Own Store vs. Marketplace: The Contribution-Margin Calculation

The question sounds like a matter of faith, but it’s maths: where does the same order earn more – in your own store or on the marketplace? The answer differs from what both camps claim, because the two channels have fundamentally different cost models: on the marketplace you pay per transaction. In your store you pay per customer. Understanding that has almost solved the channel question already.

Key takeaways

• Marketplace and store have different cost logic: take rate per transaction versus customer acquisition cost (CAC) per customer.
• On the first purchase, the marketplace often wins the calculation because a new store customer’s CAC regularly exceeds the marketplace commission.
• The store wins through the second purchase: returning customers cost almost nothing there.
• The marketplace sells transactions and keeps the customers – no contact, no retargeting, no relationship.
• The honest steering metric across both channels is contribution margin per order after all channel costs.

The marketplace side of the calculation

The marketplace takes its take rate: commission, usually fulfilment fees, increasingly advertising costs – altogether, depending on category, a substantial but predictable share of revenue.

In return it delivers what’s expensive in a store: purchase-ready demand, trust and conversion. The strategic price doesn’t appear on the fee invoice: the customer belongs to the marketplace. No contact for the next launch, no retargeting, no relationship – every order is the last one, if the algorithm decides so. The marketplace sells you transactions and keeps the customers. Anyone selling on several marketplaces at once will find a structure for weighing these channels against each other, instead of optimising them one by one, in our article on marketplace strategy as a portfolio.

Important for the calculation: the take rate is rarely the whole truth. Anyone who also books sponsored placements on the marketplace just to be visible at all adds another, often underestimated cost block that doesn’t even show up in many internal calculations, because it gets booked as “marketing” rather than as sales cost. Only commission, fulfilment and advertising costs together give you the actual marketplace margin of an order.

The store side of the calculation

In your own store, commission disappears – replaced by customer acquisition cost (CAC), store operations, payment and service. On the first purchase, the store often loses this calculation: a new customer’s CAC regularly exceeds the marketplace commission on the same order.

Store economics turn around the second purchase: returning customers come via email, brand and habit – at marginal cost close to zero. That’s why the decisive store metric isn’t campaign ROAS, but the ratio of customer lifetime value to acquisition cost (LTV:CAC). A store without a repeat-purchase mechanism isn’t a channel, it’s an expensive hobby; a store with working retention – carried by automations like the ones described in the article Email & CRM – is a wealth machine that the marketplace can structurally never be.

This calculation needs patience: LTV:CAC can only be reliably assessed once enough order cycles have run to know the actual repeat-purchase rate – for a product with an annual purchase rhythm, that takes correspondingly longer than for one with monthly consumption. Anyone who judges their store by first-purchase ROAS after just three months is systematically rating it too harshly.

Worked example: the same order, two channels

An order worth €60 shows the difference concretely. On the marketplace, a take rate of, say, fifteen percent plus a fulfilment fee deducts nine to twelve euros directly – predictable, immediate, with no further effort. In your own store, the same new customer might cost twenty euros in CAC via ads, noticeably more than the marketplace fee.

The difference shows up at the second purchase: if that same customer comes back via an email automation for another €60, that second sale in the store now costs only a few euros in shipping – while the marketplace demands its full take rate again. After two or three orders, the store has usually overtaken the initially worse calculation. The exact percentages vary by category, the principle doesn’t.

That’s exactly why the metric “cost of the first order” is of limited use on either channel when looked at in isolation. Only looking across several order cycles shows which channel actually delivers more contribution margin for which product – and that view looks structurally different for products with high repeat-purchase rates than for one-off items.

The most common mistakes in channel comparison

Most bad decisions don’t come from wrong numbers, but from incomplete calculations.

• Counting only commission against zero – fix: fully price in CAC, payment, service and operations on the store side.
• Judging the first purchase in isolation – fix: calculate LTV across several orders, not just the first transaction.
• Underestimating marketplace advertising costs – fix: ads on the marketplace are their own, growing cost block, not an optional extra.
• Steering channels against each other instead of together – fix: run one shared contribution-margin metric across both channels.
• Not clearly defining role allocation – fix: decide in advance which products do which job on which channel.

Working together instead of Either/Or

The smart architecture uses both logics instead of picking one: the marketplace as a reach machine, the store as a relationship machine.

The marketplace as a reach machine – bestsellers and search-driven products where the demand lives, calculated profitably at SKU level including all fees. The store as a relationship machine – exclusive products, bundles and services only available there, plus every legal bridge from marketplace to brand (insert cards within the rules, brand search, recognisable brand codes). And one shared steering metric across both: contribution margin per order after all channel costs – only this one figure makes store and marketplace comparable.

When store, when marketplace: the decision guide

Channel choice per product can be pinned down to a few criteria, rather than deciding it once for the whole range.

• Prefer the marketplace when: the product is strongly search-driven and interchangeable, fast scaling matters more than customer retention, or your own logistics/reach isn’t (yet) sufficient.
• Prefer the store when: the product has repeat-purchase or subscription potential, brand and story are part of the reason to buy, or you need customer data for future launches.
• Both in parallel, when: the range is broad enough to split roles – bestsellers on the marketplace, exclusives and bundles in the store.

From experience: the order often makes more difference than the decision itself. Many successful sellers start on the marketplace to gather real demand data with little lead time – which products, which price points, which variants perform – and only then build their own store deliberately around those insights. The other way round, a store with zero market validation carries higher risk, because the expensive CAC learning phase runs entirely at your own cost.

What shifts over time

The channel calculation isn’t a one-off decision – it changes with brand awareness, competition and range breadth.

A store that starts out selling almost exclusively through the marketplace often shifts that ratio gradually once its own brand builds search volume and returning customers come back – store economics then improve on their own, without anything changing in the marketplace take rate. Anyone actively watching this shift, rather than setting the channel split once and never touching it again, spots early when more investment in their own store pays off.

The bottom line

“Store or marketplace” is the wrong question. The right one is: which channel does which job for which product – and does that pay off after all costs, not just the visible commission or the visible ROAS? Anyone running both channels with the same contribution-margin logic ends the holy war where it belongs: in the finance function, not in a debate of principle.