In this keynote panel, Stephen Arnold discusses the challenges facing the search engine leaders and inquires whether they have reached barriers to their seemingly unhampered growth.  He suggests that both Google and Microsoft may be facing threats from new and more agile competitors who are seeking to reap the rewards of new opportunities that these behemoths cannot pursue, even with their billions of dollars. Following Steve's remarks, representatives from Google and Microsoft will comment, and then the session will open to take questions from the floor. This keynote promises high-impact commentary and spirited discussion spiced with audience involvement.

It is not enough just to deploy your search capability and add a taxonomy or integrate it with your content management system if your end users are still not finding what they want. Tuning or enhancing search is an arduous process, but some best practices are rapidly emerging. Avi Rappoport and the real-world situations she describes offer powerful examples of how search logs, metrics, and analytics can open a window into user behavior and allow you to improve search to meet user expectations.

Hear how a search upgrade to the knowledge management system globally used by consultants at Deloitte measurably improved user satisfaction and made it much easier for users to find content.  Improvements were accomplished by making better use of the existing search engine features and doing targeted customization.  The key aspect of the project was using a testing tool to submit raw query  language directly to the production search service and then reviewing results under different scenarios and using different query logic. This meant the tuning was validated before development got involved. Learn how the process worked and lessons learned.

This session looks at what is going on under the hood of today’s enterprise search technologies and discusses how we can optimize the information architecture to help get our customers to the information that they need to do their jobs. Learn what makes a site structure “search-friendly” (and about the HITS algorithm and authority scores), how to enhance relevance through association (outlinks and inlinks with a point of view), and about relational content (“birds of a feather,” or where to send visitors once they arrive and why).

Learn how to tune your search engine and deliver better results to your users by using best bets, search synonyms, and search logs and
how to use activity reports not only to improve search results but as a gold mine of information about your customers, your Web site, and
your company. Miles Kehoe describes how your search logs can become your ultimate business intelligence tool. Lee Romero will discuss a tool used on the Novell intranet to implement best bets in a unique way, and Patricia Quinn describes how Vision Service Plan used best bets and other strategies to refine search results based on search reports from Thunderstone’s search appliance.

Search logs contain rich user research that can provide valuable insights into the mental models of a Web site’s users. If your organization has search logs, the reports likely consist of a list of the queries and how many times those queries were executed in a given time frame. Your top queries may or may not represent the top concepts searched because of the variation in keywords that represent the same concept. This case study will walk you through a methodology for building a search log taxonomy, including how it can improve your site’s navigational structure.

In this session, attendees will hear about a variety of today’s search techniques including federated searches, query expansion, and onthe- fly foreign-language translation, as well as how to uniquely leverage the local search capabilities of each content source. The session also examines critical applications for content integration and search technologies including research and development, competitive intelligence, customer service, and electronic discovery, and suggests a framework for devising and deploying an ECM strategy.

After 12 years on the Web, people have found some useful mental models for dealing with search. But there are still far too many search engine interfaces that have confusing labels and unnecessary elements, are missing obvious functionality, or otherwise fail at being usable. This thought-provoking panel will discuss the most common mistakes, discuss both good and bad implementations, describe when to break the rules, and present some of the more recent research on search user interfaces.

Usually when people talk about “semantic search,” they mean concept-based searching, which is searching on an idea versus keyword searching. There are a number of ways to enable semantic search, including linguistic analysis, Bayesian algorithms, rules-based approaches, and even so-called “social search.” Semantic search is based on context and solutions and considers the body of content, user processes, and user roles. This informative session discusses how to make semantic search practical and useful.

Hear about Radar, an application currently in use at the MITRE Corp., that uses concept filtering to manage RSS news feeds for users. A concept search, in which multiple senses of a term are disambiguated and multiple names for a thing are known, is a good alternative to keywords. However, a problem with this approach is building a broad and deep concept space. By mining Wikipedia, MITRE created a term concept map with more than 2 million terms and more than 1 million unique concepts. Radar uses this map to associate text resources with unique semantic concepts. Learn how MITRE extracted data from Wikipedia and how other organizations can do the same by using free, open source software.

Learn how end users are leveraging semantic technologies such as relational navigation within the enterprise to reduce costs, enhance revenues, and manage information assets more effectively. Copresenter Jason Busch, who runs the popular spend management blog, SpendMatters, will describe how he has leveraged the semantic search capabilities of relational navigation to increase readership and create a more user-friendly navigation experience for his readers.

In May 2007, Steve Arnold’s analysis of Google’s efforts to capture the semantic Web appeared as part of a Bear Stearns analysis. In this session, Arnold will review: 1) Why the semantic Web is up for grabs; 2) Why Google wants to control the semantic Web; 3) What technical inventions Google has to achieve this goal; 4) What this means for content owners, SEO experts, and competitors. Google has been working on this technical initiative for more than 2 years, yet kept it hush-hush. In this presentation, you will learn what Google isn’t ready yet to let the world know.