Book of Odds Comes Out of Stealth to Make Intuitive Sense of Statistics—But Can It Sell Ads?

10/14/09Follow @wroush

Despite the fact that “Google it” is now synonymous with “look it up,” Google itself isn’t a primary reference source. It merely aggregates facts and claims from everywhere else. If you wanted to create a genuinely new reference work on the scale of a dictionary or an encyclopedia—even if it lived online—you’d still have to do it the old-fashioned way: by hiring lots of people to do lots of research, writing, and fact-checking.

And that’s exactly what Book of Odds, a startup in downtown Boston, has done. After a remarkable three years in deep-stealth mode, the company is going public today with its website, a compendium of hundreds of thousands of fascinating factoids about the statistics of everyday life, collected and collated by scores of university students from around Boston. Ever wondered what fraction of workplace deaths are caused by flying objects? It’s roughly 1 in 96. And as it happens, 1 in 96 is also the chance that a white woman between 65 and 84 will die of cardiovascular disease in a given year, and the odds that a U.S. male aged 16 or older works in the legal profession.

Amram ShapiroAmram Shapiro, Book of Odds’ founder and president, talked about the company’s plans for its storehouse of statistics for the first time on September 29 at the Web Innovators Group meeting in Cambridge, MA. His presentation won the audience favorite award, beating out social networking site Epernicus and socially networked customer relationship management portal Batchbook. The official launch party for the free, advertising-supported, privately funded website will be held tomorrow at—where else?—the Boston Museum of Science.

Book of Odds is, it must be said, a profoundly odd site. In an era of mindless, automated search engines, it’s an authoritative, highly curated guide to primary data. On an Internet full of YouTube videos and computer-generated virtual worlds, it’s stodgily text-based. In a world where the conventional wisdom says that people can’t cope with math, it’s obsessed with probabilities. It was originally going to be an actual book—but Shapiro says he changed course after an editor from Oxford University Press, publisher of the venerable Oxford English Dictionary, advised against old-fashioned paper. (“He said ‘If we were publishing the OED today it would go straight to the Web,’” Shapiro explains.)

If you’ve ever seen the Harper’s Index feature of Harper’s Magazine, you’ll be familiar with the curiosity-piquing quantities that predominate at Book of Odds. The core of the site—its equivalent of dictionary entries—are the “odds statements.” For example: “The odds a golfer will be injured while playing golf in a year are 1 in 644 (US, 2006).” Unexpectedly, that’s higher than the odds of being injured while mountain biking, which are only 1 in 839 per year, according to the site.

Each odds statement includes the specific geography and time span to which it pertains, as well as full attribution. The golf-accident data, for example, is from the National Safety Council, whose staff must keep very busy documenting the myriad hazards of everyday life. In fact, there’s a whole section of Book of Odds devoted to Accidents & Death, along with sections on Health & Illness, Daily Life & Activities, and Relationships & Society.

The site is a veritable treasure trove for worry-warts, pessimists, cynics, and paranoiacs. To wit: “The odds Massachusetts will suffer an earthquake with a magnitude of at least 6.0 in a year are 1 in 154.5 (US, 1700 – 2008).” “The odds a high school honor student will cheat on a test at least twice in a year are 1 in 2.86 (US, 2008).” “The odds a person will be murdered in a year are 1 in 17,590 (US, 2003).”

Book of Odds is also destined to become a rich source of material for Cliff Clavin-style barroom exclamations, wager-settling, and general head-scratcher trivia. Did you think triathlons were grueling? You’re actually far more likely to be kicked to death by a cow than to expire during the Iron Man competition. (There were 21.5 cattle-related deaths per year in 2003-2007, compared to only 5.1 deaths per year in triathlons.)

In short, Book of Odds has more in common with an insurance company’s actuarial tables than it does with the Web 2.0 universe; Shapiro jokes that it’s “the killer app for the 19th century.” But there’s also plenty about it that’s 21st-century. For one thing, the company built its database of odds statements using semantic middleware from Cambridge Semantics of Cambridge, MA. The system stores a lot of meta-data about each statement, which makes it easier to search, update, organize, and repurpose the information. (But just because the software is “semantic” doesn’t mean it’s drawing inferences, anticipating users’ needs, or making connections between disparate facts in the way a lot of the hype about the so-called semantic Web would have it. It’s just a way to file away data according to its meaning, rather than some abstract address in a database. “We’re somewhere in between Merriam Webster and Wolfram Alpha,” says Shapiro.)

For another thing, Shapiro and his colleagues have some ambitious plans to capitalize on Web 2.0 trends such as personalization and social networking, which create fertile ground for targeted advertising. On one level, the company can target users according to the keywords they’re searching for, just as Google does; people who search for odds statements about cardiovascular disease, for example, might be shown ads for Lipitor.

But the company also has plans to get to know its users on a deeper level. People who set up accounts on the site will be able to create a personal book of odds by bookmarking statistics relevant to their own experience, and the more of these statements they bookmark, the easier it will become for the company to sell Web advertising targeted to their interests. Says Shapiro, “As people fill out their profiles, with the proper protections and anonymization we will be able to provide value”—that is, show tailored ads—”based on detailed knowledge of what matters to them.”

Down the road, Shapiro says the company will also explore other ways to monetize its understanding of its users, possibly including contract research projects. And on a much more immediate basis, the company will earn money by selling personalized merchandise such as T-shirts emblazoned with odds statements.

Shapiro says Book of Odds was born in part out of two convictions—that it’s difficult to translate the statistical data we all hear everyday into something more meaningful, and that the traditional media underestimate most audience members’ ability to deal with numbers. “If a doctor gave you information about the odds of this or that affecting you, you have no place to turn to figure out what that means,” Shapiro says. At the same time, he says, “Everywhere you turn, you read about how people can’t understand numbers. I thought to myself, ‘How many times have we said in the past that people couldn’t do something, only to find that when they are given the right tools they can do very well.’”

In the case of the Book of Odds, one of the main tools is juxtaposition. The semantic search capabilities of the site mean that once you find one odds statement, it’s easy to find others with similar odds, giving a more intuitive feel for the meaning of each statistic. As Shapiro writes in his “Founder’s Blog” at Book of Odds, “The odds a female rape victim is under 12 [1 in 3.45] are about the same as a 100-year-old man dying in the next 12 months [1 in 3.56]. By placing similar odds beside each other without regard to their content, the juxtaposition sheds light for a reader with familiarity with any of the subjects.”

But while the content of Book of Odds reflects Shapiro’s personal interest in numbers, the venture is also the fruit of decades of study of business operations and consumer behavior. Shapiro has done stints as a director at Arthur D. Little (ADL) and Pittiglio, Rabin, Todd, & Mcgrath (PRTM), both well-known strategic and operations consulting companies based in Boston and Waltham, MA, respectively. The startup’s chief operations officer, Louise Firth Campbell, is a 30-year veteran of ADL with a background in business strategy, technology, marketing, and public policy, including healthcare and risk management. “We come with deeply died-in-the-wool instincts about how to make money and how to run successful businesses,” says Shapiro. “This isn’t a non-profit—it had better support itself or we’ll never be able to do what we mean to do.”

Book of Odds still feels a bit rough around the edges. For certain topics, such as influenza or divorce, there’s an overwhelming list of odds statements, each only slightly different from the next, so it may be hard to find the specific fact you’re looking for. For other subjects, there’s no data whatsoever: the site’s database of 250,000-plus odds statements includes nothing on the dangers of helicopters, volcanoes, or X-rays, to cite three subjects I searched at random. And quite a few of the odds statements seem to be formulaic reworkings of obscure labor statistics. For example: “The odds an employed person 16 or older in Ohio is an atmospheric or space scientist are 1 in 19,010 (OH, US, 5/2008).”

But with all the management and technology firepower behind the site, together with a huge editorial staff and a well-developed process for generating, checking, and explicating odds statements (the site is peppered with essays that draw together interesting tidbits, like this piece on vanity license plates), Book of Odds is bound to fill out its subject catalog and gain coherence fast. The odds you’ll be hearing more about it over the coming year are 1 in 1.

Wade Roush is a contributing editor at Xconomy. Follow @wroush

By posting a comment, you agree to our terms and conditions.