"Should we be on Amazon or our own site?" is the wrong question — but the margin math behind it is the right obsession. The same $60 product, sold to the same customer, can net anywhere from $8 to $24 depending on the channel carrying it. Most operators know the fees exist; far fewer have built the honest side-by-side. Here it is, line by line, with the strategic logic that falls out of it.
| Line | Amazon FBA | Own Store (Shopify) |
|---|---|---|
| Revenue | $60.00 | $60.00 |
| Landed COGS | -$18.00 | -$18.00 |
| Referral fee (15%) | -$9.00 | — |
| Fulfillment | -$8.50 (FBA) | -$9.50 (3PL + shipping) |
| Payment processing | included | -$2.00 |
| Platform/apps | -$0.50 | -$1.50 |
| Advertising (effective) | -$7.00 (12%) | -$14.00 first order / ~$1 repeat |
| Contribution | $17.00 (28%) | $15.00 first (25%) / $28.00 repeat (47%) |
Read the last row twice. On a first order, Amazon often wins — its conversion rates are unmatched and its ad auction, while expensive, targets buyers inches from checkout. On a repeat order, the owned store isn't a little better; it's nearly twice the margin, because the second visit arrives via a $0.01 email instead of a $14 ad, and there's no 15% referral toll on a customer relationship you already own. Every analysis that compares the channels on first-order economics alone — which is most of them — misses the entire game.
Amazon never tells you who your customer is. No email, no remarketing, no cross-sell, no cohort analysis beyond what the platform deigns to share. Strategically, marketplace revenue behaves like high-margin wholesale: real money, zero relationship. That's also exactly how acquirers price it — marketplace-dependent brands transact at visible discounts to owned-audience peers (the concentration penalty in our multiples guide), because the buyer inherits revenue without inheriting customers. The aggregator era demonstrated the terminal version of this trade at $15B scale.
The winning pattern across resilient brands is boringly consistent: Amazon as the acquisition tollbooth and search-demand harvester; the owned store as the relationship and repeat engine; and a deliberate handoff between them — package inserts, warranty registration, replenishment programs, exclusive bundles — moving second purchases to owned rails. Wholesale and retail then sit on top as margin-diverse stabilizers. Channel religion in either direction costs real money: pure-DTC orthodoxy forfeits the cheapest high-intent demand pool in commerce; pure-marketplace pragmatism builds a business you rent rather than own.
Amazon is where customers are cheapest to find. Your store is where they're cheapest to keep. Strategy is the plumbing between the two.
Sources: Amazon seller fee schedules and FBA rate cards; Shopify/processor published pricing; 3PL rate benchmarks; operator P&Ls (directional composite). Companion pieces: The $62B Tollbooth, Owned Audience Economics.
We talk to founders at every stage — long before they're ready to sell.
Start a Conversation →