Have you ever compared a marketplace price on your phone with the screen of a colleague sitting at the next desk? If not, try it—the result will surprise you.
E-commerce practitioner Andrey (from the "Pro Business" channel) ran a telling experiment. He logged the live price of the exact same SKU across five signed-in shoppers in Moscow, all within a 2–3 kilometer radius of one another.
The numbers on their screens diverged sharply:
- Lowest price shown: 165 rubles;
- Highest price shown: 220 rubles;
- Price spread exceeded 25% for the exact same product at the exact same minute;
- Meanwhile, the seller's average settled buyout price in their financial report landed at 192 rubles.
Behind this price carousel sits Ozon's machine-learning pricing engine. The final retail tag blends the seller's base price, promo discount, the platform's dynamic co-investment, and the personal Ozon Card discount (the green price tag).
Ozon calculates co-investment individually for each buyer profile. New shoppers or accounts with abandoned carts get maximum platform-funded discounts to trigger that first transaction. Local inventory placement also matters: if a unit already sits at the warehouse closest to the buyer, last-mile costs drop and Ozon subsidizes the order more aggressively.
For sellers, personalized pricing creates a blind spot. You can't calculate exact unit margin on an individual sale just by glancing at a weekly blended average. For how to build analytics around hidden platform discounts, read our piece on dynamic pricing and marketplace co-investment. Platforms have gone all-in on hyper-personalization, and sellers have to bake that volatility into their unit economics.
