Saturday, October 22, 2005

Cringely's Math Makes No Sense

I read Robert Cringely compulsively - he's a lively writer, a deep thinker, and totally plugged in to several frontiers of technology. But every once in a while (as with his July column on Skype) he makes claims that are not just wrong, but ludicrously wrong.

Here's the essence of his last column, which I have condensed to save you the three minutes I wasted reading it:
"A lot of money is being bet on a future user computing experience based on web services" but "the cost of keeping [user] data online all the time will be huge. It's an energy crisis in the making." "202 million [U.S.] Internet users..." * "a free Gmail account with two gigabytes of storage..." = "400 petabytes. That's 400 times the current capacity of the Internet Archive."

"that's really only about $25 million in disk drives..."
+ "total power consumption up to just under 10 megawatts, which at typical U.S. industrial power rates will cost about $5 million per year."

"This is the kind of planning and provisioning required to support FREE services...That's a heck of a lot of ads."

"My point here is that we're entering another period of Internet exuberance...the Internet will change even more than it has the ways we live and work. But it isn't going to come easy and it isn't going to come cheap."

Huh? Cringely just wrote that the big portals can create a kick-ass email service for *every single internet user in America* for a disk drive cost of $25 million, and an annual power cost of $5 million. Generously speaking, let's say that the total capex would be $125 million and the total annual opex would be $25 million (5X on both his numbers). That's a capex per user of 62 cents, and an annual opex per user of eight cents. And he says this is a problem? Dude, that's two postage stamps!

There are three great systems of analog user-created content creation and distribution in place today. They are the U.S. Mail, the phone system, and the camera/photography/film infrastructure. Voice, text, and images, analog-style. To give Cringely as much slack as possible, let's look at the smallest and cheapest, film photograpy. I called a knowledgeable friend, who told me that in the United States, there are about 25,000 minilabs for processing photographic film, and that the average cost of these labs is about $200,000. That's a capex of $5 billion. And that's just for photo prints at your local Walgreens, CVS, or Ritz Camera. If we assume one $10 per hour clerk full time to run each of those minilabs, our annual opex is $500 million. Don't forget that behind those minilabs are film plants, paper plants, distribution networks, and SuperFund sites.

By any wild stretch of imagination, the new digital infrastructure being put in place will be vastly cheaper, both to build and to run, than the old analog infrastructure that it's replacing.

The value of that infrastructure, compared to its cost, is incredible: Cringely carps that $30m is "a heck of a lot of ads" the day before Google announced that they'd just sold $1.05 billion worth of ads in their last quarter alone.

Wake up, Robert, and smell the bits.

No comments: