Selling on multiple marketplaces without cannibalising yourself

Being present on several marketplaces looks like an obvious growth move: more storefronts, more sales. Amazon first, then Cdiscount, ManoMano, Fnac, eBay: each channel opens an audience your own site cannot reach alone. Then, from the second or third marketplace onward, a worry sets in: aren't these channels stealing sales from one another, and isn't the marketplace cannibalising my own shop?

The real question is not how many marketplaces, but how to stop them from fighting each other: over price, over the same product listing, over the same Google query. Cannibalisation is real, but it rarely sits where it is feared, and the levers that contain it almost never involve "selling less".

Adding marketplaces does not mechanically add sales. Without arbitration, it mostly adds opportunities to compete against yourself: on price, on the Buy Box, on Google.

Cannibalisation: what it is, and what it isn't

The classic fear is that the marketplace eats the own site's sales. The data often points the other way: a marketplace mainly brings new customers, a share of whom would never have found the shop directly, with a modest but real halo effect on the site. Marketplace operator Mirakl shows, figures in hand, that the acquisition effect statistically outweighs cannibalisation when the assortment is built to fill gaps rather than duplicate what already exists.

Real cannibalisation therefore does not play out in volume between "site" and "marketplace". It plays out along three fault lines: margin (paying a commission for a customer who would have come direct), cross-channel price (undercutting yourself from one platform to the next), and the shared listing (ending up bidding against your own resellers on a single Amazon reference). The question shifts from "will I lose sales?" to "at what margin, and who captures the value?".

Price is the real battlefield, not the catalogue

Every marketplace watches the prices charged elsewhere, including on your own site. Amazon formalises this in its Marketplace Fair Pricing Policy: a materially lower price seen on another channel can suppress your Featured Offer (the Buy Box), or even hide the offer. Price inconsistency across channels is therefore not just a branding matter: it carries a direct algorithmic penalty on your visibility.

The consequence is a pricing policy thought through across channels, with an assumed price floor (a MAP logic, minimum advertised price), rather than a rate decided platform by platform in isolation. The reflex "let me drop a little on Cdiscount to grab volume" can break the Buy Box on Amazon the next day. Conversely, aligning everything to the lowest price everywhere is a race to the bottom: you shave margin on every channel for a volume gain on one.

The Buy Box, and the trap of competing on your own listing

On Amazon, several sellers share a single product listing and compete for the Featured Offer. If third parties resell your products, or if you sell both direct and through distributors, you can end up bidding against yourself on your own reference. Landed price (item plus shipping), fulfilment method (FBA / Prime), stock depth and seller metrics all weigh on Buy Box allocation: price is not the only lever, availability and service quality count just as much.

The lever is therefore not tariff-based but structural: control your distribution, register your brands (Brand Registry) and enforce a price floor so you do not fuel a Buy Box war over your own products. Here the logic joins that of the organic ranking of an Amazon listing (A9/A10) and the listing conversion rate: winning the Featured Offer serves little purpose if the listing itself does not convert.

Differentiate by channel rather than duplicate identically

The best antidote to cannibalisation is not selling less, it is not presenting the exact same comparable offer everywhere. Fill assortment gaps rather than duplicate SKUs: marketplace-specific packs and bundles, different lot sizes, exclusive colourways, distinct warranties or bundled services. An offer with no strict equivalent elsewhere makes mechanical price comparison impossible and gives each channel a reason to exist.

Each marketplace also has its own audience: ManoMano and Leroy Merlin for DIY, Fnac-Darty for culture and tech, Cdiscount for generalist price, Amazon for maximum reach. The map of French marketplaces shows how far these audiences differ. Adapting the assortment to a channel's audience beats pouring the same feed everywhere.

Feed management: the backbone beyond two or three channels

The real tipping point is operational. As long as you run two marketplaces, a spreadsheet is enough. By the fourth, manual management costs more than it returns: stock errors, oversells, penalties. The foundation becomes a single master product repository, mapped to each marketplace (titles, categories, attributes, price, stock) through rules rather than re-keying. That same repository is what makes per-channel differentiation sustainable: each platform receives its variant without double entry.

Near real-time stock synchronisation is not a comfort: overselling is penalised by every platform, and those penalties degrade the seller metrics that, in turn, decide the Buy Box. One physical stock shared across channels must be arbitrated, not promised twice. Without this repository, opening one more marketplace multiplies breaking points, not sales.

Marketplace and own shop: who captures the customer

On Google, the topic returns in another form. Free product listings in Merchant Center surface offers in the Shopping tab, and a marketplace listing can rank above your own site on the same product. You then risk paying a commission for a customer Google would have sent you direct, for free.

The doctrine fits in one sentence: the own site stays the channel of margin, customer data and brand; the marketplace is a channel of acquisition and discovery. You accept the commission as an acquisition cost, not as a substitute for direct. Openly diverting a buyer away from a marketplace is against their rules, but after-sales, brand content, SEO and retargeting remain legitimate levers. Well-maintained Product structured data helps the own site's listing exist in Shopping, and in generative answers, alongside the marketplaces.

Measure by channel, never in aggregate

The unit of measurement is the channel, and the metric is net margin (commission, logistics and advertising included), not aggregate revenue. A channel with high revenue but negative margin once the commission is paid is a channel that impoverishes you, even as it grows the total. For each product you also compare the marketplace acquisition cost with the direct one: it is that ratio, not the fear of cannibalisation, that should decide whether to open, or close, a channel. A dashboard that isolates each marketplace, in the spirit of good analytics reporting, beats a sum that hides the loss-making channels.

To go further on steering a marketplace's advertising, the Sponsored Products strategy on Amazon details the bidding side of this reasoning, and our marketplace services cover the whole chain, from the product repository to per-channel measurement.

At the crucible: transmute reach into margin, not into a price war. Adding marketplaces is only progress if each channel has a distinct role and a margin that holds once the commission is paid. The rest (cross-channel price consistency, assortment differentiation, a master repository, measurement at net margin) is merely the metallurgy that keeps the alloy from dissolving at the first heat on prices.