Thumbs Up for SemTech 2010

Jun 28, 2010

The annual Semantics Technology conference (SemTech 2010) is taking place in San Francisco this week, and if the first day of the sessions in the Publishing Track proved anything, it's that semantics for the content industry is poised to explode - if it can just get out of its own way.

With attendance up 30% over last year, the 1200 attendees are a testament to the fact that interest in semantics continues to swell, even in, or perhaps because of, a sluggish economy. But the day's sessions zigzagged between the types of technical three-letter-acronym laden presentations by semantic technology vendors that tend to scare off the very buyers who could benefit most from their solutions, and inspiring examples of real world semantic implementation that media and publishing executives could easily extrapolate to address their own specific business challenges.

The keynote address by Dean Allemang of TopQuadrant and Professor James Hendler of Rensselaer Polytechnic Institute, "Semantic Web for the Working Enterprise," set the stage for illuminating the contrast between semantics' undeniable cool factor, and what it will take to fit it into overall enterprise strategy. Using a thumbs up, thumbs down approach cribbed from "At the Movies," the pair reviewed four short videos that addressed both the technical and business requirements for semantics success, including a clip of Clay Shirky borrowing a quote that semantics are "a witness protection program for AI researchers."

Two sessions were aimed at helping semantics vendors get into the mindset of their clients in the publishing industry: Bob DuCharme of TopQuadrant and Luca Scagliarini of Expert System presented "Semantics in Publishing," and Rachel Lovinger of interactive marketing and technology company Razorfish talked about a recent research report, "Nimble," about publishing in the digital age. DuCharme pointed out that publishers have both data assets and metadata already in place, as well as vocabulary development experience; all are concepts found in semantic technology. The challenge is for semantics vendors to learn the publishing industry's vocabulary to present their concepts.

Scagliarini then presented a case study of how Italian financial news site il Sole 24 ORE is using semantics to increase monetization of its web site. Using Expert System's semantic technology to auto-categorize news content, apply semantic tagging, correlate, and enable semantic search, the paper's website now offers improved personalization, native language search, and better leverages archived content. This in turn has driven advertising revenue, as guided browsing has resulted in a higher average page number per user visit.

The Razorfish report, available for free download at,  summarized interviews with participants from major media organizations like the New York Times and The Wall Street Journal. It may not have been groundbreaking to SemTech attendees working in those industries, but is potentially a useful guide for semantics vendors trying to understand the business challenges that semantics can help address, from making sites stickier to doing a better job of rights management to leveraging user generated content.

For their part, publishers were reminded that using semantics needn't be an all or nothing proposition. Both in a session debunking the myths behind the Semantic Web and in the Media and Publishers panel moderated by Tom Tague of OpenCalais, incremental implementation and starting small was encouraged. Panelist Hannah Eaves, LinkTV's director of new media, demonstrated the beta website, a platform for videos from and about the developing world using semantic analysis of video transcripts and automated tagging. "Without semantics, we couldn't have done this," Eaves said. "It's making connections between videos that we wouldn't have even thought of." Now that is approaching launch, the non-profit organization will embark on building a similar video platform for news, thanks to a grant from the Knight Foundation.

It's clear that semantic technology for publishing is here to stay. But for it to move from early adopter to mainstream implementation, SemTech vendors need to take the technology out of the AI box and focus on proving its benefits in business, not technical, terms.


Related Articles

In today's enterprise, information discovery is often treated more as a survival mechanism than an opportunity. And that's only for the companies that have solid information discovery practices in place. For organizations that still rely on simple search engines to uncover data assets within their IT infrastructure, the opportunity costs can be catastrophic.