A startup has got to have faith in its product and a lot of nerve to claim it can do what Google and Yahoo! cannot — or at least have not as yet done.
Stamford-based Semantifi, formerly ExeCue, hopes to make oceans of cash from its Resource Description Framework-driven (RDF) semantic search platform capable of harvesting the Deep Web.
“Google, Bing and Yahoo! can search web pages very well. But they can’t search datasets in their finance, shopping or travel portals and get knowledge-based results,” says Shree Pragada, 39, Semantifi’s CEO and founder.
By contrast, Semantifi’s search engine — which unlike Yahoo! or Google goes beyond keyword and symbol search — seeks to narrow down the number of results to match user needs, pulling up information from multiple datasets from what’s called the ‘Deep Web’. Michael K. Bergman, widely recognized as the creator of the term Deep Web, compares the process of retrieving and collating data to diving below the ocean’s surface for hidden resources.
In a white paper by Bergman titled “The Deep Web: Surfacing Hidden Value,” published in The Journal of Electronic Publishing (Vol. 7, No. 1, August 2001), public information on the Deep Web is 400 to 550 times larger than the World Wide Web, according to data collected in March 2000 by BrightPlanet Corp. The study, which was spearheaded by Bergman, found that the Deep Web consists of around 550 billion individual documents compared to just one billion in the World Wide Web. The majority of these are highly specialized and subject-specific. Additionally, the quality of the content in the Deep Web is up to a whopping 2,000 times greater than the surface web.
It’s a concept that has a vast potential market but there’s also huge competition and challenging that will take significant capital. A range of risk takers — including venture capitalists and Connecticut Innovations (CI), the quasi-public investment group charged with helping nurture startups in the state — are circling.
Larry Cannell, senior analyst at the Stamford-headquartered technology research firm Gartner, says the idea of having people build, publish and share apps is the dream of every technology. “If Semantifi is able to get people to find interesting ways of searching datasets and draw insight that would otherwise have been missed, than yeah, it’s exciting!”
Pragada says Semantifi has this capability. He offers an example. “Neither Yahoo! Finance or Google can provide a meaningful result when you enter U.S. companies with market capital of over $1 billion,” he says. When this HBJ reporter followed up with a Google search, the first five listings pulled up a definition of market cap in Wikipedia, an unrelated Bloomberg story, an article on Indian ADRs, an investment bank portal, and the criteria for S&P index membership.
Pragada says if users create relevant apps and add to it on Semantifi, then its search engine can, within seconds, pull up a list of U.S. companies with market cap of over $1 billion, and present the data to users in the form of graphs and charts. But to be able to this, the company must harness the power of the community in a gargantuan scale. Launched just two weeks ago, the search platform expects to attract individuals, corporations, government entities and research firms to publish and offer vertical apps for a fee or free of charge.
To date, Semantifi hosts around 50 free apps such as ‘bailout recipients’ and ‘analyst ratings,’ many of which require additional datasets culled by users. When asked about the accuracy of user-generated information, Pragada says he trusts crowd-sourced content over that put out by any one company or person whose motives could be biased.
Some data, however, can be directly integrated within business intelligence models. For example, Zacks Investment Research, Inc., a leading Chicago-based investment research firm, is in the process of publishing paid search apps on Semantifi, expected to be available publicly next month.
“You can ask questions like ‘which companies have insider transactions of more than 2% of total transactions?’ Another vendor can build better apps and compete with Zacks. It’s an open marketplace for data owners, content owners and Internet users to publish data and share search apps,” Pragada says.
Zacks is in the process of providing a number of data feeds, which will be updated on a weekly or monthly basis. The data feeds are converted into search apps, where users ask questions and get answers with automated visuals within seconds. Semantifi will pay Zacks a share of the revenue that results from that. End users will pay a $20-$30 monthly subscription fee for each Zacks search app.
“The value of our data is not so much in the raw data itself, it’s in the applications that extract insight from the data,” says Bryant Sheehy, director of business development at Zacks. “It’s like the value added by a skilled diamond cutter who applies precise technology to create a beautiful sparkling gem from a rough stone. We’re hoping the Semantifi applications will help customers extract some useful insights from our data that may have been overlooked by other investors.”
The investment firm will provide data related to Wall Street earnings estimates, company financial statements and insider stock trades.
“Our target market is primarily sophisticated investors who want to mine our data to find good stocks to invest in. We also hope to target corporate finance customers who want to keep tabs on their own customers, partners and competitors,” says Sheehy.
The big boys are eyeing this market as well. Microsoft, which launched the federal government’s portal Open Government Initiative, enables users to create solutions-based apps for Africa, apps on improving education in U.S. schools, and apps that explore how kids in America can stay healthy.
“Microsoft’s offering consists of hosting datasets which third parties can download, but you can’t search data because Microsoft doesn’t yet have data search technology,” Pragada points out.
According to Anil Vasagiri, one of CI’s two members on Semantifi’s board,
“Semantifi’s technology trumps other technologies by its ability to process multiple terabytes at Internet speed.”
The company’s revenue model consists of advertising, revenue-share from paid search apps, and licensing fees from private instillations such as banks. Pragada says early revenues are flowing in from pilot deployments. Semantifi raised funds from an angel round in 2007, series A in 2008 and a bridge round in 2010. CI pumped in $1 million in 2008 and $450,000 in August of this year. Private investors include Stamford-based Launch Capital. According to Pragada, much of the $3.5 million raised to date was spent on technology development.
Vasagiri declined to offer an estimate on expected returns, the exit value and likely additional funding by CI. “CI’s ROI threshold is 35% to 40%,” he explains. “The average time to exit for a startup is around seven years. It’s too early to predict potential exit value for Semantifi considering the size of the market for this technology as well as the IP that the company has built in the last few years.”
He says the company’s future capital needs will be driven by the speed of technology adoption, the success of its business model and competition between some of the bigger players in this space. “Considering these business variables as well as the typical investment profile/bite size of some of the later stage venture funds that the company is in discussions with, it’s hard to predict how much CI will invest in the company in the future,” he says.
Pragada too declined to comment on Semantifi’s revenue targets and instead referred to the opportunities in monetizing datasets in the Deep Web. “It’s far bigger than the current search market of $32 billion,” he says.
Still, not everyone is excited.
Robert McCartney, assistant professor of computer science at UConn, points out that the semantic web concept has been floating around for years but relatively little has happened yet. “Some would say that’s a good thing, that there’s opportunity in it,” he says.
Joe DeMartino, president of the Stamford-based, 55-member Angel Investor Forum, evaluated Semantifi several months ago. “Eventually we just couldn’t get enough people who were interested. We typically proceed if a minimum of six members are interested. But we didn’t get to that level,” he says.
DeMartino is keen to add that Semantifi offers a different slice at traditional business intelligence. “It’s a hot area that’s growing rapidly. But it’s a long, uphill slog to build marketshare. They’re battling against some fairly entrenched players. It’s going to be three to five years before they start getting traction,” he says.
Pragada, formerly a senior vice president at Citicorp and a first-time entrepreneur, responds by citing a success story that changed the face of the Internet. “I’m hugely inspired by Wikipedia which became a phenomenon in just a year-and-a-half,” he says. “If they could do it, why not others?”
