The full Business Model Canvas, block by block — rebuilt in StartupKit from Amazon's public filings. The canvas most founders draw for Amazon is a store; the real one is a machine that turns internal capabilities into external businesses — and lets the high-margin ones fund the thin ones.
Nine blocks, exactly as they'd sit in the tool — each one ends with why it matters.
Why it matters — The pivotal partner decision was opening the store shelves to competitors: third-party sellers cannibalize Amazon's own retail — and Amazon let them, because seller fees plus FBA plus ads earn more per sale than selling the item itself. If a partner can serve your customer better than you can, charge them rent instead of fighting them.
Why it matters — Notice the pattern in the list: almost every activity started as an internal necessity — warehouses for its own books, compute for its own site, shelf placement for its own products — and was later industrialized and sold to outsiders. Amazon's core activity is really 'turn cost centers into product lines.'
Why it matters — Bezos's founding logic still holds: build on what won't change — nobody will ever want higher prices, less selection, or slower delivery. Value propositions anchored in permanent human wants survive every platform shift. Ask what will be true about your customer in ten years, and build the moat there.
Why it matters — Prime inverts the loyalty program: instead of rewarding purchases with points, Amazon charges upfront and lets sunk-cost psychology do the rest — a member 'making the fee worth it' consolidates their shopping without another discount. Paid loyalty beats earned loyalty when the perceived value is overwhelming.
Why it matters — Amazon monetizes the same shopper three times: once as a buyer, once via the seller competing for that buyer's screen, and once via the advertiser paying to win the comparison. Multi-sided segment design means every new consumer makes the other segments more valuable — the flywheel is really a segment-linkage machine.
Why it matters — Amazon's resources are mostly things software companies refuse to build: warehouses, trucks, planes, chips. Capital-heavy moats are unfashionable and nearly unassailable — a competitor can copy an app in a quarter but not a thousand fulfillment centers in a decade. Sometimes the deepest moat is the one that requires pouring concrete.
Why it matters — More product searches now start on Amazon than on Google — meaning Amazon owns the channel every retail competitor must pay to access, including through its own ads business. When your storefront becomes the industry's search engine, distribution stops being a cost and becomes a revenue line.
Why it matters — Amazon deliberately runs retail near breakeven — the famous 'your margin is my opportunity' — because thin retail margins starve competitors while scale feeds the businesses that DO carry margin. Reading Amazon's cost structure as inefficiency misses the strategy: the costs are the moat, paid for by AWS and ads.
Why it matters — Look at where the money is made versus where the revenue is booked: retail generates the volume, but AWS and ads generate most of the operating income. That's the cross-subsidy engine in one line — monetize the infrastructure and the attention, keep the storefront brutally cheap. Every mature platform eventually discovers its ads business; Amazon just did it with the highest-intent audience on earth.
The one thing to copy
Amazon's flywheel is taught everywhere; the copyable part is the cross-subsidy engine underneath it. Amazon audits what it built for itself — compute, logistics, shelf space, purchase data — and turns each into an external product (AWS, FBA, ads) whose fat margins fund the deliberately thin retail that keeps the flywheel spinning. Do the same audit on your own startup: what did you build to serve yourself that others would pay for? Your next revenue stream is usually already on your cost sheet.
Clone Amazon's canvas into StartupKit's free Business Model Canvas tool and replace its answers with yours — the annotations above tell you what each block has to prove.
Free account · no card required
A flywheel of interlinked businesses: low-margin retail (first-party and a marketplace where third-party sellers move 60%+ of units) builds scale and customer data; that scale feeds high-margin businesses — AWS cloud computing, advertising, seller services, and Prime subscriptions — which fund keeping retail cheap. Each layer strengthens the others.
Mostly not from selling you things. Of its $638B in 2024 revenue, the operating income is dominated by AWS ($108B revenue at roughly 37% margins) and advertising ($56B at software-like margins), plus seller fees. Retail runs intentionally thin — scale and customer lock-in are its job, not profit.
The self-reinforcing loop Bezos sketched: lower prices bring more customers; more customers attract more sellers; more sellers add selection, which improves the experience and enables more scale, which lowers costs and prices again. Amazon's later businesses (AWS, ads, FBA) attach to the flywheel and monetize its spin.
Because it began as internal infrastructure: Amazon industrialized the compute it built for its own store and rented it out — the pattern behind FBA (its warehouses) and its ads business (its shelf space) too. The lesson: internal capabilities, productized, can dwarf the original business.
No — Amazon is not a StartupKit customer. This canvas is an editorial reconstruction from public sources: Amazon's SEC filings, shareholder letters, and executive interviews. It exists to teach the pattern, not to speak for the company.
Clone this canvas into StartupKit's free Business Model Canvas tool and replace Amazon's answers with yours. Then run the cross-subsidy audit on your key resources block: list what you built for yourself, and ask which of it others would pay to use.
Sources
Reconstructed from public sources for educational purposes. Amazon is not a StartupKit customer and has not endorsed this page.