First Look

What you don't know can hurt you - especially when it's Web 2.0 information.

It's estimated that enterprise IT systems can now access at best 20 percent of pertinent data generated by employees, customers, analysts and competitors.

"These people go the Web, they blog about it, they write customer reviews, they write notes on email and all that information is not readily accessible by enterprise IT systems, so there's sort of a gap that's opened up between the enterprise and the Web 2.0 generation data," said Denodo Technologies VP of Marketing Suresh Chandrasekaran. "And also, there is a ton of relevant structured information in the Hidden Web like industry trends and projections, demographic data and competitive technical and pricing data, which all goes untapped today.

Finding and structuring this data and merging it with traditional IT systems is the challenge.

"In the past, a lot of the feedback about the products would come in form of warranty forms and customer service forms. But today people will send an email or go out on digitalphotographyreview.com and write about what problems they have, what features they like and what competitors they compare you to," Chandrasekaran added.

"All this information would be very valuable to a product manager at a company like Canon, or to a warranty manager to prevent costly blowups they could have predicted if they knew about it," he noted. "We are able to search for that information but also structure and aggregate that information so we can answer questions like 'How often was the Canon 20D compared to the Nikon' or 'What were the most frequently reported three problems?'

"And this is a summary of information across, let's say 3000 blog entries or we could ask the same question from your internal CRM systems' call notes," Chandrasekaran said.

To display the data, Denodo uses Web 2.0 techniques that follow a SOA-related paradigm to provide data mashups that feed traditional applications, data warehouses, BI and single views.

"We use a combination of search indexing and data keyword analysis, Hidden Web data extraction and virtual integration of structured databases from multiple heterogeneous stories and Web extraction," Chandrasekaran said.

Once the information is combined, it appears to consuming applications like just another database, presented as data services consumed by SOAs or as analytics or reports or portals that provide unique insights.

Chandrasekaran also noted Denodo's interesting origins at a Technical University of Madrid project.

"t was all funded and driven by problems faced by businesses and governments in Europe," he observed. "One of our applications is for the EU's Homeland Security Agency which is asking us to analyze all sorts of information across Europe and related it back to terrorist activity."