Your Google point (besides being a contrast to amazon's surprising athena) is an interesting way of describing things.
Youtube, Android and even adwords & web analytics were bought fairly formed, and clearly were quite good already. That's a skill in itself. When you think of examples of ebay, skype and how few buy-outs really do succeed like that.
Google could make the web work too. Remember how slow yahoo a basic webmail service was. The idea that they'd make an excel that worked.. it wasn't trivial.
But I think you're right on the lead skill. If it works best with an exabyte of data, a billion CPU-years... it works best on a google.
So, is Amazon really more B2B, while Google is B2C?
Is Amazon using e-commerce like a US university looks at undergrads for funding?
Stage 1: Railroads are going to be the next big thing. Smart play: Be one of the first with a railroad.
Stage 2: Everyone knows railroads are the next big thing. Tons of people are raising money to build railroads. Smart play: Sell those people the stuff to build railroads.
So now - what is Stage 3; once everyone has a railroad - then what?
Stage 2: make it easy for everyone to have a railroad by leasing them yours. Lease yours multiple times over, each time for less than it would cost the customer to build a new one.
Stage 3: everyone has a railroad which they rent from you. Collect rent on all railroad activity forever.
Welcome, you just reinvented the problem of rent-seeking.
Then it's a long time after the initial Industrial Revolution, and you're in a fantastic world where it's commonplace for ordinary people to own a steel machine which can travel as fast as a train at "high speed."
They realised other people will have this problem (maybe at different times), and they already had the technology.
Granted its easy to speculate with hindsight.
For example, many were sick of Wal-Mart. Was it because of the way they treated their workers? No. If it were, they wouldn’t have ran to Target and Amazon instead. The actual shopping behaviors are guided by convenience but judtified by cultural associations. Target capitalized on the same image appeal without offering anything new at all. Amazon is, culturally, Target + Silicon Valley. Ad agencies all understand this very well. Amazon is neoliberalism in the flesh. There’s nothing a bit surprising about it.
They also appeal more to women, and eschew product categories that men and poor folk flock to like car parts, guns, fishing, Jesus books, crafting, etc.
Target does not eschew most of those categories (guns and car parts are the only ones I think they consistently don't carry). Also, as an aside, crafting isn't a “men and poor people” category, anyway, but a famously huge interest for middle class women.
Target probably has about 40 feet of that stuff, mostly scrapbooking and similar stuff. No yarn, no knitting needles, etc.
Fishing in my local Target in a single 4-ft section, next to 12ft of water bottle and Thermos jars. Walmart is probably two half aisles of stuff.
Other than guns and car parts, I've seen all of the above at Target.
(However I'm in Houston, so product selection may vary regionally)
Not accurate to say that Amazon is selling its own infrastructure, or not accurate to say that they had the internal competencies to launch AWS due to building their own infrastructure? I've always assumed it's the latter -- AWS is an abstraction of tools they needed to build anyway -- rather than literally sharing compute etc. with Amazon services.
Look at any individual AWS service by itself, none of them are all that groundbreaking or innovative. What is groundbreaking is that Amazon has managed to take all of these services, and distribute them through a massively complex and infinitely scaling ecosystem. That’s logistics.
A lot of what Google does, or fails at, makes a lot more sense when you actually think about what its core competencies are: big data, anything that works best with an exabyte of data, a billion CPU-years, and some crazy machine learning / good ol' fashion statistical elbow-grease.
It's obvious how this relates to their original business, Search, and how it relates to their more successful spin-offs and their failures. Spam filtering and self-driving cars are more big data problems (though spam filtering is less impressive nowadays), and the Cloud and YouTube make a lot more sense when you realize that Google already runs a really efficient data-center. Meanwhile, eg, Google Reader was something that didn't really need a Google to run, and suffered for it.
It's not as obvious to me how AWS grew out of Amazon; it doesn't seem obvious to me that you need that much data center to run Amazon (or at least, Amazon-at-that-time).