I was pretty surprised at the response to my observation on how Amazon had turned in its a rare quarterly profit:
Amazon made a $92 million profit last quarter, or about 19 cents per share. That’s peanuts compared to other tech giants like Google, which netted $3.93 billion last quarter, or eBay, which made $682 million. But analysts expected Amazon to lose 14 cents per share. In fact, it’s unusual that Amazon, a 21-year-old company, actually turns a profit at all. Last year the company lost 27 cents a share in the second quarter. So investors are feeling pretty good about Amazon right now: shares are up 15 percent in after-hours trading.
My observation was that Amazon had made money on cloud services and the reaction was does it matter? The answer is, yes, it matters.
First, a little background. For those of you who aren’t particularly tech-y, Amazon got into the cloud services business back in, I believe, 2002. It began selling the technology it had developed for its own use. It began to sell what has become known as cloud services. The 25 words or less description of that is that you can put your computing infrastructure, your processing power, storage, and so on, in “the cloud”, i.e. the Internet. Instead of having whole buildings full of servers, environmental control, backup power and so on, you just pay a fee to have someone else do it for you. For the purposes of this post I’m not going to go into the technical differences between what’s called “colocation” and true cloud services.
It’s actually quite cheap. Right now you could browse on over to Amazon Web Services and they’d give you the cloud services equivalent of a small server of your own for free for a year. After that it would be just a few bucks a month. It’s no wonder that all of the Fortune 500 companies are using cloud services at some level and some have completely put their information infrastructure “in the cloud”.
Amazon didn’t make its profit in retail sales. That continues to lose money. It makes it in cloud services.
A shrewd manager will devote more resources, particularly management attention, always a scarce commodity, to where he’s making money and, at the very least, try to control his losses in other business segments. She’ll either do less of it or find out why it’s losing money and deal with it.
Amazon has been trying, unsuccessfully, to make money in retail sales for more than 20 years. Amazon’s weird business model moved Matt Yglesias to waggishly characterize the company as a “not for profit operated for the benefit of consumers”.
So I wouldn’t be at all surprise if Amazon gradually de-emphasized its online sales effort in favor of cloud services. But there’s more to the story. Amazon may have been the first mover in cloud services but it’s no longer alone. Amazon has about 29% of the cloud services market and it’s growing fast but Microsoft, IBM, and Google have carved out substantial chunks of that market and are growing even faster. Have you wondered why Microsoft’s software offering have been so lackluster? Maybe it’s because that market segment is languishing while cloud services is growing fast.
At any rate Amazon is actually making money in cloud services and it’s facing increasingly stiff competition. It will want to maintain its position in cloud services but whether it will be able to in the face of increasingly stiff competition is another question.
Meanwhile it’s good news for consumers. Your next PC may be “in the cloud”.
Just …. news for consumers. Saying that “your next computer will be in the cloud” is on a par with “your next sports car will come from Uber!” Gosh wow gee.
The thing of interest about cloud-based computer usage is that it will likely erode consumer choice, leading to mediocrity.
Consider the mid 1980s — a Golden Age of software anarchy. The dominant operating system was DOS, in half a dozen flavors, and a variety of word processors were in use: MS Word, WordPerfect, WordStar, XyWrite, Brief, Emacs, and others, generally in several versions, plus a hodgepodge of clones and text editors. Plus some “office suites” like Lotus Symphony and related applications such as HTML editors and desk publishing utilities. And pretty much, people could make their own choice of editor — as long as it could get output to print on 8-1/2 x 11″ paper, that was commonality enough for most offices and most home users.
But these were very different word processors, appealing to users with different tastes and skills, some quite arcane. You still find devoted users of WordPerfect 5.1 or XyWrite 3.5 or Nota Bene hanging on, running their software on obsolete dust-encrusted PC-ATs or perhaps via DOSemu in Linux or vDOS in Windows or in some virtual machine.
But by the late 1980s, standardization was bringing a heavy hand — nay, a fist — down on this libertine sprawl. In retrospect, the warning sign was when employment ads for secretaries began to stress that employers would consider only applicants experienced in WordPerfect 5.1; typists who had been trained in WP 5.0 just weren’t good enough. About the same time, I remember talking to my brother in law about the contrasting merits of dBase 3 and Paradox, since he was in a mood to buy a database program. As it happened, Microsoft came out with Access just two days later, and the name on the box was all the selling point he — and several million other folks — needed. Database software has pretty much vanished from most users’ PCs these days — Oracle and MySQL and a couple other high powered packages rule the commercial world, but if you run a small office or want a relational database system on your home PC, odds are you’ll use Access or a clone mimicking the last-but-one edition of Access.
Ditto for word processors. On paper, it looks like variety is still alive, WordPerfect is still going, under the Corel label. WordStar and XyWrite are dead, but Nota Bene still exists. We’ve lost Lotus in the Suite category but gained GoBe Productive. And every PC comes with Notepad and Wordpad, and there are a thousand clones of both available for feee downloading, all over the internet.
But the sad reality is, we’re all using Microsoft Word, and we don’t have any choice about it. Oh, some choice — We can run MS Works, an obsolete compilation of DOS era programs, or MS Office 97, or MS Office 2000, or MS Office 2003 or 2007 or 2010 or 2013, or a close such as Open Office or Libre Office. In principle, there are significant differences between these versions. In practice, they all seem about the same, since few users bother to learn more than the basic capabilities of their office suites, We get enough product differentiation from MS that it can justify marketing a new version of Office every three years or so; at the same time, there’s so little differentiation that a majority of office workers could continue doing their work with their eyes shut.
My guess is, this is The Future. Standardized software, with a user interface changing at glacial pace, and a mass of users doing very similar things with very sketchy training, I understand the benefits — no one’s pushing for a return to the Good Old Days when six different power providers in every major city offered their own
varieties of amps and wattages and every household outlet could be Electrifyingly Different. But I suspect our New & Improved Computer Future is going to be about as exciting as washing clothes in a communal laundry.
]]]]]]]]]]]]]]]
It’s not just end users. It’s software developers, too. The latest crop are mostly just skilled manipulators of software development tools. The software development equivalent of using Word or Excel.