a benchmark?
Three scenarios.
ROAS is higher than good (more than 6.5x in e-com clothing). Check two things. Is the first one a first purchase or repeat purchases? If with repetitions, you consider LTV-ROAS, not classic performance-ROAS. The second is not a “brand” campaign on searching? Brand traffic is people who already know you, they would have come without advertising. A purely cold audience is rare gives ROAS above 5x in clothing.
ROAS is around the median (3.5–5x). You're in the market. Next, you check whether it fits into the margin. Make it convenient in break-even ROAS calculator: you plug in the margin%, it shows the break-even point. If break-even ROAS = 3.6x with a margin of 28%, and yours is 4.2x 0.6x margin on marketing investment. It's working, but narrow economics.
ROAS is below bad (less than 2.5x). The channel is unprofitable in the vast majority of cases. Three reasons usually. First - cold audience that does not know the brand (it has low ROAS always). The second is broad interests instead of LAL, the cabinet spends budget for irrelevant ones. Third - landing page with a conversion of 0.8% (rather than 2–3%, like the median). Landing treats the bid, but the bid does not treats the landing page.
sources and methodology
Figures compiled from:
- 6 of my projects 2024–2026 - basic data on e-com clothing, MedTech and B2B SaaS.
- Correspondence with 12 colleagues in performance by niches where I have no personal experience - FMCG, fintech, auto.
- Open reports Yandex Advertising Q1 2026 by e-com and Calltouch by service niches.
Method: for each niche 3–5 sources, median and quartiles, anomalies (ROAS above 15x or below 0.5x) are rejected.
ROAS is not financial advice. "Good" ROAS determined by the project margin, not the market median. Break-even ROAS with a margin of 25% = 4x, with a margin of 60% = 1.67x — median 4.2x for the first case is critical, for the second redundant.
related metrics and tools
- Landing page conversion 2026 — median 2.5%. ROAS directly depends on landing CR.
- CPC in Yandex Direct 2026 — median 32 ₽ in e-com on Search. ROAS = AOV × CR ÷ CPC.
- LTV/CAC in SaaS 2026 — for subscription models, ROAS is calculated by LTV.
Calculators: ROAS calculator, break-even ROAS calculator, unit economics calculator.
On the blog: marketing budget 2026, channel attribution 2026, unit economy 2026. From glossary - ROAS, AOV and CAC.