A Seller Squeezed by Marketplace Fees

What happened

A seller launched a kitchen-goods brand on Amazon and sales came fast, because the marketplace supplied the demand. Then the fee stack grew: a referral fee, fulfillment, storage, and advertising that had become mandatory in practice, together taking about half of revenue. A fee change and rising ad costs pushed several products below break-even, and the seller realised the platform owned the customer. He built his own site and email list for repeat purchases, kept Amazon for discovery, and his blended margin recovered.

The case, start to finish

No single fee was unreasonable. The stack was, and a stack never announces itself the way a price increase does.

Selling where the buyers already are

This is an anonymized seller carrying a documented pattern: the rising all-in cost of selling on Amazon, as tracked in Marketplace Pulse's published analysis of seller economics. The seller is a composite. The fee structure is not.

The first year felt like validation. A new kitchen-goods brand went up on the marketplace and sales came quickly, because the platform supplies the one thing a new brand does not have, which is people already standing there intending to buy. That is a real service and it is worth paying for. The question is what the payment becomes once the business depends on it.

The stack

By year two the bill had assembled itself. A referral fee of roughly 15% on each sale. Fulfillment and storage through the platform's own logistics. And advertising, which is nominally optional and functionally mandatory, because visibility is an auction and the sellers who decline to bid become invisible to the ones who do. Marketplace Pulse puts the common all-in take at around half of revenue.

The reason this erodes a business quietly is that it never arrives as an event. Each fee is defensible on its own, and each was introduced or raised separately. Nobody receives a notice saying the platform now takes half. They receive a sequence of individually sensible adjustments that arrive at half. And because the fees scale with revenue rather than sitting as a fixed cost, growth does not dilute them. The platform's share is structural.

Underneath the fees sat the more expensive fact. The customer belonged to the marketplace. Every repeat purchase from a buyer the seller had already won paid the full take rate over again, indefinitely. No relationship existed outside the platform through which that buyer could be reached.

The move that failed, and the one that worked

The first instinct was to price around it. Raise on-platform prices, restore the margin. This fails mechanically: the platform's ranking and Buy Box logic answer uncompetitive prices with invisibility, and the fee stack scales with the higher price anyway. The seller kept pennies of each raised dollar and lost the volume that had made the math work at all. A fee stack cannot be out-priced. Only the dependence can change.

By year three a fee change and rising ad costs had pushed several products below break-even, which is what forced the reframing. The marketplace stopped being the business and became a channel. Pay it generously for what it is genuinely excellent at, which is putting a new brand in front of a first-time buyer, then route the second purchase somewhere the fees do not follow. A note in the box and a reason to hand over an email address did the converting, and blended margin recovered.

Rented distribution

The useful frame is ownership. A marketplace, an app store, a referral network, a franchise territory, a search ranking: each of these is rented distribution, brilliant at reach and permanently expensive, on terms the renter does not set. There is nothing wrong with renting. The mistake is mistaking the rental for the business, and the exposure has a name, which is platform risk.

Two questions follow, and they are worth asking of any channel rather than only this one. What is the all-in take once every fee, commission and effectively mandatory promotion is counted? And who owns the second purchase? A business whose customer list lives entirely on somebody else's server finds out the answer on the day the terms change.

Timeline

• Year 1 Seller launches a kitchen-goods brand on Amazon; sales come fast, because the marketplace supplies the demand.
• Year 2 Fee stack grows: referral fee + FBA fulfillment + storage + the now-mandatory-in-practice advertising. All-in take reaches ~half of revenue.
• Year 3 A fee change and rising ad costs push several SKUs below break-even; the seller realizes the platform owns the customer.
• Year 4 Rebalances: builds a DTC site + email list for repeat purchases, keeps Amazon for discovery. Blended margin recovers.

You're in the owner's chair

Year 3. A fee change pushes several SKUs below break-even, and you realize every repeat purchase pays the full fee stack again because Amazon owns the customer. What do you do?

• Amazon for discovery, your site for repeats
• Quit Amazon entirely — the fees are extortion
• Raise prices on Amazon to restore margin

Where each $100 of Amazon revenue went

• Referral fee (~15%): $15
• Fulfillment + storage (typical FBA): $20
• Advertising: the visibility auction: $15
• Left for product cost AND profit: $50

Illustrative split of the commonly-reported ~50% all-in platform take (Marketplace Pulse). Each fee looks reasonable alone; the stack quietly consumes the margin.

Business model

Sell physical products where the buyers already are. The marketplace supplies traffic, trust, and fulfillment, and it charges for each, in fees that compound quietly.

Revenue model

Marketplace sales at retail price, minus the stack: referral fee (~15%), fulfillment, storage, and advertising (the auction for visibility). Marketplace Pulse pegs the common all-in total around 50% of revenue.

Cost structure

Product cost plus the fee stack. Because fees scale with revenue, growth doesn't dilute them: the platform's share is structural, and ad costs rise as more sellers bid for the same shelf.

Strategic challenge

Margins eroded year over year without any single dramatic event. Each fee looked reasonable alone; together they consumed the profit. And the customer relationship belonged to the platform, so repeat purchases carried full fees every time.

Key decision

Stop treating Amazon as the business and start treating it as a paid discovery channel: move repeat purchases to owned channels (site + email) where the fee stack doesn't apply.

What worked

Using the marketplace for what it's genuinely best at (new-customer discovery) while routing the second purchase to owned channels; package inserts and email capture converted platform buyers into direct customers.

What failed

Trying to price around the fees first: raising prices on-platform just lost the buy box and rankings. The fee stack can't be out-priced, so the dependence itself had to change.

Risk factors

Fee increases and policy changes (unilateral); ad-auction inflation; account suspension risk; the platform launching competing private-label products; single-channel revenue concentration.

Lesson summary

A marketplace is rented distribution: brilliant for discovery, expensive forever. Know the all-in take rate, and own the repeat purchase. A business that owns no customer relationships owns very little.

Key data

• ~50% of revenue (Marketplace Pulse) Common all-in Amazon take
• ~15% Referral fee alone

Sources & basis

The business in this story is a stand-in, not a company you can look up. This case is an illustrative composite: the operator, the people and most of the dollar figures represent a pattern rather than reporting one firm's history. What the list below cites is the other half, the documented industry data and public reporting the composite was assembled from, including any real company whose published figures the case draws on by name. The mechanism and the arithmetic are real even where the business is not.

1. Marketplace Pulse research on Amazon seller economics — referral + FBA + advertising commonly totaling ~50% of seller revenue View source ↗
2. Published Amazon fee schedules (referral ~15% in most categories; FBA fulfillment and storage rates)