Skip to content
The Founder's Notes

Founder index · Amazon · 1964–, founded Amazon 1994

Jeff Bezos

The lesson that travels

If you can persuade shareholders to judge you on a different metric, you can make decisions your competitors are structurally forbidden from making.

Bezos left a quantitative hedge fund in 1994 with a list of twenty categories that might sell well online and chose books, not out of affection but because print carries more titles than any physical store can stock. The real founding act was the 1997 shareholder letter, in which he told investors he would optimise for long-term free cash flow per share rather than quarterly earnings, and would make decisions that looked bad on the income statement. He then attached that same letter to every annual report for twenty-four years. It reads as sentiment and functions as a contract — it bought Amazon permission to lose money for six years while Walmart and Barnes & Noble were being punished for a single soft quarter.

The two decisions that built the modern company share one shape. In 2000, after the failure of zShops and Auctions, Amazon opened its own product pages to third-party sellers — putting rivals' offers, sometimes cheaper, on the same page as its own. The retail team fought it because it cannibalised first-party margin directly. Third-party sellers are now roughly 60% of units sold and the higher-margin half of retail. In 2006 Amazon took the internal storage and compute layer it had built for its own engineers and rented it to anyone with a credit card. AWS is now roughly 58% of the company's operating income.

Both are the same move: take something built for internal advantage, expose it to outsiders on identical terms, and let outsiders pay for your scale. Very few companies do it, because it requires arming your competitors.

§01The hard calls

  1. 1997

    Tell public shareholders that reported profit is the wrong metric — in writing, permanently

    Six years to first full-year profit, tolerated. No competitor operating under conventional earnings expectations could have matched the strategy.

  2. 2000

    Let third-party sellers list on Amazon's own product pages, undercutting Amazon's own retail

    Approximately 60% of paid units today. Seller services revenue reached about $156B in 2024 on inventory Amazon never bought.

  3. 2005

    Launch Prime at $79 for unlimited two-day shipping, with no analysis proving it could be profitable

    Converted the largest friction in e-commerce into a sunk annual cost, and made every subsequent purchase feel free at the margin.

  4. 2006

    Sell Amazon's internal infrastructure to the public as AWS

    Approximately $108B of revenue at 37% operating margin in 2024, funding a retail business that earns close to nothing.

  5. 2014

    Ship the Fire Phone

    A $170M write-off. Hardware without a platform is a gift to customers. The failure directly informed how Amazon later positioned Alexa devices as loss-leaders rather than products.

One of these is set up as a tutorial — you make the call before you find out what Jeff did.

Face the decision