Selling on your own site versus a marketplace
Compare the total take rate rather than the commission, then weigh instant audience against owning the customer. A decision that usually ends in both.
· 5 min read
Compare take rate, not commission
The comparison is almost always made on the commission percentage, which is one deduction among several and not reliably the largest.
A marketplace sale can carry: a referral or commission fee that varies by product category, a fixed closing fee per order, weight-based shipping charges if you use the platform's logistics, storage fees if you use its warehousing, handling costs on returns, and advertising spend that has become effectively necessary to be visible in a crowded category. The sum of those against your selling price is the take rate, and that is the figure to compare.
This is also why quoting a general commission range for a marketplace is unhelpful to the point of being misleading. Fees differ substantially by category — the schedule for apparel is not the schedule for electronics or for grocery — and they are revised. Every major marketplace publishes a category-wise fee schedule, and the only figure worth planning against is the one you calculate for your own products from the current schedule, including the advertising you will realistically need.
Your own site has a different cost shape
On your own site the costs are a fixed platform or hosting fee, a payment gateway percentage, and the cost of getting people there. The first two are small and easy to quantify. The third is the one that is routinely left out of the comparison, and it is the item that decides it.
A marketplace commission is partly a payment for demand. Buyers arrive already searching for what you sell, and the fee includes access to them. On your own site that demand does not exist by default: it is content you produce, advertising you buy, an audience you have built, or customers who already know you. None of it is free, and the fact that it is not itemised on an invoice is what makes the comparison look lopsided in your site's favour.
So the honest version sets the marketplace take rate against your own site's fees plus a realistic customer acquisition cost. For a business with no existing audience and no marketing budget, a low-fee channel that delivers no visitors has a cost per sale that is undefined rather than low.
Audience now versus audience you own
The trade at the centre of this decision is speed against ownership.
A marketplace gives you buyers immediately. Listings can be live in a day, and within a fortnight you have real information about whether anybody wants your product at your price. As a way of validating demand before committing to inventory or a brand, that is genuinely valuable and difficult to replicate.
Your own site accumulates slowly and belongs to you. Search visibility, an email list, returning customers and a recognisable name compound over years, and the asymmetry with the marketplace is what matters: platform traffic can stop without warning if a listing is suppressed, a category's ranking logic changes, a competitor undercuts you, or an account is suspended over a policy dispute. Your own audience does not disappear on somebody else's decision.
The strategic risk of marketplace-only selling is not the fee, which is at least predictable. It is concentration — a business whose entire revenue depends on one platform's continued goodwill has a single point of failure it does not control and cannot negotiate with.
The customer data question
On your own site you hold the customer relationship: the email address, the phone number with consent to use it, the order history, and the ability to contact somebody again with a relevant offer.
On a marketplace you generally do not. The platform mediates the relationship, and depending on the arrangement you may receive little more than a shipping address and no permission to market to it. The repeat purchase belongs to the platform, and the buyer's loyalty attaches to the marketplace rather than to you — many customers genuinely do not recall which seller fulfilled their order.
For products with real repeat purchase this is the single most important factor in the whole comparison, and it is where the arithmetic diverges most sharply. If the first order is roughly break-even after fees and the margin lives in the second and third, then a channel that does not let you reach the customer again is not competing on take rate at all. It is competing on the value of one transaction against the value of a relationship, and losing badly.
For genuinely one-off purchases, this consideration nearly vanishes.
Payment timing, price competition and brand
Three further differences are worth pricing.
Settlement timing. Marketplaces pay on their own cycle after deducting fees, and may hold amounts against expected returns. On your own site the gateway's cycle applies and you can compare providers on it. For a business funding inventory from sales, a longer or less predictable cycle is a genuine working-capital cost even though it appears nowhere as a fee.
Price competition. A marketplace listing sits beside near-identical alternatives in a list influenced by price, which compresses margins on anything commoditised and rewards whoever can accept the least. On your own site you are not adjacent to a cheaper version of yourself, and the comparison the customer makes is against the effort of looking elsewhere.
Brand and presentation. You control the entire experience on your own site — the photography standard, the copy, the packaging story, the follow-up. On a marketplace you control a listing inside somebody else's template, and the customer's experience of buying is the platform's experience rather than yours. For a business whose differentiation is how it does things rather than what it stocks, that constraint is substantial.
A decision rule, and why most end up doing both
Two clear cases and a common middle.
If you sell largely undifferentiated products where discovery is the whole battle and repeat purchase is limited, marketplace economics usually win. You are buying access to demand you could not generate, and the fee is the price of that access.
If your products are differentiated, your brand is part of the value, and customers buy repeatedly, your own site is where the value accrues. Paying a take rate on every order forever to a channel that keeps the customer relationship is the more expensive arrangement over any reasonable horizon.
Most businesses end up running both, with different jobs. The marketplace handles discovery, reaches buyers who will never visit your site, and clears surplus stock. The site serves customers who already know you, carries the full range, protects margin, and holds the relationship. Packaging inserts inviting a marketplace buyer to order directly next time are the standard bridge between the two.
The mistake is running both with identical pricing and identical strategy, then concluding from the resulting numbers that the marketplace is simply better — when what has been measured is which channel had an audience.
Common questions
Should a new business start on a marketplace or build a site first?
Starting on a marketplace answers the more urgent question — whether anyone wants the product at your price — with less capital and within weeks. Building a site first means spending on infrastructure and traffic before demand is proven. The usual sequence is marketplace to validate, then your own site once you know what sells and to whom.
Can I sell the same products at different prices on each channel?
Different pricing is common, and check the platform's terms, because some marketplace programmes impose parity or price-competitiveness conditions that affect your visibility rather than being outright prohibitions. Where you have freedom, the argument for pricing lower on your own site is that you keep more of it, and the argument against is that the marketplace listing then becomes the reference price customers remember.
How do I move marketplace customers to my own site?
Within platform rules, which generally restrict marketing to buyers directly, the workable route is what goes in the box: a packaging insert, a card with a reason to order directly, or product registration and warranty pages. Read the specific programme terms first, since some prohibit inserts that divert customers, and a suspension costs more than the transfers are worth.
Is it worth being on more than one marketplace?
It reduces the concentration risk of depending on one platform and multiplies the operational work: separate listings, separate fee schedules, separate return handling and the perennial problem of keeping inventory synchronised. It is usually worth it once your volume justifies tooling to manage listings and stock centrally, and premature before that.
Related pages