You Haven't Bought Anything on Amazon in Years. You're Still Paying Amazon.

Say you buy a $3 pen at the stationery shop down the street. No Prime membership involved. No app, no algorithm, no one-click checkout. And yet a slice of that $3 may still be flowing, indirectly, toward Seattle.

That’s the claim behind a phrase marketer Seth Godin has been pushing back into circulation: the Amazon tax. It’s worth stating up front that this isn’t a fresh scandal. It’s a slow structural drift that’s been running for years, which is exactly why it’s easy to stop noticing. So let’s skip the outrage and go through the numbers.

Half the sale price never reaches the seller

Start with the headline figure. Amazon’s take rate — the share of a third-party seller’s revenue that ends up with Amazon — has climbed for years and now sits around 50%. That estimate comes from a nonprofit research group reverse-engineering Amazon’s own disclosures. Amazon disputes the methodology.

But the arithmetic isn’t mysterious. Base referral fees run 8-15% depending on category. Fulfillment by Amazon layers on pick-and-pack fees. Storage is billed separately, with surcharges for inventory that lingers. And then there’s advertising — the piece that quietly turned into the biggest line item.

Sponsored placement is technically optional in the way that breathing is technically optional. Skip it and your product lands on page three of search results, where roughly nobody goes. Amazon’s ad business now clears $50 billion a year in revenue. That money comes almost entirely from the people selling things on the platform.

The “just sell somewhere else” argument doesn’t survive contact with the market

The standard rebuttal writes itself: if the fees are too high, leave. Markets sort this out.

Except the exit isn’t real. A majority of US product searches now start on Amazon rather than Google. When the buyer opens the Amazon app before they open a browser, Amazon isn’t one channel among several for the seller — it’s the market itself. That’s not a negotiating position. That’s a landlord.

There’s also the pricing side. For years Amazon enforced price parity clauses: sell cheaper on your own site and face consequences. Regulatory pressure in the US and EU forced the explicit language to soften, but sellers describe the same mechanism operating informally — undercut Amazon elsewhere and you quietly lose the Buy Box, which is where the overwhelming majority of sales actually happen.

The downstream effect is predictable. Sellers price their own storefronts as if the Amazon fee applies, because effectively it does. When Amazon’s rates rise, prices rise across every channel that seller touches. Shoppers who never open the app still get handed a price list Amazon shaped. That’s why “tax” is the word people reach for.

The metaphor is only half right, and the wrong half matters

Here’s where I’d push back on Godin slightly. Taxes are, at minimum, disclosed. Rates are published. Budgets are audited, badly, but publicly.

The Amazon tax has no such property. Spend $30 and nothing on the receipt tells you how much was product and how much was platform. Sellers can’t clarify it either — publicly itemizing platform costs at checkout tends to violate the platform’s terms.

And the counterargument deserves a fair hearing. Amazon delivers real infrastructure: warehousing, two-day logistics, payments, returns handling, fraud protection, buyer trust, and search traffic that would cost a fortune to buy directly. A small merchant assembling all of that independently might well spend more than 50%. Department stores historically took around 30% of the sale price for shelf space and foot traffic alone. A higher number online isn’t automatically predatory.

So the real question isn’t whether the fee is large. It’s whether the price is set by competition or by position. A rate negotiated where there’s no viable exit isn’t a market price, no matter how good the service attached to it is.

The same machine, running in every market

None of this is a uniquely American problem. Korea’s Coupang holds a dominant share of domestic e-commerce, and while its direct-purchase Rocket Delivery model differs structurally from Amazon’s marketplace, the pressure on sellers rhymes: fees, ad spend, ranking visibility, lowest-price demands. Korea’s antitrust regulator has already moved against search-ranking manipulation there. In Europe, the Digital Markets Act codified self-preferencing rules for exactly this reason.

Food delivery apps are the most honest version of the phenomenon, because they gave up hiding it. Dual pricing — one price for the dine-in menu, another for the delivery app — is now standard across Korea, the US, and much of Europe. That’s the Amazon tax, itemized and visible.

Visible is better. A shopper who can see the markup can decide whether it’s worth it.

The version that should worry you is the invisible one: the platform cost already dissolved into the price of something you bought from a neighborhood store that has never listed a single product online. There’s no receipt line for that. There’s no way to check.

So how much are we actually paying

Platform fees look like a private matter between a seller and a marketplace. They aren’t. They reprice the entire category, including the parts of it that opted out.

The cost of convenience was never the $3 shipping charge. It’s a few percent spread thin across everything, everywhere, permanently.

If you’re curious, there’s a test you can run yourself. Next time you’re about to check out, open the manufacturer’s own store and compare. If the price is identical, that’s not a coincidence — that’s the answer.