Print
Email
Print Edition Stories
Sandie Allen
MHT's Efrain Viscarolasaga: How social networking is changing business
MHT's Galen Moore: Charities score with online fundraising
MHT's Jim Connolly: Social network sites focus on niches, sports
What if you could ask the web a question — and get an honest answer?
If you’re a financial analyst, a Boston startup called MyRoar Inc. says you can. The bootstrapped venture is building a semantic search product designed to answer analysts’ questions directly, like, “Is anyone backing this bond?” or “How many hedge funds are there in the U.S.?”
Co-founder and CEO Kate McDonough started the company in 2007 while still an analyst with Lehman Brothers. Last May, she left Lehman and in October merged technologies with another nascent startup, founded by MyRoar CTO François Schiettecatte, former chief scientist at web search company Feedster Inc.
McDonough based MyRoar’s offering, now in development, on her own experience as a Lehman analyst. Now 27 years old, she had been thinking about the problem long before Lehman went into bankruptcy. While a computer science major in college, McDonough saw her mother, a WorldCom Inc. employee, lose her retirement fund. After that, all she wanted to do was help protect investors from exposure to events like WorldCom’s collapse.
“There are so many inputs analysts don’t have time to consider,” she said. Automated natural-language queries of research databases and the web will provide answers faster, allowing analysts to monitor more datapoints and see a more complete picture of each investment they’re watching, she said.
McDonough cited as an example a time when she was looking to see if a third party had propped up a particular bond to prevent it from default. She spent two days trying different keyword searches. Coming up with nothing, she made a sell recommendation, only to later find an obscure blog post indicating that particular bond had a corporate backer. The only keyword that could have turned up the information was the name of the backing company, she said — a Catch-22 situation that natural-language search can eliminate.
From Ask Jeeves Inc. in the late 1990s, to San Francisco-based Powerset Inc., bought by Microsoft Corp. (Nasdaq: MSFT) last year for more than $100 million, many companies have pursued the promise of semantic search across the web.
In MyRoar’s case, accurate natural-language search results are possible because the company has restricted itself to the financial domain, said Mark Finlayson, the MIT Ph.D. candidate who developed the machine-learning algorithm on which MyRoar’s application is based.
“On the one hand, you have your technique for doing the learning, but on the other hand you have some structure that you know holds in the domain,” he said. “You’ve reduced the ambiguity. In the finance world, ‘bank’ never means ‘riverbank.’”
Even so, MyRoar’s potential customers will not want to hear how the company’s technology works, said Gartner Inc. (NYSE: IT) analyst Whit Andrews. They’ll want to see the return on investment the software promises. The average computer worker spends five hours weekly looking for information online, Andrews said, citing a Gartner study. “If you can reduce that time or make it more efficient, you win,” he said.
MyRoar Inc.
Location: Boston
Employees: Two founders
Business: Natural-language search for financial analysts
Founded: 2007
RSS Feeds
Comments (1)
Please Login/Register to post comments.