Marketplace or Your Own Store: Rozetka, eMAG, Amazon and the Margin Maths

The marketplace question is usually asked as a strategy question and is actually an arithmetic question. A marketplace charges a commission you can calculate to the cent. Your own store charges you the cost of acquiring every visitor, which you can also calculate, and which most owners have never written down. Until both numbers exist on the same page, the debate is about preferences.

The short answer for most small and medium sellers is that this is not a choice between two channels but a decision about sequence and proportion. What follows is how to do the arithmetic, what each side buys you beyond the numbers, and the point at which the balance shifts.

What each channel actually charges

Marketplace commissions vary by category more than by platform, and the differences are large. Electronics, where margins are thin and competition is intense, sits in single digits. Fashion, accessories and home goods run far higher — high-teens to around thirty percent on some categories, which is more than the gross margin of many sellers.

Transparency also varies. Amazon publishes its referral fees and fulfilment charges openly, so you can model a business before signing anything. eMAG publishes its structure and holds detailed category tables for registered sellers. Rozetka discloses category rates through the seller cabinet and a manager rather than publicly, alongside a small monthly platform fee charged only in months when the account is used.

The practical consequence: before you can compare anything, you need your exact rate for your exact categories, in writing. A quoted range is not an input to a decision.

The comparison that matters

Here is the same order seen from both sides. The numbers are illustrative; the structure is the point.

The row that decides most cases is traffic cost per order, and it is the one nobody has. It is not the cost per click; it is total marketing spend divided by orders won. A shop paying €18 to acquire an order is, on this example, worse off than a fifteen percent marketplace commission — until the second purchase, when the marketplace charges again and the shop pays only the cost of an email.

What the commission buys that is not traffic

Treating the commission as purely a fee for visitors understates what is included, and sellers who ignore this build a worse business case in both directions.

• Trust you have not earned yet. A first-time buyer will order from an unknown brand on a familiar platform long before they will order from an unknown site. This is the single largest advantage and it decays as your own brand becomes known.
• Checkout and payments that already work. No integration, no card processing contract, no cash-on-delivery reconciliation.
• Logistics, if you use it. Fulfilment services move the warehouse problem off your balance sheet at a price you can compute per unit.
• Demand data. Search volumes, competitor pricing and conversion by listing, which is genuinely useful even if you eventually leave.

What it does not buy is a relationship. The customer is the platform’s, the reviews are the platform’s, and the email address usually is too. That is the cost that does not appear in the commission line.

Where marketplaces get expensive

Three costs surface only after a few months of selling, and all three are structural rather than accidental.

Price competition on identical listings. Where several sellers offer the same product, the platform’s ranking rewards the lowest price. This is efficient for the buyer and compresses margin for everyone else. Sellers of unique or own-brand products avoid this entirely, which is the strongest argument for own-brand goods on marketplaces.

Advertising inside the platform. Organic visibility declines as sponsored placements expand, so the commission increasingly buys eligibility rather than traffic. Budget for it as a second, growing line, not as an optional experiment.

Rule changes you do not control. Category commissions change, return policies change, an account can be suspended over a metric. A business whose entire revenue sits on one platform has accepted a risk it cannot price. This is the same dependency problem as building on rented infrastructure, discussed in our note on who owns your website.

Where your own store gets expensive

The symmetrical honesty is that an own store is not free, and the costs are front-loaded at exactly the moment cash is tight.

Build and running costs are the visible part, and they are usually the smaller part — the realistic three-year picture is set out in our note on what an online store really costs. The larger part is demand: someone has to produce it, every month, through advertising, search, content or reputation. The channel decision behind that spend is examined in Google Ads or SEO.

Then there is the operational work a marketplace absorbs invisibly: payments, delivery integrations, returns handling, customer questions before purchase. None of it is difficult. All of it is time, and time is the resource small sellers actually run out of.

The hybrid that most sellers should run

The version that works in practice is not fifty-fifty. It is a deliberate division of labour.

1. Marketplaces for discovery and for products that survive price comparison. Standard items, entry-level products, seasonal overstock. Accept the commission as a customer-acquisition cost and stop resenting it.
2. Your own store for margin, range and relationship. Bundles, configurable products, own brand, anything with a story that a listing template cannot hold.
3. A deliberate bridge. Everything legitimately permitted inside a marketplace parcel — branded packaging, a printed card with a reason to visit, warranty registration, a manual, a spare part. Read the platform rules and stay inside them; the goal is recognition, not diversion.
4. A price policy you can defend. Do not undercut the marketplace listing on your own site if the platform’s terms prohibit it. Differentiate instead through bundles, wider range, service or loyalty terms, so the two channels are not selling an identical unit at different prices.

Once the second and third purchases start arriving through your own channel, the arithmetic changes decisively: the marketplace paid for the introduction and you keep the lifetime.

Three cases where the answer is not a hybrid

The balanced recommendation is right for most sellers and wrong for three recognisable ones. It is worth checking whether you are one of them before building a plan that assumes you are not.

The reseller of widely available goods with no exclusivity. If forty other sellers list the identical item, an own store has almost nothing to offer a buyer who can compare prices in one place. Here the marketplace is not a channel, it is the market, and the money is better spent on procurement and logistics than on a site nobody has a reason to visit. The strategic project is finding something to sell that is yours.

The seller with a real brand and repeat purchase. Consumables, cosmetics, coffee, pet food, anything bought monthly. Here the marketplace charges a commission on every one of those repeats forever, and the customer it introduces is a customer it keeps. Use it as a controlled acquisition experiment with a measured budget, and put the serious effort into the channel where the second and twentieth orders are nearly free.

The B2B or high-consideration seller. Long specification conversations, technical selection, quotes, contracts. Marketplace listing formats cannot hold this and the commission on a large order is disproportionate to a service the platform is not really providing. The right instrument is a site that answers technical questions well and a sales process behind it.

Everyone else — most retailers with a mix of standard and distinctive products — belongs in the hybrid above, with the proportion set by the arithmetic rather than by preference.

How to decide this quarter

Four steps, no consultants required.

Get your exact commission for your exact categories, in writing. Calculate your true cost per acquired order from twelve months of marketing spend. Compare them as percentages of average order value, then look at repeat rate — if a typical customer buys three times a year, the own-store case is far stronger than a single-order comparison suggests. Finally, ask which channel your competitors are winning in and why; if the top listings in your category are all own-brand and yours is a resold product, the marketplace is a price war you will enter without an advantage.

Whichever way it lands, keep your own store alive even if it is small. It is the only asset in this comparison that you own, and the only one where the second sale is nearly free.

Key takeaways

• This is arithmetic, not strategy. Compare marketplace commission against your real cost per acquired order, both as percentages of average order value.
• Get the exact rate in writing. Amazon publishes fees openly; eMAG and Rozetka disclose detailed category tables to registered sellers. Ranges are not inputs to a decision.
• Commission buys borrowed trust, a working checkout, optional logistics and demand data — but not a customer relationship.
• Marketplace costs grow structurally: price competition on identical listings, mandatory-in-practice advertising, and rules that can change without you.
• Own stores cost demand generation, which is monthly and unavoidable, plus the operational work a platform absorbs invisibly.
• Run a deliberate hybrid: marketplace for discovery and comparable goods, own store for margin and range, a compliant bridge in the parcel, and a price policy that does not put the two channels in conflict.