A bit OT:
For Start-Ups That Aim at Giants, Sorting the Data Cloud Is the Next Big Thing
By MALIA WOLLAN
New York Times
December 25, 2011
SAN FRANCISCO — The idea of big data goes something like this: In a world of ever-increasing digital connectivity, ever larger mountains of data are produced by our cellphones, computers, digital cameras, RFID readers, smart meters and GPS devices. The huge quantity of data becomes unwieldy and difficult for companies and governments to manage and understand.
“My smartphone produces a huge amount of data, my car produces ridiculous amounts of really valuable data, my house is throwing off data, everything is making data,” said Erik Swan, 47, co-founder of Splunk, a San Francisco-based start-up whose software indexes vast quantities of machine-generated data into searchable links. Companies search those links, as one searches Google, to analyze customer behavior in real time.
Splunk is among a crop of enterprise software start-up companies that analyze big data and are establishing themselves in territory long controlled by giant business-technology vendors like Oracle and I.B.M.
Founded in 2004, before the term “big data” had worked its way into the vocabulary of Silicon Valley, Splunk now has some 3,200 customers in more than 75 countries, including more than half the Fortune 100 companies.
Customers include the online gaming company Zynga, the maker of FarmVille and Mafia Wars, which uses the software monitor game function to determine where players get stuck or quit playing, allowing Zynga to tweak games in real time to retain players.
Macy’s uses Splunk’s software to observe its Web traffic in order to avoid costly down times, particularly during peak holiday shopping. Edmunds, an automotive research Web site, started using Splunk to troubleshoot its information technology infrastructure and now uses the software to analyze all its customers’ online actions. Hundreds of government agencies use Splunk to monitor suspicious activity on secure sites, and a Japanese tsunami relief organization used it to track aid and monitor road and weather conditions.
The amount of data being generated globally increases by 40 percent a year, according to the McKinsey Global Institute, the consulting firm’s research arm. And while Splunk has a lead in selling software to analyze machine data, big data is big enough to create new opportunities for a multitude of start-ups, many of them using the open-source software Hadoop.
“ Venture capital is absolutely foaming at the mouth over big data,” said Peter Goldmacher, an analyst and managing director at Cowen & Company. “The volume of data being created now is not 10 times bigger, it is like a thousand times bigger.”
While skyrocketing valuations for social networking sites like Twitter, LinkedIn and Facebook have kept Silicon Valley investors betting heavily on the next social start-up, investors are increasingly looking at companies that build software for other companies. Worldwide revenue from enterprise software reached $244 billion in 2010, according to the research firm Gartner. Splunk is seen by some investors as proof that a wily start-up can chip away at some of that market.
“For a while there, people felt like everything that needed to be solved had been solved and that big companies would inevitably find all of the white space in enterprise,” said David Hornik, an investor at August Capital, which invested $3 million in Splunk in 2004. “Splunk is really the poster child for thinking differently about an enterprise challenge and creating a platform that ends up really being disruptive and valuable.” The start-up got a total of $40 million in venture capital at that time from August Capital, Ignition Partners, JK&B Capital and Sevin Rosen Funds.
From the start, Splunk’s founders — Mr. Swan and Rob Das, 52, who is the company’s chief architect — set out to shake up what they saw as the stodgy, top-down world of enterprise software. “Big software is sold on the golf course, not sold to the people who actually use it,” said Mr. Das. Instead of aiming at the golf-playing chief information officer, the company took a quirky name that sounded like “spelunking” and zeroed in on the culture and tastes of everyday I.T. employees, the ones who actually had to use, and program around, enterprise software.
In 2005, when Splunk unveiled the first version of its software at the LinuxWorld conference in San Francisco, its booth was in an obscure corner, hidden by “rows and rows of vendors plastered with stock art of guys in suits and ties,” remembered Mr. Das. Nothing about enterprise software seemed hip or even vaguely playful, said Mr. Das, who spent more than a decade working in I.T. at companies like Lotus and Sun Microsystems. “We wanted to make enterprise software cool again.” So they decorated Splunk’s booth in all black and gave away T-shirts that said, “Take the SH out of IT.”
“People were stacked up 10 deep,” said Mr. Swan. Everyone, it seemed, wanted a T-shirt.
“Our customers, especially at the start, were I.T. people,” said Mr. Swan, who had worked at Apple and Disney Online, before becoming a co-founder of Splunk. “We’re talking about the guys in the basement, the guys in kilts and Mohawks. Those are our people.”
The company says it has been profitable for two years, and though executives will not comment on its exact plans to go public, Mr. Swan says, “We will be the first one to get shot out of this big data thing like LinkedIn got shot out of the social media space first.”
In another sign of an impending initial public offering, in 2008, the company hired Godfrey Sullivan, formerly of enterprise software companies like Hyperion, as its chief executive.
“There is a lot of money chasing this new world of unstructured data,” said Mr. Sullivan. “I would call Splunk the first mover in big data because we have been at this for years now.”
nytimes.com |
