Marketplaces
Date: 2026-08-16
Selling on someone else’s platform. You rent their demand and give up the customer relationship, the data, and control of the economics — and the last one can change without notice.
What it is
A marketplace is a third-party platform where you list products alongside competitors, and where the platform owns the customer, the traffic and the rules.
The trade
What you gain: existing demand at scale, no acquisition cost per order beyond the fee, fast market entry, and access to customers who search on the marketplace rather than the open web.
What you give up, and this is the half that’s understated:
- The customer relationship. Usually no email, no ability to run lifecycle marketing, no repeat purchase you control
- The data. Limited visibility into who bought, why, or what they browsed
- Margin. Referral fees, fulfilment fees, advertising to be visible at all
- Control of the economics. Fee structures and algorithm changes happen to you, not with you
- Price control, where the platform enforces parity or reprices for you
- Brand presentation, constrained to the platform’s template
The economics differ fundamentally
The critical point: a marketplace order is a one-off transaction, not a customer acquisition.
own site marketplace
CAC £22.50, contribution £15 no CAC, fee ~15% of £50 = £7.50
loses £7.50 on first order contribution £15.00 − £7.50 = £7.50
profitable immediately
1.6 orders → LTV £24 no repeat you control → LTV ≈ £7.50
LTV:CAC 1.07 no ratio — there's no lifetime
Marketplace orders are profitable per transaction and have no lifetime value. Own-site orders lose money initially and compound.
Which means comparing them on first-order profitability gets the answer backwards, and comparing them on LTV:CAC is meaningless because one of them has no CAC. They’re different businesses sharing a warehouse — Customer Lifetime Value, Payback Period.
Where it makes sense
- Incremental demand you couldn’t otherwise reach — customers who start their search on the marketplace and would never have found your site
- Clearing excess stock without discounting on your own site and training your own customers
- Market or category testing before committing to inventory
- Categories where the marketplace is where search happens, and absence is the greater cost
Where it doesn’t
- Cannibalising your own site. If customers who’d have bought direct buy on the marketplace instead, you’ve paid a fee to lose the relationship. This is the central question and it’s rarely tested
- Competing on price alone in a category where the platform can see your margins and may compete directly
- Becoming dependent. A concentration risk with an owner who can change the terms — Channel Mix
Testing cannibalisation
Harder than most incrementality questions, and worth attempting:
- Regional or category listing holdouts — list in some areas or categories and not others, compare total demand — Geo Holdout Tests
- Watch own-site orders when marketplace listings launch. A fall concurrent with marketplace growth is a signal, though confounded
- Match customers where you can. Same name and address across both channels indicates substitution rather than incremental demand
Managing it
- Report marketplace separately. Blending it into site metrics corrupts conversion rate, AOV and every cohort — the populations aren’t comparable
- Compute contribution after all fees, including advertising to be visible — Contribution Margin
- Convert where permitted. Packaging inserts and follow-up within the platform’s rules are the only route to a relationship
- Set a ceiling on marketplace share of revenue as a deliberate policy, not as an outcome