Is it just a coincidence that Google dumped the “Froogle” name for its shopping service immediately after shelling out a wallet-whomping $3.1 billion in cash for DoubleClick? As Marketwatch’s John Shinal notes, Sergey Brin and Larry Page’s veneration of Warren Buffet no longer extends to copying the Oracle of Omaha’s philosophy of buying companies and other assets on the cheap. Larry and Sergey are turning into Buffet’s prodigal sons. Froogle, they ain’t.
But there is one place where Google continues to display a Buffetian skin-flintiness: in the buildout of the company’s data center network. And that brings us back to Omaha. Mounting evidence suggests that the Omaha-Council Bluffs area, on the Nebraska-Iowa border, will be the site of yet another of Google’s mammoth server farms. According to a report in yesterday’s Des Moines Register:
Two Iowa lawmakers say a large Internet company is weighing whether to locate part of its operation near Council Bluffs – and the House is moving ahead with incentives for that company. “The best information we have: This could be at least a $600 million investment with 100 jobs at $50,000 to $130,000 per job or as much as 10 times that number,” said Rep. Phil Wise.
“$600 million investment” is the new code for “Dude, we’re getting a Google data center!” Google’s Lenoir, North Carolina, center, announced earlier this year involves a “$600 million investment.” Its Goose Creek, South Carolina, center, announced a few weeks ago, involves a “$600 million investment.” There’s also a “mystery plant” being built on an 800-acre site in Pryor, Oklahoma, that has Googlish fingerprints all over it. When it’s announced, I’m guessing it’ll involve a “$600 million investment.” (And let’s not forget the possible Google data center in Blythewood, South Carolina, with an estimated price tag of “between $200 million and $800 million.”)
Add up all those $600 millions, and you’re talking a whole lot of money. But if you look at the sites, you see a very frugal approach at work. The common threads are cheap land, cheap electricity, cheap labor, cheap water, cheap bandwidth, and rich tax breaks. An article in the Omaha World Herald describes some of the attractions of the region for a Google plant, and the list applies equally well to the other sites:
The Omaha-Council Bluffs metropolitan area would give a company like Google access to relatively inexpensive land and labor compared with other parts of the country … The metro area has a large network of fiber-optic cables, which came with Offutt Air Force Base … The potential Bluffs development is near several electrical grids from the Mid-American Energy plant in southern Council Bluffs. David Sokol, chairman and CEO of MidAmerican, said he doesn’t know what company may be coming, but he did say the Lake Manawa area has plenty of electrical redundancy … The industrial foundation has approached the Bluffs water works about dramatically increasing the amount of water available for the site. Water can be used as part of “chilled air” systems that keep computer servers and other heat-generating equipment cool.
Shinal argues, with good reason, that the purchase of DoubleClick was essentially a defensive move for Google. When Google builds data centers, in contrast, it’s playing offense. That hints at the strategy underlying Google’s frugal-prodigal bipolarism: Spend like drunken sailors when necessary to block competitors, but pinch pennies when building what the company hopes will be the foundation of its dominance in the long run. Prodigal frugality? Frugal prodigality? Let’s just call it Google Proogle.
Spend like drunken sailors when necessary to block competitors, but pinch pennies when building what the company hopes will be the foundation of its dominance in the long run.
The words ‘short-term’ and ‘long-term’ come to mind, as indeed do the words ‘one-off’ and ‘recurring’. I don’t think this is one of life’s great mysteries, in other words.
I was more interested in this:
The common threads are cheap land, cheap electricity, cheap labor, cheap water, cheap bandwidth, and rich tax breaks.
I’m not too bothered about items 1, 2, 4, 5 and 6; if state X wants to spend its tax revenue subsidising a multi-million dollar corporation, that’s up to its legislators (and their voters). But, um, cheap labour? As in, people working long hours for low wages? I’m sure that’s not what they said Web 2.0 was all about…
Things are cheap when their is competition, companies with a large market sahre aren’t sheep becasue they are unique. Bargaining power at its simplest, I would say.
Phil, if anything, Web 2.0 is about cheap crowdsourcing, and well-paid innovators; what Google is getting by going to the heart of the land is cheaper labour because a coffee in Arkansas is less expensive then a frappucino in LA. What really matters though is the capital expenditure compared to the number of employments: Google remains a investment company.
A critical component of google’s dominance is its engineering and management talent. Google has elaborate interview process and hires/retains top-notch talent at much higher than average rate in some of the most expensive engineering labor markets. So, I think the rationale behind google’s penny-pinching not so much offensive vs. defensive move as good-old maximizing of cost/benefit equation across the board.
I’m struck by the water aspect because I’ve been looking a lot lately at water wars and global warming effects.
I imagine those guys are thinkng long term on this but, if they’re not, they better get started!
Clyde, see Google.org.
Re “cheap labor”:
The 200 jobs Google is creating in Lenoir, North Carolina will average about $48,000 a year. Early reports about the Council Bluffs project indicate that ‘Company X’ has told officials that the average salary would be $60K.
The labor piece is the key to making the economic incentives work. Data centers don’t bring a lot of jobs due to the level of automation, but offer a better salary range than many other industries . And once one major data center gets established, others follow. This “clustering” effect has been more pronounced with Microsoft sites, first in Quincy, Wash. and now in San Antonio.
Surely most Google queries must be answerable from a dataset of less than 2TB or so. Most people switched to Google because it returned less of the web than AltaVista! Most of Google’s value-add is in spam filtering the web, not indexing it. Why do they need these huge centres?
Speed.
(And of course they’re not just doing search anymore.)
Thanks, Michael.
Not to be cynical or anything but doing that kind of philanthropic development must provide amazing visibility into global resources.
Smart.