April 28, 2008
New search technology helps execs make sense of all that’s being said (good or bad) about a company in the blogosphere. Why everyone from hedgies to Motorola is signing on.
By Carleen Hawn
April 28, 2008
For finance executives, staying on tiop of market information has become a daily ritual. Check the Bloomberg terminal, check the wire services, check the stock price. CFOs who fall behind the curve can find themselves getting tripped up on conference calls by research analysts, or worse, misjudging how corporate news will play in the marketplace.
Wading through the daily dose of news, information and flat-out gossip can be remarkably time-consuming, however. Corporate managers have long had to cope with data from traditional sources such as stock exchanges and regulatory bodies, as well as news reports. Now they must also deal with the swell of opinions and rumors appearing daily on Internet discussion boards and blogs.
That’s no small task—but it’s a vital one given the growing influence exerted by Web commentators. To get a better sense of what’s being said in the blogo-sphere, some companies are turning to a new technology that sorts so-called “unstructured data”—that is, info that can’t be easily categorized by a traditional database (blogs, for instance). In most cases, the software relies on sophisticated algorithms and search tools to trawl the Web and aggregate hard-to-categorize information. The info is then packaged into e-mail feeds or sleek dashboards that display how corporate information is portrayed. In other cases, “user-generators” deliver a fresh financial perspective that isn’t available from traditional sources (see the story on Wikinvest on Page 13).
The result? Users—institutional investors, asset managers and corporate executives—receive organized financial information in a hurry, and in a format that can give them a qualitative edge.
Hedge funds and institutional investors have been the early buyers of the information. But the market appears poised for rapid expansion, with interest from corporate clients picking up. Said Penny Herscher, CEO at FirstRain, an unstructured data specialist: “If you’re a CFO, you have to see the same information that your investors are seeing.”
Sean O’Dowd, capital markets senior analyst at Financial Insights, a unit of IDC, puts total revenue from unstructured data research at $100 million right now. But over the next five years, O’Dowd thinks, the market will top $250 million. “It will mushroom,” he predicted.
Jay WatsonCOLOR-CODED CONTENT “Once it’s in the Wall Street Journal, it’s old news,” says Kevin Pomplun, founder of SkyGrid, which aggregates information and then analyzes its tone. Sentiment journey
If corporate customers and institutional investors are suddenly glomming on to the importance of unstructured data, so too are providers of more traditional research. Last year, Reuters bought Waltham, Mass.-based ClearForest, while Goldman Sachs made a minority investment in Connotate Technologies of New Brunswick, N.J. And Dow Jones agreed to distribute its news feed on the InfoNgen platform—owned by Instant Information—which monitors content from roughly 15,000 sources, including e-mail, blogs and search engines.
Venture capitalists appear to be picking up the scent as well. SkyGrid, an unstructured data search specialist launched in 2005, has raised $2.25 million in venture capital in two rounds, from Tim Draper of Draper Fisher Jurvetson, a VC firm in Menlo Park, Calif., and New York-based angel investor Esther Dyson. The company came out of stealth mode in February.
SkyGrid uses Web crawlers and search algorithms to amass thousands of financial news stories on a topic or company. But SkyGrid does something no other aggregator does: It applies another set of algorithms that filter for semantics and natural language, to quickly convey to clients the actual tone of the news. Most important for investors, SkyGrid also indicates how that mood is changing in real time.
The idea here is that corporate news alone doesn’t move a company’s share price. The investing community’s perception, as well as the media and Internet portrayal of that news, also drives the stock price. Thus it can be equally important for CFOs and institutional investors to track the sentiment about a news item and not just the item itself.
HOW IT WORKS: SkyGrid’s first column displays headlines from traditional news outlets. The center column culls from blogs and other Internet sources. All items are color-coded, indicating positive, negative or neutral stories. The fever chart below reflects the prevailing sentiment about the company. A screen shot of a SkyGrid dashboard (at left) shows how this works. The dashboard displays a list of items about Apple Computer. At first, the search results look a lot like what you might see on wire service terminals, including chronological lists of stories about the company.
But the similarity ends there. Besides mainstream news sources, which appear at left, SkyGrid’s center column displays content from blogs like Gizmodo. “We know from experience that news breaks faster in the blogosphere,” said SkyGrid’s 26-year-old founder, Kevin Pomplun. “Once it’s in the Wall Street Journal, it’s old, so for our clients the information we pull off blogs is more valuable.”
Look more closely and you’ll also see that the results are color-coded. Positive stories about the iPod maker are identified by a green icon. Negative stories are red; neutral news stories are coded white. SkyGrid breaks down the positive, negative and neutral news in percentages at the top left of the screen as well. So without even stopping to read the headlines, an investor or CFO can assess in seconds whether the overall sentiment on a company is bullish or bearish. SkyGrid also produces a fever chart showing the linear development of the mood on a company. Unlike similar graphs on Bloomberg or Yahoo Finance, SkyGrid’s charts are based on news items, not past stock price performance. Such a feature could prove beneficial to CFOs or IR chiefs, who might be able to arrest negative news sentiment before it hits a company’s stock price.
Greg Parsons of CP Capital, a SkyGrid client, said the sentiment tracker can prove invaluable. A number of times, he noted, “a shift in sentiment has preceded negative news” coming from companies.
SkyGrid clients pay $500 a month per user for the service, which isn’t much (by comparison, a Bloomberg terminal costs $1,800 a month). The company has less than $1 million in annualized revenue, but sales are growing. One signal: When the company launched the technology following its beta test in February, it had just 30 users. Now it has more than 100.
Mr. Pomplun feels confident that revenue will pick up dramatically. That belief, he says, stems from SkyGrid’s product—and the amount of unparsed information that bombards corporate managers and institutional investors every day. Said Mr. Pomplun: “We want to accelerate the wisdom of crowds.”
Fascinating algorithm
FirstRain, a Foster City, Calif., company, has a similar agenda. The company already boasts several corporate subscribers, including the finance departments at Motorola, E2open and Talecris Biotherapeutics. FirstRain also provides customized daily and weekly news feeds to more than 500 money managers and professional investors.
The company applies several patented algorithms to the data it pulls from blogs, wikis, discussion boards, industry trade publications, even far-flung international resources and very small regional newspapers. “It’s the obscure sources of information that investors need to make money on the margins,” explained Ms. Herscher. “For our clients, it’s not really about speed, but the fact that there is richer [qualitative] information on the Web.”
A feed can cover a host of subjects, including obscure information about supply chain vendors or corporate partners. To supply that, FirstRain relies on more than 100 research analysts in Gurgaon, India, whose sole mission is to review Web content for substantive value.
In January, First-Rain launched a new product, Management Monitor. One of the application’s strongest features is its management turn-over report, which pulls news about executive departures from around the Web, then represents it in a graph. Relevant Web items about companies in the turnover report give context to the numbers. Management Monitor also pulls together “comments” and “no comments” from executives, as well as topic-driven Web items.
The goal is to distill unstructured data into concise data points and trends, revealing competitive intelligence rivals don’t possess.
For example, FirstRain recently ran an item indicating that Netflix might be expanding its offerings. It culled the info from a blogger who had posted details about his experimentation with game downloads for Microsoft’s Xbox 360 that he’d pulled off Netflix. A large mutual fund manager in New York City who had a stake in Netflix read the piece in his FirstRain news feed—and immediately called Netflix CEO Reed Hastings. The investor hadn’t known that the movie rental service was developing a game console strategy.
In another example, FirstRain recently pulled a story from a regional Alabama paper about a local McDonald’s experimenting with a new test product. That bit of information came as news to some investors who follow the restaurant chain.
David Rosenfeld, the director of research at New York-based William Jones Investment Management, is a FirstRain client. “We try to get information from [a company’s] management, but they are…restricted on what they can actually say,” he said. “So that makes secondary information that much more important.”
FirstRain’s service is not cheap. Investors pay anywhere from $10,000 per person per year (for professional investors) to as much as $75,000 for organizations such as corporate finance departments.
It’s money well spent, said Mr. Rosenfeld. “There’s too much going on [on the Web], and it would be too hard to create an inventory of relevant potential information. It wouldn’t be realistic.”FW
