Amazon Unbound

Amazon Unbound

I read The Everything Store by Brad Stone three years ago, and I loved it. It provided an incredible window into Amazon, one of the most enduring companies in the world. Hearing the stories, anecdotes, and leadership principles gave me a lot of ideas for how we might function at Segment.

But, the story wasn't done there. The Everything Store came out in 2010. AWS was only a nascent business, there was no mention of Alexa, Kindle, Whole Foods, or Amazon Go. In the 10 years since, Amazon has grown to be an unstoppable force.

Amazon Unbound is the sequel (also by Brad Stone). In it, he tells the story of everything that's happened in the last 10 years. It's truly remarkable how many times the company has reinvented itself.

Alexa, Fire, and Kindle

I didn't know this, but Bezos designed the first version of the Alexa in his office. He had the idea that the future of computing would all take place over voice.

There were a few competing efforts to get Alexa to be really good at understanding the human voice. The first efforts were downright atrocious. When launching the Alexa, Bezos asked all teams to consider "how would they support Alexa?".

Prior to the book, I don't think I'd considered the strategic importance of Alexa. The Alexa serves as Amazon's "gateway to the home", and Bezos saw the future of Amazon's orders coming directly via Alexa. In a similar manner to Google's search product, Alexa also gets better the more that its used. While other natural language products struggle to get enough training data, Alexa is constantly collecting data that then enhances other AWS offerings like Polly.

Amazon Go

If you've ever tried out the Amazon Go stores, you can probably attest that the experience of leaving the store feels magical. You just "walk out".

The book tells a new store around Amazon Go's origin–it was originally supposed to replace grocery stores rather than convenience stores.

But Bezos had a different version (and in my opinion, great taste in describing the problem): "I should be able to just walk out!". The book gives the impression that the store is largely still an unproven experiment, but you could see how they could be used as a jumping off point for competing with Doordash or "instant" deliveries.

The Washington Post

Another new learning for me: Bezos has a soft spot for the idea of independent journalism, and also helped turn around The Washington Post by adding in Amazon-style leadership principles.

Apparently the paper was losing money all over the place, and didn't really have a software team. Bezos had the paper hire digital advertisers, push for a subscription model, get rid of local news reporting, and brought in several former Amazon execs. As a result, The Washington Post grew more quickly than even the NYT for a few years.

Loops upon loops

The book has made me even more impressed with Amazon's virtuous product loops that work at scale. There's the basic: "lower prices > more customers > more sellers > lower prices" that's easily understood. But beyond that, there's almost a fractal ecosystem of loops:

  • "more customers > better private label data > better Amazon Basics > more customers"
  • "more customers > better recommendations > more customers"
  • "more customers > increased ad sales > lower prices > more customers"
  • "more customers > more logistics data > robot-powered FCs > lower prices > more customers"

It's a fantastic reminder of the advantages a single company can leverage when it is willing to invest in technology at the scale that Amazon does.

One non-obvious loop is their private label offerings. Because they get the best data on what customers want, they can offer recommendations, and straight up better products. A good example in the book was around dog poop bags... most of them were hard for customers to figure out how to use, so Amazon released one labeled "open this end".

Amazon can then take those better products, and make label them as the "Amazon recommended" cohice. Because Alexa uses the "recommended" item as the default, they can further increase sales and decrease the cost of goods.

Another is the robotic delivery. Because Amazon sees such scale, they can automate the parts of their warehouse delivery in a way nobody else can.

Blue Origin & SpaceX

I hadn't known much of the history of Blue Origin, but it's interesting to contrast the results with those that Elon has achieved at SpaceX.

Blue: took money mostly from Bezos, required no profit model, had a scientist rather than a CEO, took a long time to R&D everything.

SpaceX: raised money from VCs, signed numerous government contracts, was led by a CEO with the grand the vision to occupy Mars.

The DNA of SpaceX is arguably much better, it focuses on delivering results, and incrementally lining up bigger and bigger launches, funded mostly by government contracts. No amount of talent can really overcome the incentive mismatch.

Anti-trust and working with government

The book ends with a large discussion about whether Amazon is really a pro-consumer company, or whether it pushes the bounds of anti-trust monopolies. There seem to be pretty legitimate concerns around Amazon abusing it's power over sellers... though it's also difficult to get any sort of hard facts on them. Even inside the government testimony, the complaints of sellers are purely anecdotal.

The most interesting part of this section was revealing how little muscle Amazon has built around working with the government. Organizations like Microsoft and Facebook have been dealing (and sometimes collaborating) with governments for years. Even Elon has managed to win numerous subsidies and contracts that helped both SpaceX and Tesla.

Solving the Principal-Agent Problem

One meta-theme I've noticed throughout the book: Jeff Bezos is incredible at getting his teams to think bigger. He's sort of a sci-fi futurist geek, so he pushes his teams to think about what they might achieve, whether that's voice-activated software, or 3d holographic phones.

I think one of the biggest assets that Bezos brings to Amazon is almost a quasi-solution to the principal-agent problem. Bezos wants teams to take a lot bets, because it's better for Amazon even if only a handful of those risky bets end up as a success. He's constantly asking how the business can invest more funding in exchange for bigger returns–and he normalizes talking about his own personal failures (fire phone, healthcare efforts, etc).

Overall, I've come away from the book with even more respect for Bezos and what he has built. He's a genius at driving the proper organization-wide incentives, and instantly drilling down to various key parts of the business (there's numerous anecdotes about him asking some key numbers question at s-team meetings).

In some ways, he comes across as a bit of a mercurial leader. But what's most impressive to me is that Amazon seems like it will be totally fine without him.