A few unmonitored years can make a huge difference in search relevance. A project in 2009 to tune the E&Y search engine—using measures of precision, recall, and mean reciprocal relevance—was revisited 3 years later. In the intervening time, precision decreased, as did the mean reciprocal rank measures. That’s what happens when you ignore relevancy over time. This presentation clearly demonstrates that search is not a “fire and forget” project and provides some specific tips so you can avoid this situation at your workplace.

We are starting to see a slow convergence of business intelligence and enterprise search in the world of corporate IT. One catalyst is Big Data, but the continued focus on unstructured data remains a powerful force. Analyst and industry observer Stephen Arnold gives us his views on where enterprise search and business intelligence intersect, how the lines are being blurred, and how the information discovery elements of both can further the goals of the organization.

Two case studies are presented in this session. From Hewlett-Packard we have expert findings, which can be both a problem and an opportunity. Everyone from sales teams to project managers to human resources to frontline problem solvers benefit from a versatile engine that taps into a search engine that will rapidly identify expert knowledge. Fried shows several successful systems that use social behavior to provide strong cues to expertise finding. Applications in recruiting, collaboration, and team creation are highlighted.

Big data poses big problems and demands big answers: How do you scale your search mechanism to sift through billions of bits, index them and then and retrieve large result sets in a matter of milliseconds? Hadoop can store terabytes of data, and Lucene/Solr can help with the search tasks, but how do you create the best applications and make the results immediately available to multiple search clients? This session looks at common use cases for generating search indexes from big data, typical patterns for the data processing workflow and how to make it all work reliably at scale.

As our ability to capture data increases, our appetite to consume data also increases. This produces rising demand for lean and agile solutions that provide fast and flexible access to data. Q-Sensei unveils a preview of its new integrated solution to organize, search and discover Big Data.

Kapow Software will present case studies on how they were able to help several global enterprises provide a seamless and consistent user search experience by mapping legacy and user-generated, unstructured data across multiple knowledgebases to a standard taxonomy and controlled vocabulary for improved search results.

How to optimize findability in collaborative information stores was the issue facing Oak Ridge National Lab. A critical assessment of the SharePoint environment, including analysis of emails, personal interviews, weblogs, and web-based user surveys identified findability problems so they could be solved. Consultants at Consejo believe that search is not the answer. Finding is the solution. Novel and compelling uses of search boost findability, with practical implementations ranging from Google to SharePoint/FAST to FunnelBack.

Large companies, such as Pfizer and Northrop Grumman, have correspondingly large repositories of content. At Pfizer, the challenges of separate search portals and multiple repositories was addressed by combining intranet search with specific business line search portals. Faced with a diversity of information collection platforms and growing volumes of data, including that from social media, Northrop Grumman developed some open technology development (OTD) projects that necessitated decisions about governance as well as technology.

Balancing the need for less IT complexity, cost, quick development time, standardization, and user satisfaction is a tightrope walk. In this final session, a panel of technology buyers and vendors argue for which to choose—the all-encompassing platform or a set of purpose-built applications to fill the need for ediscovery, publishing, marketing, site search, or business intelligence. The audience is invited to join the discussion.