It's no coincidence that Amazon calls a unit of storage space in its FCs a "bin" and also happens to pseudo-randomly distribute products in the FC. I am convinced that Jeff Bezos' long term plan is to automate everything, and treat every problem like a CS problem so your CDN analogy is spot on. Pricing is automated, ordering is automated, and I'm pretty sure with Kiva storage positioning is dynamically optimized too. [0]
The thing is, the further invested Amazon is along this path, the cheaper and more profitable it becomes to continue. Moreover, the harder it becomes for new players to catch up.
Will be interesting what the plan is with these planes, I note Amazon also is investigating entering the shipping industry, a space pretty ripe for disruption[1][2]
[0] http://m.youtube.com/watch?v=UtBa9yVZBJM
[1] https://news.ycombinator.com/item?id=10907163
[2] http://www.bloomberg.com/news/articles/2016-02-09/amazon-is-...
You're partially correct in your overall assessment of Amazon, but they aren't a tech company first. They're a logistics company that was forced to become a tech company to support their retail efforts.
http://thenextweb.com/insider/2011/10/05/amazons-cto-amazon-...
I don't suppose other companies call their product classifiers browse nodes and call their most granular classifiers leaf nodes...?
I was actually thinking the etymology of the word bin in CS is probably from someone who worked in a warehouse. It can't be a coincidence.
What I have found more amazing is the selection of books in Amazon "brick-and-mortar" bookstore compared to Barnes & Noble (B&N) bookstore in same Seattle shopping complex previously. I typically visit that shopping complex few times a month. Every time, I have visited Amazon bookstore, I tend to pickup couple of books to browse that I find interesting. When I used to visit B&N in same complex previously, it was rare for me to pickup books to browse in repeat visits. Amazon is able to achieve in 1/4th of the one floor what B&N couldn't do with books on two floors, make me pick up books on repeat visits. Amazon seems to refresh books more often in its store. Amazon might also be using the browsing/purchasing habits of users living in the area to determine what books to stock at the store.
For fiction and any other type of linear reading material though, e-readers have been a godsend.
You'd be amazed how quickly you can find something in a book (or in a few cases for me, a blueprint) when you can just hit F3 or ctrl-f and search for it.
However, I often find myself in situations where I don't know precisely the term to search for and am simply browsing to see if there is some information applicable to the problem I'm attempting to solve. That's where the ability to flip rapidly back and forth shines.
1) the current Apple trend to get rid of all PDFs has made my Mac time a pain given I cannot get good print outs of documentation or use it on something other than an iPad.
I can relate with the sentiment on paper books though. It's quite great being able to quickly flip back and forth to several places, or place notes throughout. I've had luck with the highlighting feature on my Kindle, which gives a decent interface for going back and looking at all the things I've highlighted; I've developed the habit of highlighting all the important parts in a book and going back and re-reading just the highlighted bits, which has helped tremendously. Unfortunately, highlighting across pages is rather aggravating.
Apologies if I seemed flippant -- things are changing so quickly nowawdays that apart from a few core texts on algorithms and programming languages, I've pretty much stopped expanding my library. It seems to be stackexchange and blog posts across the board now.
I suspect I'll pay for this choice sooner or later...
I did buy a copy of The C Programming Language a couple of years back though, but that's just because I feel like it's just one of those books you must have.
It's unlikely that it would be worth doing this if one person looks at a book; the return-and-buy-later rates wouldn't be high enough. But if ten people look at a book, it makes sense to move one or two copies to a closer warehouse -- and that's where Amazon's scale really wins. With O(N) customers and good historical data they can predict future purchases to within O(sqrt(N)) thanks to the central limit theorem.
Yeah... But then again, if a truck is leaving soon from warehouse A to warehouse B, why not put that book on the truck, if no one near A has looked at it recently?
With good data you can make the right small decisions.
source: work in Supply Chain for Amazon
But there a again, I don't collect patterns about my behaviour so it could be true.
Sometimes I look at Amazon and then buy the product elsewhere, say if I spent 3 occasions over a few days looking at a product, does that breach some threshold somewhere for them to say "we have one on the hook, lets start moving this to a warehouse near him just in case"
But again, I know nothing, other than thinking about it a bit when the patent app became known.
I missed a lot in my own childhood and upbringing to have people in my environment that we're or we're in contact with people at the top of their field, having deeper understanding of something instead of just struggling at their level.
Facebook can gauge the difference between different types of likes from how you read a page[1], Amazon does too from their harvested clickstream.
[0] http://www.bbc.co.uk/news/technology-29609472
[1] http://www.slate.com/articles/technology/cover_story/2016/01...
For Prime orders I can see it making a huge difference. Two-day shipping gets really expensive compared to slower methods when shipping over large distances, but for short distances it can be done entirely on normal ground delivery trucks and is cheap. If prepositioning wastes a few bucks sometimes and saves a bunch of money on expensive two-day shipping other times, it could be a big win.
Amazon is probably never shipping a single book, they are using analytics to decide if they should ship a case of books. If they think you are 90% likely to buy a book tomorrow but you are the only one who will it isn't worth the risk. If they decide 100 people are 50% likely to buy a book, sending a box of 100 books is a good investment - if 50 people buy tomorrow that implies 40 will buy the next day and 30 the day after and then 45 the next day and so on for a few weeks/months as sales slowly drop to zero.
Amazon is way ahead in the logistics game, even further than some of your flagship logistics companies.
That sounds cool, but I doubt it actually happens. At least not for one person.
Supposedly, Amazon does tons of backend real world optimization on shipping.
Say you, a SF resident, look at a book today on Amazon.com. Based on your past actions, Amazon knows very well how likely you are to purchase right away, or tomorrow, etc.
Now imagine the book isn't in a warehouse close enough to you for prime, but Amazon's data tells them you are much more likely to buy it when you come back to Amazon tomorrow if Prime is offered.
They'll ship the book to the warehouse in case you come back tomorrow to look at that book. Of course there's tons more modeling behind it to make sure they have very accurate estimates of the cost to pre emptively ship vs the payoff of you buying it, but that's the gist of it.
That was a few years ago, and it blew my mind at the time to think of all the optimization you can do in shipping to individuals when you're working at scale.
Amazon is more and more essentially running a CDN, but for the real world.