There was a general sense of "it's just a bookstore".
Now I know everything is more mature and the situation is different, but I also remember feeling very confident that Amazon was waaay overvalued.
Besides, Amazon, Kodak and Boeing were never exactly in the same market. The parent is comparing companies that ostensibly offer similar services.
If I can trust Wikipedia by 1999 Amazon "also sold video games, consumer electronics, home-improvement items, software, games and toys in addition to other items." That meant that it was already an online retail store, not a mere bookstore.
How long before Amazon sounds of its retail business?
Would not surprise me if a majority of both inventory and sales are actually marketplace sellers by now, that pay Amazon for fulfillment.
However, their cloud (and other IT) services do qualify them as a "tech" company.
WebVan is another such example.
Timing was wrong, too early. Chewey sold for 3 billion in late 2017 to petsmart, a less online savvy box store https://www.recode.net/2017/12/6/16681040/ryan-cohen-chewy-r...
The biggest difference I see this time is these unicorns have actual revenue. Heavily subsidized but revenue still. Last time we had monster companies that didn't even show that given 2 dollars they could make 1, so progress?
What made Amazon survive (and achieve valuations way in excess of those in 1999) was not the fact that it had something super valuable then, but that it relentlessly innovated, stayed flexible (an online bookstore offering a computing cloud services, what a weird idea) and executed well enough. And had enough money and controlled costs well enough to reach profitability. For every Amazon there are hundreds of Enrons and WorldComs.
So as long as AMZN was >= 20% of your basket, then it was fairly valued. Sounds about right.
and there's the rub, knowing, a priori, what the proper balance should be.
Yes. This has been extensively studied. Almost every investor who deployed new capital in the late 90s lost money on those investments.
> as long as AMZN was >= 20% of your basket
You’d have to torture causality to come up with a portfolio that would have made sense in the 90s and would have been 20%+ Amazon. It wasn’t even in the top 10 most valuable public companies by market cap [1].
[1] https://en.m.wikipedia.org/wiki/List_of_public_corporations_...
[1] https://data.bls.gov/cgi-bin/cpicalc.pl?cost1=1&year1=199901...
It's also worse case because it assumes you buy at the $107 peak. It didn't spend much time above $100...
It's like saying "All houses built in 1900 are built SOLID because look at all the 100 year old houses that still exist" and then completely ignoring all the 100 year old houses that don't exist...
Hell, look at the companies that were undervalued. GIANTS! Companies that will never fall... like, say, Blackberry and Palm?
'Valuation' these days seems to be less about determining the free cash flow of a company's operations. It's the promise of great innovation that's just around the corner. Or, more likely, it's the promise that a bigger company will acquire this company at a markup above whatever you paid.
I wonder how these profitless companies will fare when the next recession arrives.
It's difficult to decompose a stock price, but I think Amazon gets most of its from AWS.
As the old adage goes -- the pioneers get all the arrows.
It's more accurate to say that venture capital allows Uber to subsidize rides more aggressively. I see no evidence that Uber is substantially more efficient (when you include the driver's costs) than traditional taxi companies.
Edit: Uber has also managed to skirt a bunch of labor laws by claiming their employees are contractors. I guess that's a kind of "efficiency".
At its most efficient, an Uber model can automatically and predictively allocate rides based on realtime rider and driver current and predicted locations with realtime traffic/travel data. There's simply no comparison.
The extent to which rides are cheaper due to skirting of regulations, extracting money from VCs, pushing costs onto drivers, is effectively a separate issue.
That is to say, Uber can break many laws, screw over drivers, flood the market with way too many cars, all while losing money, but none of that means that the dispatch model isn't inherently more efficient.
one, the convenience of the app model.
two, subsidizing rides to flood the market with drivers.
three, offloading risks like employment law and fleet management onto drivers.
None of the stuff you said about traffic routing is currently true nor is it likely to be relevant. Taxi and Uber drivers do the same thing -- show up where and when there are predictable flows of people. They do this using human intelligence. It is unlikely that all the crazy big data analytics is going to do significantly better than this.
Occasionally they work off hours and pick up a few people that need rides.
That's it.
Their only realistic "exit as a world champion" play was to get to self-driving and own a fleet themselves (being able to create economies of scale on vehicle purchase and maintenance that would beat local taxis). That dream ended when they ran over a woman in Tempe.
Their best hope now is to be one of several large commodity players-- modest margins, some limited ability to spread fixed costs like software engineering over large volume.
This seems like a bizarre claim. Precisely when a specific rider will call for an Uber is inherently unpredictable (at least without a lot of surveillance). Meanwhile Uber drivers are not paid to drive from place to place without a rider - so Uber has limited capability to make them do something like head to a high-volume area in anticipation of demand. This is especially true since almost all drivers I talk to now drive for Uber, Lyft, and whoever else at the same time.
The micro-optimizations of dispatching drivers based on what side of the street I'm on or how many left vs. right turns are on the route are nice but only have a meaningful impact on wait time when there are already numerous drivers in the area to choose from (i.e. when the wait time is already short).
Edit: On second glance I realized you may mean that the dispatch is predictive in terms of anticipating traffic conditions and factoring this into the matchmaking between riders and drivers. While this is valuable, the key innovation here is the same kind of predictive routing algorithms that both Google and Apple already provide. Building this yourself may be hard, but is far from novel.
In the US, taxi drivers were always considered contractors, Uber just followed the existing model. In fact, they used to pay to work[1], starting each day in the hole. Only after Uber et all appeared have drivers started unionizing nationally and suing taxi companies to change that.
[1] https://www.nytimes.com/1995/04/09/nyregion/driving-a-taxi-d...
Amazon, despite losing money, had very competitive gross margins from day one. I think pets.com was literally losing money on every product they sold (but don’t quote me), which is not sustainable.
Not to mention when the dotcom bubble popped Amazon's stock price took a beating and didn't really surpass its dotcom bubble peak until fall 2009, 9 years later...
Which, while some of the examples sometimes quoted survived, was generally validated by the 2000 crash.
It's a heavily leveraged company with high capital requirements. AWS makes all of the money.
It’s all relative