March 5, 2010 | 2
Launching a successful start-up these days takes more than a good idea—in the digital age a revolutionary quantitative approach can be just as effective as a solid business plan, if not more so. Take Google—it was hardly the first search engine to come along (remember Lycos? AltaVista?). But the company’s PageRank algorithm vaulted it over superficially similar competitors in just a few short years.
The symbiosis between upstart companies and math-minded individuals is a win–win, as Columbia University applied mathematician Chris Wiggins sees it: The start-ups can optimize whatever service they set out to provide, and the mathematicians can make a living without turning to the dark side—the world of finance. Wiggins waxed evangelical on the topic Wednesday at a Columbia meet-up for members of the math community and New York City start-ups, the latter group brought to the table with help from Jonah Peretti, a co-founder of the Huffington Post. It’s all part of Wiggins’s plan, as he put it in describing a new internship program, to "keep the kids off the street … off Wall Street."
At Wednesday’s brainstorming event, representatives from four start-ups described what they do and what they would like to do better. But despite Wiggins’s entreaty that the start-up staffers talk algorithms, not business plans, it often seemed that the two groups were speaking different languages.
During a presentation by Fred Benenson of Kickstarter (launched 2009), which facilitates small-increment funding of creative projects, some audience members wondered in murmurs about what the y-axis represented in Benenson’s graphs. Another person in the audience expressed confusion about the shifting metrics the company seems to use to gauge a project’s success.
And Tom Quisel from OkCupid (2004), a free dating Web site, took only half of Wiggins’s plea to heart. Quisel did talk algorithms but also sprinkled in hearty servings of biz-speak, explaining to the audience he planned to "get you up to speed" so he could then "put you in the driver’s seat."
Wiggins later challenged a team from Simulmedia (2009), which uses data from set-top boxes to match promotional advertisements for TV programs to promising time slots, to quantify the effectiveness of their approach. The team said their method worked better than non-analytical, scattershot placements of promo spots but acknowledged that they did not have enough data to make that claim with statistical significance.
Chief technology officer Mark Angelillo of Snooth (2007) explained that his company is "trying to build the IMDb of wine." Angelillo did win some approving nods by explaining how Bayesian filtering can help Snooth to match messy wine names to the products they describe—how to tell that "2006 Mondavi Reisling Pvt Select (Monterey CA)" means "Robert Mondavi Winery Riesling Private Selection 2006." (Benenson had also invoked Bayes’s theorem.)
Each presenter only spoke for about 30 minutes, so a real dialogue could never quite take shape, and it did not appear that any of the start-ups’ problems had been conclusively solved by the end of the evening. But most brainstorming sessions are generally just starting points, and perhaps some follow-up conversations will bear collaborative fruit. Indeed, Peretti said that such quantitative assistance had made a big difference for his current venture, BuzzFeed. And Quisel, who had announced that his company was hiring, seemed to have no problem drawing a small crowd of potential collaborators/colleagues after the event.
Who knows? Maybe one of the students in attendance will choose love over money, so to speak, and go on to play matchmaker at OkCupid rather than taking an analyst gig at Goldman Sachs. That would no doubt please Wiggins. "I’ve been watching very talented students go price derivatives and have their souls sucked dry," he said. "You can stay on the island [of Manhattan] without inventing weapons of financial mass destruction."
Photo credit: ©iStockphoto/mddphoto
Get 6 bi-monthly digital issues
+ 1yr of archive access for just $9.99