For years, I have seen the same take over and over about Amazon.
First it was “Amazon doesn’t make money.”
Then it was “Amazon doesn’t make money without AWS”.
Now we have a new version: “Amazon doesn’t make money without AWS and advertising.”
Say it with me now…
Amazon’s retail business is profitable and very much so.
They use the branded 1st party retail sales as a loss leader to print money off of their marketplace which is, unlike retail, riskless profit.
I know because I’ve looked at the unit economics of their business for years.
So then why does their business look unprofitable after separating out AWS and advertising?
First reason is they’re aggressively depreciating every plane, truck, and warehouse they buy.
Second reason is this crap. How much of it makes money? It’s all in “retail”.
Amazon’s marketplace is actually their highest moat business. It’s a better business than AWS which has less market share significantly more competition and stronger competition.
I suspect they understand this and deliberately conceal it but who knows?
/end rant
• • •
Missing some Tweet in this thread? You can try to
force a refresh
One of the dudes I play bball with is an engineer for ercot, the texas agency that is often blamed for the Feb 2021 freeze.
His explanation of last years freeze:
1. Nat gas power plants were not built for such cold temperatures and they broke, not the green wind turbines
1/n
2. There just was not enough power generation. Ercot’s job is to best use the power they are given by the power plants and there was not enough so they had to rotate it between the residential areas.
3. Depending on whether or not the city’s residential grids were shared with critical infrastructure that never loses power eg hosptials, fire, and other city stuff, their experiences were different
I college, I was taught and convinced that data analysis could do anything and that if you weren’t using it, it was because you weren’t smart enough to.
A decade later, I would say that besides throwing the datapoints up on a graph, it’s almost always useless.
1/n
The world is too complex to gain understanding from a regression or other complicated mathematical formula.
Even randomized control trials, which people who operate in the real world almost never have access to, are often bullshit.
Today, I think data analysis is mostly how people, namely academics and some professions, choose to communicate.
It’s sometimes useful, but also signaling to show that you’re part of a group.
If you want to be heard in that group, you must speak their language.
1. Not really because 9 times out of 10, a new bolt on company is a new supplier. Most products where you’re getting a benefit from container scale are already there when they sell to thrasio. Combining two factories products is…
inefficient and a headache. Like I said, weak returns to scale. The other problem is timing. You’ve got to have two shipments ready at the same time and they’ve got to fit together.
2. 3pl savings are very minimal because of huge assortment of product (headaches) and inability
to negotiate better pricing with FedEx ans ups because you’re using Amazon.
3. By and large the so called “mom and pop” business owner grinds way harder than the thrasio 9-to-5er and is probably more knowledgeable and smarter. Plus it’s easier for the mom-and-pop to blackhat