Amazon Marketplace: The Flywheel That Funds Itself

Amazon generated $156.1 billion of third-party seller services revenue in 2024. The category includes commissions, fulfillment, shipping, and related services; it grew 11.5% from the prior year. Independent sellers supplied more than 60% of sales in Amazon’s store, placing most of the merchandise volume outside Amazon’s own inventory balance.

The arrangement changes a retailer’s central constraint. A conventional merchant must predict demand, buy stock, and accept markdown risk before offering a product. Amazon lets independent businesses provide capital, selection, and product judgment, then charges when demand passes through its store and infrastructure.

Amazon invited independent sellers to list alongside its own retail inventory around 2000. The decision was controversial because outside merchants would compete for product-page space and customer spending. It also allowed Amazon to offer items that were too narrow, seasonal, uncertain, or numerous for a single retailer to buy efficiently.

Third-party sellers own their inventory even when it sits inside an Amazon fulfillment center. They choose products, fund production, and absorb much of the risk that demand fails to appear. Amazon supplies discovery, transaction infrastructure, rules, and optional services. The catalogue can therefore grow faster than Amazon’s willingness to purchase stock.

The seller share of physical gross merchandise sales rose from 3% in 1999 to 58% in 2018, according to Amazon’s 2018 shareholder letter. By 2025, the company said independent sellers accounted for more than 60% of store sales. First-party retail continued to grow during much of that period, so the marketplace expanded the total system rather than simply transferring a fixed pool of sales.

A larger seller base increases the chance that Amazon has the exact model, size, color, spare part, regional brand, or new invention a customer wants. Multiple offers on one product page can improve price competition and availability. The customer can compare offers without visiting separate merchant sites or managing several checkout systems.

Selection matters because shopping habits form around expected success. A customer who believes Amazon will have an item starts the search there, creating traffic even before a purchase occurs. More searches and transactions generate observations about demand, price, conversion, returns, and delivery promises.

Amazon uses aggregated activity to improve search, recommendations, merchandising, fraud controls, and inventory placement. Sellers receive access to a large pool of purchase intent and tools for listing, pricing, analytics, and brand management. Each side contributes evidence that makes the store easier to operate at scale.

Reviews add another shared asset. A seller benefits from product feedback collected inside a destination that customers already consult, while buyers gain evidence from prior transactions. Amazon can surface recurring defects, answer common questions, and remove abusive behavior across many categories. The system becomes harder to reproduce because trust attaches to individual products, merchants, buyers, and the marketplace at the same time.

Traffic lowers the cost of reaching buyers for a merchant that lacks a national brand or its own acquisition engine. Sellers enter an environment where customers already have accounts, stored payment methods, Prime expectations, reviews, and trust in Amazon’s service policies. Referral and related fees arrive against that distribution.

That distribution advantage supports a high-margin adjacent business. Amazon produced $56.2 billion of advertising-services revenue in 2024, up from $46.9 billion in 2023. Sponsored placements allow sellers and brands to compete for visibility near a purchase decision. Marketplace traffic creates the inventory; seller competition creates demand for the ad product.

Advertising can help a new product find customers, but it also changes seller economics. Organic ranking, paid placement, fees, and fulfillment costs all claim part of the transaction. Amazon must keep the combined package attractive enough that merchants continue adding useful selection rather than shifting their best products elsewhere.

A marketplace fragments when every seller offers different delivery speed, returns, packaging, and customer support. Fulfillment by Amazon lets merchants send inventory into Amazon’s network, where the company stores, picks, packs, ships, handles service, and processes returns. Eligible products can carry a familiar Prime delivery promise.

Shared infrastructure pools volume across many sellers and Amazon’s first-party business. Denser flows can improve utilization of fulfillment centers, sortation, transport, and last-mile routes. Sellers gain capabilities that would be expensive to build alone, while Amazon earns fees and receives more volume over which to spread fixed network costs.

The company said entrepreneurs had shipped more than 80 billion items through FBA since its 2006 launch. Amazon also allows seller-fulfilled offers when merchants meet service requirements. Optionality broadens participation, while the Prime standard keeps the customer experience from dissolving into unrelated storefronts.

Seller commissions, fulfillment fees, advertising, and subscriptions create several revenue streams around the same transaction. Amazon can reinvest in warehouses, delivery capacity, fraud prevention, seller tools, and the consumer experience. Faster and more reliable service raises conversion, which increases the value of joining the marketplace.

First-quarter 2025 data shows the scale of the service layer. Third-party seller services produced $36.5 billion of net sales, advertising produced $13.9 billion, and subscription services produced $11.7 billion. Online-store revenue was $57.4 billion, yet Amazon recognizes third-party transactions on a net basis rather than recording the full merchandise value as revenue.

The accounting makes comparisons easy to misread. A marketplace dollar represents Amazon’s fee, while a first-party dollar generally represents the product sale. The economic advantage comes from earning several service revenues without purchasing every unit carried through the store.

Working capital adds a quieter benefit. Customers usually pay at checkout, while Amazon remits seller proceeds later under its settlement terms. The company does not need to fund the underlying merchandise purchase in the same way as first-party inventory. At scale, the timing between cash collection, seller payment, returns, and service charges affects how much capital the retail system consumes.

Software can accept seller listings quickly. The difficult sequence is attracting enough buyers to interest sellers, enough selection to interest buyers, and enough transaction density to support fast fulfillment. Reviews, seller histories, fraud systems, stored payment credentials, logistics capacity, and advertising demand accumulate over years.

A rival can subsidize one side, but weak service on the other side breaks the promise. Inviting sellers without traffic creates empty shelves with no customers; attracting shoppers without reliable sellers creates poor availability and trust. Amazon entered the loop early and repeatedly reinvested transaction revenue into both sides.

Seller economics. Rising fees, advertising dependence, returns, and inventory charges can remove the merchant’s profit even while gross sales rise.

Customer trust. Counterfeits, manipulated reviews, unsafe products, or inconsistent service weaken the shared reputation that all sellers borrow.

Governance conflict. Amazon sets marketplace rules while also selling products and advertising. Perceived self-preferencing can drive merchant exits and regulatory action.

A marketplace compounds when each new supplier improves customer choice and each new customer improves supplier economics. Measure selection that converts, repeat purchasing, seller contribution after all fees, and service quality before celebrating listing count. If merchants must keep buying more visibility to reach customers they helped attract, extraction is outrunning reinforcement. The loop will eventually lose useful supply.

Historical cutoff: July 27, 2025. No event or financial result published after that date is used in this analysis.