🇺🇸 Jeff Bezos(b. 1964)

One-way doors and two-way doors

Bezos's most portable idea is a classification, not a process. Type 1 decisions are irreversible — sell the company, change the pricing model permanently — and deserve deliberation, dissent and delay. Type 2 decisions are reversible, and treating them with Type 1 seriousness is a form of organisational decay he associated with size. His stated target was to decide with roughly seventy percent of the information you wish you had, because waiting for ninety percent means being slow in most cases that matter.

• Speed and quality are only in tension for the small share of decisions that cannot be undone.
• Most organisational sluggishness is misclassification, not caution.
• 'Disagree and commit' resolves stalemates without pretending consensus was reached.

Working backwards: the press release before the product

Amazon teams write the launch press release and the customer FAQ before building anything. If the release is not compelling in plain language, the product is not worth constructing — and the document, unlike a slide deck, cannot hide vagueness behind bullet points. Bezos banned PowerPoint in leadership meetings for this reason and replaced it with six-page narratives read in silence at the start of the meeting.

The flywheel: why Amazon accepted low margins for decades

The flywheel Bezos described is causal, not motivational: lower prices bring more customers, which brings more third-party sellers, which improves selection, which spreads fixed cost over more units, which permits lower prices again. Once you believe that loop, sacrificing near-term margin is arithmetic rather than courage. It also explains why Amazon's most profitable business, AWS, emerged from internal infrastructure — spare capacity turned into a product because the loop rewarded utilisation.

Day 1 and the specific decay he was guarding against

'Day 2 is stasis, followed by irrelevance, followed by excruciating, painful decline.' Bezos named the symptoms precisely: proxies replacing customers, process becoming the point, slow decision-making, and resistance to external trends. His countermeasures were mechanical — single-threaded owners, small teams, written narratives, customer anecdotes weighted alongside metrics — because he did not trust culture to survive on intention alone.

Failure as a budgeted input

Bezos argued that outsized returns are inherently asymmetric: a failed experiment costs its budget, a successful one can return many multiples, so a company that wants big wins must fund a high absolute number of failures. The Fire Phone's write-down is the example he cited himself, alongside the observation that AWS and Prime came from the same appetite. The uncomfortable implication is that a division with no visible failures is probably under-experimenting.