The Maze: Scale is usually the first number ecommerce teams worship. ECDB's 2026 retailer benchmarks make that look lazy. Zalando has EUR 14.4bn in online GMV, more than four times Zooplus. But Zooplus converts at 4.15%, while Zalando converts at 2.90%. That is not a rounding error. It is the category model showing up in the KPI.
The biggest retailer is not the most efficient retailer. In the visible ECDB sample, Zalando leads the group with EUR 14.4bn in 2025 online GMV, followed by Otto at EUR 7.9bn and MediaMarkt at EUR 7.1bn. All three sit below 3% conversion: Zalando at 2.90%, Otto at 2.86%, and MediaMarkt at 2.86%. The report page frames the exercise around conversion-rate benchmarks and KPI interpretation, which is the real point. Scale tells you where demand aggregates. Conversion tells you how naturally that demand turns into an order.
Specialists win because the shopper mission is cleaner. Zooplus does EUR 3.4bn in GMV, but its 4.15% conversion is the strongest in the view. Flaconi is tiny by comparison at EUR 0.7bn, yet converts at 3.60%. Sephora sits in the middle at EUR 5.0bn and 3.52%. The post's logic is straightforward: pet supplies have refill behavior; beauty has high purchase intent; fashion has more browsing and hesitation. That means a category can have a lower ceiling for GMV and a higher ceiling for conversion at the same time.
The blended benchmark is the management trap. A 3.4% ecommerce conversion average sounds useful until it is applied to all models. It is not. A fashion marketplace, a pet replenishment retailer, a beauty specialist, and an electronics player ask shoppers to do different jobs. MediaMarkt's EUR 178 AOV, cited in the post, helps explain why electronics can carry intent but still convert at only 2.86%. Big-ticket categories create comparison, consideration, and delay. Smaller baskets can create habit.
ECDB's methodology makes the comparison harder to dismiss. The public ECDB page says the report covers conversion benchmarks, AOV, cart abandonment, return rate, and checkout behavior across categories. It also says ECDB's figures are built from real purchase data, payment-processor and issuer signals, retailer data, traffic signals, public market data, and analyst modelling. That does not make every benchmark universal. It makes the category caveat more important.
The operating question changes from 'how big?' to 'what mission?' Ferenc Bartha's visible comment put the sharper strategic frame on it: manufacturers and brands should ask which category model they are entering, not just which retailer is large. That is the better question. If the mission is refill, win availability and retention. If the mission is discovery, win assortment and inspiration. If the mission is big-ticket confidence, win information, trust, and checkout certainty.
Why it matters: Conversion targets are often copied across businesses because benchmarks feel objective. This evidence says the opposite. A bad benchmark can make a healthy category model look broken, or make a weak one look fine. Operators should compare Zalando with fashion logic, Zooplus with replenishment logic, Sephora with beauty logic, and MediaMarkt with electronics logic. Scale is the scoreboard. Category economics are the rules of the game.
