Story Play

[RULiquid]

Google Secrets
Google doesn't broadcast its strategic plans. Here are eight critical areas where its decisions will shape the industry.
By J. Nicholas Hoover, InformationWeek, Jan. 30, 2006

When Microsoft or Oracle barrel into a market, they serve up a grand strategy and a product road map to get you there. With Google, you might get a beta product that shows up on the Web, sometimes one that's less advanced than similar offerings in the market, and often with a narrowly defined goal, even if the potential for grand expansion seems obvious. CEO Eric Schmidt has said the company doesn't have a sweeping strategic vision, that "we delight in not having such strategy." Not no strategy, but one to innovate in many interesting areas and not to build just one thing.

What follows aren't the kind of secrets we dug through Google's trash to get. Google even talked a little about many of them. These are unknowns in the sense that the typically reticent Google hasn't told the world just how it intends to move down these roads. Here are some insights.

How big in the offline world?

Google is getting into something new all the time: Google Earth, a partnership with NASA, Google Pack software, voice over IP, scanning the world's books. But when it comes to revenue, the company's business can be understood on three lines: 56% from online ads running on its sites, 43% from other Web sites and magazines where Google places ads for a share of the revenue, and 1% other. Its dependence on online ads is one of the reasons Google paid $1 billion for 5% of America Online--to protect the ad revenue that partner AOL generates for Google and keep that from going to Microsoft.

Understandably, the company wants to diversify, and this month's acquisition of dMarc Broadcasting shows one way it intends to do so: by placing ads in conventional broadcast and print media. DMarc does business the way Google likes it--using software that helps automate the buying and placing of radio ads. Google plans to integrate that with its AdWords platform for placing Internet ads. How much does it like this business? It paid $102 million in cash, but the price tag could rise to $1.1 billion if the business hits all its targets the next three years. That's a lot of upside.

Google also is experimenting with print, placing ads for customers through a limited trial in PC Magazine, Maximum PC, Budget Living, and the Chicago Sun-Times. One advantage from such partnerships may be to access partners' local ad sales forces.

Google is now focused on the Internet, radio, and print media, says Patrick Keane, head of advertising sales strategy. But his description of the company's vision--delivering "accountability, efficiency, relevance, and scale to advertisers"--isn't so limited. With Google's recent decision to start selling video content such as CBS TV programs and pro basketball games, it's all but certain the company will move into enabling video advertising as well.

What about further afield, like telemarketing and direct mail? If Google goes there, it would have to do so more cautiously. The company's instant-messaging and Internet telephony application, Google Talk, is free and ad-free, but it's doubtful both of these conditions will continue indefinitely. Telemarketing over IP may not sound appealing, but marketers have yet to meet a medium they won't try.

The caveat here is that brokering the world's information presents as many pitfalls as opportunities. Given the wealth of information that Google collects about its users, figuring out what not to sell will be one of Google's most vexing challenges. --Thomas Claburn

If not a PC, then what?

Google co-founder Larry Page's speech at the Consumer Electronics Show in Las Vegas this month couldn't help but be a letdown. Page made significant announcements, detailing Google's plans to sell video online and distribute a consumer software bundle called Google Pack. But with the rumor mill speculating about a "Google PC" that would be sold in partnership with Wal-Mart (which Google denied all along), nothing short of a major hardware unveiling was going to be enough.

Network computer maker Wyse Technology says it has discussed the idea with Google, though the device they mulled isn't quite a PC capable of running an operating system like Windows. Wyse CEO John Kish says the companies talked about a $200 Google-branded machine that would likely be marketed in China and India through local telecom companies. Specula- tion abounds, with the biggest question being whether Google would have any interest in building some kind of consumer device.

And the answer is, almost certainly. A look at its existing hardware business suggests why.

The company in 2002 introduced its Search Appliance, a rack-mountable chassis that offers Google's search technology for businesses. A few thousand companies have bought this box, priced at $30,000 and up, or its smaller cousin, the Google Mini. But Google personnel who discuss the company's enterprise goals suggest the appliances are something of benign Trojan horses, a way to get an invitation inside the corporate firewall, typically a hostile environment for advertisers. Provide business-search technology and hope to have a future platform for more Google services and ad access to people--employees--who have jobs and money.

For Google, hardware is a doorway to the network. It doesn't make sense for the company to sell a conventional PC in established markets. But where doors to the network are scarce--in China and India, or just maybe in the multimedia living rooms of developed markets--it makes sense for Google to help supply them. Google's goal "to organize the world's information" is often cited, but without the ending "and make it universally accessible and useful." If it takes building hardware to do that, don't discount Google. --T.C.

Is there an Achilles' heel in search?

Search-related advertising is the fastest-growing segment of the online advertising business, forecast to grow 41% this year in a report from Piper Jaffray. And Google is king of that market. In December, it had 16.5 trillion sponsored link impressions, compared with Yahoo's 9 trillion, Nielsen/NetRatings says. In 2005, Google grabbed about 70% of all paid search advertising, according to researcher eMarketer.

Google must broaden its horizons. Although Yahoo also gets the bulk of its revenue from advertising--85%--those dollars are split 50-50 between search and display or "branding" ads, notes Martin Pyykkonen, an analyst with research firm Hoefer & Arnett. It had 28% of the $12.9 billion pie for all online advertising in 2005, eMarketer says.

"Google needs to make that leap into display advertising," says Charlene Li, a Forrester Research analyst. "Right now, their advertising base is really comprised of direct marketers, not traditional brand marketers, and they really don't have the expertise and credibility that Microsoft and Yahoo and AOL have in that area."

It's by no means clear that Yahoo and Microsoft, the next-biggest search-ad rivals, can catch Google, but they're continuing to invest in doing so. And don't dismiss Ask Jeeves, now that it's owned by IAC/InterActiveCorp, warns Scott Kessler, an analyst at Standard & Poor's. "I see a big, aggressive, and well- capitalized company. You have to take [CEO] Barry Diller seriously. At Fox, he was successful when people thought there was only room for three networks," Kessler says.

Google has made the most of its opportunity. Now it needs more. "What people underestimated about Google was how much growth was going to be in search-engine advertising," says Mark Mahaney, a Citigroup analyst. "But at some point, people will want to see signs of traction in other business areas." (For more, see "Google's Achilles' Heel".)--Alice LaPlante

What's next in desktop software?

Google has had client-side software for a few years, starting with the Google Toolbar in 2000. But with the Web 2.0 under development, the Internet has become a software-delivery platform, and Bill Gates is shifting Microsoft's strategy to make a set of applications, called Office Live, more Internet--and Internet-ad--friendly. Will Google respond with a full-blown set of productivity apps such as word processing?

First, it's instructive to understand each company's approach. Microsoft moves into markets with a public proclamation, a road map, and a platoon of developers. Google offers the story of co-founder Page's frustration that a new home computer didn't come with enough easy-to-use applications, and its answer of the humble Google Pack, unveiled at the Consumer Electronics Show. It contains Google software, including Desktop Search and the Google Talk instant-messaging client, plus Norton AntiVirus, Ad-Aware anti-spyware, the FireFox browser, and the Adobe reader.

Google hasn't spelled out how Google Pack will evolve, other than to say it will center on organizing information and making it more accessible. "To help our end users that way, there are instances in which it makes sense to write a piece of client software, and whenever it makes sense, we'll do that," says Sundar Pichai, the company's project manager for client software.

Think of Google Pack as an access road to people's desktops. It will get people used to getting software, not just information, from Google. "It was more important that they got the installer and updater on your desktop than anything else in there," Gartner analyst Allen Weiner says. "In the Web 2.0 world, Google can position itself to be the delivery mechanism for software as a service. They could send you productivity apps that update on a regular basis."

Microsoft's "Live" ambition includes using its apps to cut into Google's market by offering targeted online ads. That's too big a threat for Google to ignore. --J. Nicholas Hoover

Your next phone company?

Voice over IP is an example of how Google enters markets with technology, rather than a strategy statement. Last year it started offering voice service through Google Talk, its instant-messaging platform, as a free beta service without ads. This month it made Google Talk compatible with several other services through open standards, but it didn't spell out a business model. "We're working hard to add features and make improvements," its Web site says. "We're just not quite ready yet to reveal the other cool things we've got planned."

Google has shown other interests in telecom. It started a minor firestorm early last year when one of its job listings sought someone with telecom experience who could buy dark fiber for the company. Then came a flurry of wireless initiatives: It began offering Wi-Fi in its Mountain View, Calif., hometown; it released a Wi-Fi VPN client called Google Secure Access; and it submitted a proposal to give all of San Francisco's residents wireless Internet access. Then Google joined with other companies to invest about $100 million in Current Communications Group, which provides broadband over power lines. The company characterized all this activity not as a passion for selling Internet access but for getting people onto the Internet, describing it with almost missionary zeal in a Wi-Fi proposal last fall: "... ubiquitous, affordable Internet access is a crucial aspect of humanity's social and economic development, and working to supply free Wi-Fi is a major step in that direction." Oh yeah, people can't be Google customers without Web access.

The company is looking to bolster its VoIP offering. One executive, who asked not to be identified, says he recently met with the company to discuss technology that could advance its VoIP services, and he says he's only one of many doing so. Among the obvious potentials for Google Talk: ad-supported voice-calling services, integrating phone number and Web searches into the interface and allowing third-party plug-ins.

Google may have a different agenda from VoIP vendors such as Skype. The company is using standards from the Jabber Software Foundation's open-source messaging initiatives that aim to make messaging clients more interoperable. That could help integrate voice with other communication methods and make voice a much bigger part of the Web, thus giving people another path to Google's door. "It's not even necessary that Google dominates," says Chris DiBona, Google's open-source program manager. No, it's just necessary for communication to thrive for Google to survive. --J.N.H.

What's under the hood?

Resisting the federal government's subpoena for search data won Google praise from privacy advocates. But that position reflects keen self-interest: guarding trade secrets about how it runs the world's best-known information factory.

Google's secrecy extends to its hardware. The company no longer volunteers details about its IT infrastructure, something IT managers would love to learn more about. Google used to be more open. In a 2003 paper, "Web Search For A Planet: The Google Cluster Architecture," Google engineers Luiz Andre Barroso, Jeffrey Dean, and Urs Holzle described the machines that answer all those search queries. Then, Google managed more than 15,000 custom-designed, commodity-class x86 servers. Today, it still runs Linux-based servers, but Google won't say how many servers and data centers it has.

Much of Google's operating infrastructure is custom made or highly customized. But one challenge it likely can't solve with its own ingenuity is heat. In 2003, its server racks drew 400 to 700 watts per square foot, considerably more than the power density of 70 to 150 watts per square foot typically found in commercial data centers. That's why Google developed and patented a special cooling baffle for its servers and why, in 2005, Barroso cautioned in a paper published by the Association of Computing Machinery, "If performance per watt is to remain constant over the next few years, power costs could easily overtake hardware costs, possibly by a large margin."

Barroso argued that chip multiprocessor technology represents the best and perhaps only way to improve the performance-per-watt ratio. With the release of energy-efficient multicore chips such as Sun's Ultra- Sparc T1, Google's appetite for commodity hardware may diminish as it opts for pricier silicon that delivers better performance per watt. The company's strategic relationship with Sun, revealed last fall, no doubt plays into this calculation.

The advantage Google gains from its customized hardware and software is performance, both in terms of the efficiency of its applications and its personnel. In a business dependent second to second on people, servers, and power, the data-center infrastructure--which some companies relegate to commodity--is the start of everything. --T.C.

Is a megadeal in the cards?

With Google sitting on more than $7.6 billion in cash and securities and about a $128 billion market capitalization, it's got the currency for a big, splashy acquisition. But Google already must integrate the 10 people a day it hires, and, at a company that values culture, mega-acquisitions could become a megadistraction. Then there's the history of such deals: AOL-Time Warner looms over any Internet match-made-in-heaven thinking.

Google hasn't been entirely on the sidelines, but it has limited acquisitions mostly to small companies that round out its technology portfolio or stretch its business model. There's the $102 million earlier this month for radio advertising firm dMarc. Back in March it bought Urchin Software, a maker of Web-analytics software that Google offers free as Google Analytics. In October 2004 it bought Keyhole, a digital-mapping startup it integrated with Google Maps six months later.

Google looks ready to deal if its market position is threatened. When it put up $1 billion for 5% of AOL in December, the deal helped expand the companies' online advertising alliance but also kept Microsoft from moving in. Google may not have the look of a serial acquirer, but it will likely do more and larger deals. "They still don't have stickiness," contends Ian Campbell, CEO of research firm Nucleus Research, citing the hit-and-run nature of search.
What might a blockbuster deal look like? A content-delivery company in another medium, such as satellite radio or cable TV? A VoIP or other tech-based service provider? Or with Disney buying Pixar, giving Apple CEO Steve Jobs new entry to a content provider, does Google need to buy closer ties to a content provider?

Google shouldn't bank on out-innovating everyone, Campbell says. "Microsoft has yet to take advantage and give Google a run for its money," he says. But Microsoft is an example of a company that kept its culture, made tactical acquisitions, enjoyed runaway success, and grew without being tempted into the megadeal. --Tony Kontzer

What does 'Don't be evil' look like in China?

Google's approach to China looks a lot less secretive after last week. The company said it will follow government rules to censor certain information, in return for being allowed to offer a Google.cn Web search site in the country. And it won't offer blog or E-mail services in China, fearing the government would ask it to turn over personal information. Yahoo and Microsoft offer those services, and the government has forced them into embarrassing disclosures.

But last week's decision only starts to answer one of Google's toughest questions: How will Google translate its mission and values to fit in foreign markets?

Google is catching flak for cooperating with the government. "The tricky question is, what's the gray area between obey or stay out?" says John Palfrey, executive director for Harvard's Berkman Center for Internet and Society. Governments lay down standards, like "no subversive content," and leave so much to interpretation that companies err on the side of censoring, so as not to risk being shut down, he says. So, the real test is how close Google or others come to crossing the line.

Microsoft caught heat recently for taking down a Weblog posted on MSN China. But at least that blogger had the chance to offend.

It raises the question of whether the leading search engines will differentiate themselves based on their information policies in totalitarian regimes. "It will tell you a lot about these companies," Palfrey says.

The world needs U.S. companies mixing it up in the gray areas of restricted regimes and emerging markets. No company is perfect, but Google has done a whole lot right to earn the Web community's trust. China needs the Google it can get, for now. Seems like we'd be better off having Google slogging knee deep in the gray areas, risking getting its pants muddied, rather than sitting it out and keeping pristine clean. --Chris Murphy

http://www.informationweek.com/internet/showArticle.jhtml?articleID=177104744