We were recently working on a project that required us to read through Amazon’s financials, and what we found really shocked us. In our many, many years of reading financial press releases we have never come across anything like Amazon’s. Amazon does a lot of things. Their earnings releases read like a catalog with pages of updates across dozens of businesses. We struggled to parse just those pages, so we have to wonder how does anyone manage a company like this?
And over the holidays, we actually found several Twitter threads touching on this subject. First this one talking about the early days of Amazon, AWS and the last Internet Bubble. And then this one comparing AWS’s go-to-market versus GCP’s. And both of these reminded us of Steve Yegge’s famous oops-this-shouldn’t-be-public rant about the management history of Google versus Amazon. The summary of that one is basically Amazon has succeeded because they abstracted every business unit and infrastructure component into corporate (and digital) APIs.
Our point here is that Amazon seems to have created, or possibly just stumbled onto, an important evolution in the way businesses can be run. Amazon seems to run all of its business units as semi independent entities. The thinking about the history of management and corporate finance turned against conglomerates years ago. In the 1970’s the US seemed to be awash in giant do-everything conglomerates, and then the 1980’s wave of hostile takeovers broke those apart. Since then the common wisdom has held that companies should focus on their core competencies. This is the best way to align shareholders and employees’ interests.
Amazon clearly defies that logic. And we think there are two ways to understand this.
The first is that a big criticism of conglomerates holds these corporate structures create too much friction which takes away from the overall value of the holdings, 2 + 2 < 4. Every business unit needs to have its own accounting functions and reports up and down the management chain become an immense time and reource drain. Amazon seems to have solved this through the use of digital systems. With this model, the various pieces have a low-cost, low-friction way of interacting. Management can isolate responsibility in a way that simply wasn’t possible to companies prior to the Information Age. This is an important innovation.
The other way to view this is that Amazon’s core competency is running disparate businesses. Jeff Bezos is famously averse to many of the common forms of corporate bloat – excess travel, endless bloated PowerPoint presentations for internal consumption, etc. Instead, they have a system that allows for all of the businesses units and support functions to inter-operate in a way that requires no (expensive) human oversight. By dong this, they have probably eliminated three layers of management that a traditional company would require. All the reporting, evaluations, accounting and controls are automated.
Now it is possible that this works because the company has essentially existed in a bull market for the past 20 years. When (not if) the broader economy turns down, the tide may go out and reveal some critical flaw in their system. And the fact that employees get something like 50% (or more) of their compensation in the form of Amazon’s ever-increasing stock is definitely something that could become a problem when (not if) the stock market turns down.
That being said, we believe pretty firmly that Amazon has a model that marks an important turning point in how companies can be run.
Pingback: Intel Selects Option C – All of the Above | Digits to Dollars·
Pingback: Intel 2.0’s customer dilemma | Digits to Dollars·
Pingback: Intel 2.0's customer dilemma: the foundry move - Moloft·
Pingback: Intel 2.0's customer dilemma: the foundry move - SimpleNews·
Pingback: Intel 2.0’s customer dilemma: the foundry move – JERVIE.com·