- Amazon's 2026 Fulfillment by Amazon (FBA) increase averages $0.08 a unit, under 0.5% of a typical price, but it sits on a system of 40+ distinct fee types. [1][4]
- Referral is 6-15% on both platforms, yet operators put the all-in effective take on Amazon near 30-40% of gross merchandise value (GMV) once fulfillment, storage, and ads stack. [4]
- Walmart is structurally lighter: $0 monthly, flat $0.75 storage, lower ad CPCs, but less demand and reach. [2][3][4]
- Ad spend is now a fee, not marketing on top. Sponsored Products buys the rank, so budget it as cost of goods sold (COGS). [4]
- Third-party (3P) versus first-party (1P) selling is a margin-against-volatility trade, and inventory discipline decides it. [4]
Amazon told sellers its 2026 fee change would be slight. The average unit goes up $0.08, less than 0.5% of a typical item's price. [1] Read on its own, that is nothing. The problem is where it lands: on top of a fee system with more than 40 distinct charges once surcharges, removals, returns processing, and inbound placement are counted. [4] The referral rate you agreed to stopped describing what selling actually costs a long time ago.
The headline rate hides the real one
The referral fee is 6-15% for most categories on both platforms, with outliers: one operator breakdown puts Amazon as high as 45% on Device Accessories and Walmart's cap near 20% on Jewelry. [3] But referral is the smallest line on the bill. Add fulfillment, storage, surcharges, and the ads it takes to stay visible, and the number that actually governs your margin is much larger.
That is the gap worth planning against. Amazon frames the 2026 increase as a rounding error, and per unit it is. Compounded across fulfillment, seasonal storage, and ad inflation, it is the difference between a stock-keeping unit (SKU) that clears and one that quietly loses money while revenue still grows.
Amazon versus Walmart, line by line
| Fee | Amazon | Walmart |
|---|---|---|
| Monthly plan | $39.99 (Professional) [3] | $0, no listing or setup fees [2] |
| Referral | 6-15% most categories, up to 45% on Device Accessories [3] | 6-15%, 15% on apparel over $20, ~20% cap on Jewelry [2][3] |
| Fulfillment start | ~$3.22/unit, small standard [6] | ~$3.45/unit, ~$4.95 at 2 lb [3][5] |
| Storage | $0.78/cu ft off-peak, up to $2.40 in Q4 [4][5] | $0.75/cu ft Jan-Sep, higher in Q4 [3][4] |
| Ad cost per click (CPC), operator benchmark | $0.80-$2.50 [4] | $0.40-$1.20 [4] |
On fulfillment the two are close: Amazon starts around $3.22 for a small standard unit on the published rate card, and Walmart Fulfillment Services (WFS) starts around $3.45, roughly $4.95 at 2 pounds. [6][3][5] The gap opens in storage. Amazon charges about $0.78 per cubic foot off-peak but runs Q4 peak rates cited as high as $2.40 and adds aged-inventory surcharges starting at 181 days, about six months. [4][5] Walmart's storage is $0.75 per cubic foot January to September, with higher peak-season rates. [3][4] For fast-turning goods it barely matters. For slow-turn or bulky SKUs, Amazon's seasonal and aged-inventory penalties can flip a profitable unit into a loss.
Ad load is a fee now, not a line item
The cost that moved most is ad spend. Sponsored Products is the default path to visibility on Amazon, so treating it as marketing on top of the fee card understates the real number. Operator benchmarks put Amazon Sponsored Products clicks around $0.80-$2.50 and Walmart Connect around $0.40-$1.20, category-dependent and not a published tariff. [4] It belongs in the same budget line as the referral fee, which is the case we make in full on retail media. Book it in SG&A (selling, general and administrative expenses) instead and your channel margin looks healthier than it is.
Amazon or Walmart buys your inventory and sets the retail price. Margins are thinner but simpler, sell-in is guaranteed, and you cede control over price compression.
You keep the list price and the gross margin, and you absorb the whole stack: referral, fulfillment, storage, surcharges, and the ads it takes to hold rank. Higher ceiling, far more volatility.
None of this makes owned direct-to-consumer (DTC) selling free. A Shopify storefront skips the referral, but payments, shipping, returns, and paid acquisition rebuild a take of their own, plausibly into low double digits before a dollar of media. [6] The honest trade is margin for demand: marketplaces hand a new brand traffic and trust it cannot buy cheaply, while DTC pays off on repeat purchase and higher-margin SKUs, if you can carry the acquisition cost and the logistics. If your DTC conversion is the weak link, that is a separate problem to fix before you blame the channel.
What to actually do with this
- Model the all-in effective take per SKU, not the referral rate. Referral is the smallest line on the bill.
- Move ad spend into marketplace COGS, not SG&A. On Amazon it buys rank, so it behaves like a fee.
- Match the platform to the SKU. Slow-turn and bulky inventory gets punished by Amazon's Q4 and aged-inventory surcharges. Walmart's flat storage may suit it better.
- Reprice or cut SKUs that only clear on 1P terms. Surcharges tied to size, price, and age make sloppy 3P assortment the fastest way to lose the margin you switched to 3P to capture.
The slight increase was never the story. The story is that the fee card grew a second layer no rate card advertises, and the only defense is to price against the whole stack, not the number on the sign.
Sources
- Amazon Selling Partner Communications · Update to U.S. referral and Fulfillment by Amazon fees for 2026holds the $0.08-per-unit average increase and the under-0.5% framing
- Walmart Marketplace · Marketplace pricing$0 monthly, 6-15% referral, 15% apparel over $20
- Aura · Walmart seller fees 2026: $0/month plus 6-15% per saleoperator analysis; the 45% and 20% caps and the WFS-cheaper claim are operator estimates, not official rate cards
- NovaData · Amazon vs Walmart Marketplace 2026: an operator's guide40+ fee types, storage rates, effective take, and CPC ranges; CPCs are operator benchmarks, not published tariffs
- Sellify · Amazon vs Walmart 2026: the profitability showdownQ4 peak storage up to $2.40, 15-month aged surcharge, WFS at 2 lb
- SellerTransparency · Marketplace fee rate cardFBA small-standard fulfillment near $3.22; DTC payment and shipping take
Frequently asked questions
What is the real take rate on Amazon in 2026?
The headline referral fee is 6-15% for most categories, but that is not what selling costs. Once you stack fulfillment (from about $3.22 a unit), storage with Q4 and aged-inventory surcharges, and the Sponsored Products spend it now takes to hold rank, operator estimates put the all-in effective take near 30-40% of GMV, and higher in competitive categories. That figure is an operator estimate, not a platform-published metric, so model your own profit and loss (P&L) per SKU rather than trusting a blended number.
Is Walmart Marketplace cheaper than Amazon for sellers?
Structurally, yes, on the fees you can see. Walmart charges no monthly subscription, referral of 6-15%, WFS fulfillment from about $3.45 a unit, storage at $0.75 per cubic foot January to September (higher in Q4), and generally lower ad CPCs. Amazon adds a $39.99 monthly Professional plan, Q4 storage surcharges, and pricier clicks. The catch is demand: Walmart has less traffic and weaker international reach, so lower cost per unit does not automatically mean more profit.
Should I treat marketplace ad spend as marketing or as a cost of goods?
As a cost of goods, at least on Amazon. Sponsored Products has become the default path to visibility, so the spend buys rank the way a fee buys placement. Booking it in SG&A rather than marketplace COGS flatters your channel margin and hides the real contribution per SKU. Operator CPC benchmarks run roughly $0.80-$2.50 on Amazon and $0.40-$1.20 on Walmart Connect, category-dependent.
What is the difference between 1P and 3P selling?
In 1P (vendor), Amazon or Walmart buys your inventory wholesale and sets the retail price. Margins are thinner but simpler, and sell-in is guaranteed. In 3P (marketplace/seller), you keep the list price and gross margin but absorb the entire fee stack, including ads. 3P has a higher ceiling and much more volatility, and in 2026 the surcharges tied to size, price, and inventory age make inventory discipline, not channel loyalty, the thing that decides whether it pays off.
How big was Amazon's 2026 fee increase really?
By Amazon's own figure the average unit went up $0.08, less than 0.5% of a typical item's price, so on a single unit it is close to nothing. The catch is that it lands on top of a system with more than 40 distinct fee types once surcharges, removals, returns processing, and inbound placement are counted. Compounded across fulfillment, seasonal storage, and ad inflation, the slight increase is not the real story.
Does selling direct-to-consumer on Shopify avoid these fees?
Not really. A Shopify storefront skips the referral fee, but payments, shipping, returns, and paid acquisition rebuild a take of their own, plausibly into the low double digits before a dollar of media. The honest trade is margin for demand: marketplaces hand you traffic and trust you cannot buy cheaply, while DTC pays off on repeat purchase if you can carry the acquisition cost and the logistics.
How do Amazon and Walmart storage fees compare?
This is where the two platforms diverge most. Amazon charges about $0.78 per cubic foot off-peak but runs Q4 peak rates cited as high as $2.40 and adds aged-inventory surcharges starting at 181 days (about six months), while Walmart charges $0.75 per cubic foot January to September, with higher peak-season rates. For fast-turning goods it barely matters, but for slow-turn or bulky SKUs Amazon's seasonal and aged-inventory penalties can flip a profitable unit into a loss.
Which product categories carry the highest referral fees?
Referral runs 6-15% for most categories on both platforms, but there are outliers. One operator breakdown puts Amazon as high as 45% on Device Accessories, while Walmart caps near 20% on Jewelry and charges 15% on apparel over $20. Those high-end figures are operator estimates rather than a single official rate card, so confirm the rate for your specific category before you model margin.
Mikołaj Salecki
Writes about media, tech, and AI business for people who actually run digital. Former agency lead. Skeptic of frameworks that read better than they perform.
Get the next issue by email.
One letter, once a week. Sharp coverage of media, tech, and AI business. No filler.
