Business at Cyberspeed
By Leslie Walker
For Gary Culliss, the hardest part was telling his mother.
After earning his law degree from Harvard last June, Culliss called home to Kansas to break the news: Instead of practicing law, he was starting an Internet company.
His mother was not pleased. "She wasn't sure if I was opening a doughnut shop or something," recalled Culliss, 28, who had spent years preparing to be a patent attorney. "She doesn't really understand the Internet. But she's caught on now and has been supportive."
How Culliss ended up scrapping his law career and starting a company called Direct Hit Technologies Inc. speaks to the heart of the Internet, which beats so fast and so far outside the traditional world of business that it allows young people with clever ideas to turn their dreams into multimillion-dollar companies almost overnight.
Within three months of his graduation, Culliss had founded a company dedicated to making Internet searching easier, raised $1.4 million in venture capital and landed contracts with America Online Inc., Apple Computer Inc. and Wired Digital Inc. Without stopping for fancy planning and the usual business school stuff, Direct Hit raised another $2 million in the fall, hired 30 people, signed up more high-profile customers, released a string of new search features and last week moved into 9,000-square-foot digs near Boston.
The founding of Direct Hit illustrates not only how the global computer network is speeding up the world of business, but also how the Internet is keeping its cutting edge sharp: by embracing original thinkers in their twenties. Dozens of youngsters like Culliss are able to walk through the doors of big companies and bring fresh clarity to problems that have stumped their elders (often people in their thirties), and have been handed millions of dollars for their effort.
Culliss's vision for a better Internet began in a patent records room in Florida, where he conducted searches for inventors before he went to Harvard. Culliss found he got better results by asking colleagues what they had learned from doing similar patent research in the past.
When he began searching the World Wide Web at Harvard, Culliss experienced deja vu looking at the thousands of irrelevant search results -- like typing "Armageddon" into an online search engine and retrieving thousands of documents about the end of the world, instead of about the movie. Why not train a computer, he wondered, to do what he had done with the patents -- ask other searchers what they had learned? Wouldn't that harness the Web's essence -- its network of millions of users -- to make it a smarter medium?
Culliss's "popularity engine" (www.directhit.com) is a software program that ranks Internet search results by calculating how popular each Web site has been with other people who ran similar searches. It weighs things like how long people spend viewing each site they select.
Early last year, Culliss wrote a plan offering to license his program to the Internet's leading search companies, then entered it in a competition at the Massachusetts Institute of Technology. Direct Hit won first place in MIT's prestigious $50K Entrepreneurship Competition a few weeks before Culliss got his law degree.
The Internet industry's rapid embrace of Culliss's idea reflects how the Web's evolution is driven by the people who use it, often working alone or with a few friends. There was Marc Andreessen sitting in his Illinois dorm room in 1993 designing the first "browser" that displayed graphical images on the Web. There were Jerry Yang and David Filo at Stanford University in 1994 organizing a directory of Web documents they called Yahoo! And now Gary Culliss at Harvard.
His is not the only smart-searching idea to emerge from the dorms in the past year. Another, called Google Inc., sprang from the minds of Stanford grad students, who rank search results by the number of other Web sites that link to each site -- like the time-honored citation analysis of scholarly research.
"The reason a lot of these ideas come from the guys outside the big companies is they are unencumbered by legacy," suggests Warren Packard, a director at Draper Fisher Jurvetson, the Silicon Valley venture capital firm that backed Direct Hit. "Once these companies are built, people tend to get fixed on their own companies, and then it's hard to come up with another idea."
Culliss didn't bring his idea to market alone. He had help from -- who else? -- another young man with a plan. Culliss was smart enough to seek a chief executive with real-world business experience.
Mike Cassidy, at 35 an older version of Culliss, was on a hunt of his own that had followed a similar path. He had attended Harvard's business school and while there placed first in the same MIT contest that Culliss later won. The victory bestowed chief executive status on Cassidy at age 22 by attracting enough venture capital for him to start a technology company. Cassidy sold his Stylus Innovation Inc. in 1996 for $13 million. The success financed a year of relaxation at the Berklee College of Music in Boston, where he played jazz piano.
"Then I got restless," Cassidy recalls, talking fast from his cell phone as he drove up a hill in Silicon Valley to meet with potential business partners last week. "It kind of gets in your blood. I wanted to do it again."
So last March, Cassidy went back to the scene of his first big score: MIT. How he found Culliss shows the Web's growing role in shaping business careers: "I logged onto the MIT-50 competition Web site and looked over the 82 entrants. I saw two I thought were great ideas. One was Gary's Direct Hit. So I sent him an e-mail and said, 'I think you've got a great idea; I'd like to meet you.' Two days later we met in Harvard Square. We ended up spending four hours together, and he offered me the job right there as CEO of the company. I accepted."
Since then, Direct Hit has scored almost a hit a week, either a new contract or feature for its search technology. Its popularity engine is deployed or will be soon at HotBot, LookSmart, America Online's ICQ.com and Apple Computer's Sherlock search engine, among others.
Cassidy and Culliss achieved their success partly because they were mindful that cyberspace business moves at the speed of e-mail; they wasted no time.
First was the money. Tapping an old colleague, Cassidy landed an appointment at Draper Fisher Jurvetson. "We presented to the venture capital firm at 8:15 in morning," Cassidy recalled. "At 4:30 the same day, they said, 'Fine. here is your money.' "
The next day, Cassidy flew back to Boston and picked up the phone, called several old friends and said, 'Quit your job. I got money; we are starting in two weeks.' "
Packard said he and his partners at Draper see 100 business plans a week, but few as original as Direct Hit's. "It was so compelling to us. We have never given a term sheet so fast."
The money men quickly saw how the popularity ranking of Internet search results could be extended beyond helping consumers find knowledge faster by elevating relevant documents in search returns. It could also, for instance, help online shoppers identify the most popular products and services -- something Direct Hit already is working on.
This week, Direct Hit (which has four patents pending) will start promoting its new personalized version that customizes popularity rankings for users who provide age, gender, address and other information. A woman searching for shoes, for example, might see Web sites selling women's shoes move higher in her returns.
Direct Hit's backers think most of the big search portals like Yahoo eventually will license the technology.
So why didn't the established search companies think of it themselves?
Culliss said he thinks it is partly because the search engines, after they were launched on the Web and became publicly traded companies, concerned about ad sales and profits, "got distracted from the goal of the search engine, which is to provide really relevant information. They started becoming portals, adding things like free e-mail and chat."
Cassidy agreed. "But it's also a question of how everyone tends to think based on the mind-set they started from," he added.
When the Web's first search engines were created, there were few users on the Internet. It wasn't until the Web popularized the network, and millions of people began adding documents and searching them, that anyone thought much about how computers could collect and analyze data on searching habits.
So while the popularity engine now seems obvious, it was the Internet that made it possible. And in a way, the popularity-ranking software magnifies the Internet's ability to boost human recall by creating a kind of collective memory about human searching -- in this case, data on what people learned each time they searched.
"If only we had been able to trace the searches of the millions of people who went to libraries over the years and looked up things," Packard said. "Now the Internet makes that possible."
Gary Culliss will be online to answer questions from readers of washingtonpost.com in a moderated discussion on Thursday from 1 to 2 p.m.
© Copyright 1999 The Washington Post Company