9:00 am – 9:45 am

Listen to a Preview (8 MB)

Despite the billions of dollars invested in information technology, organizations are still failing to realize the latent potential of their information. To be successful, organizations need to take a different approach—one that views information as a critical business asset, not an afterthought. Organizations must put exploiting information at the heart of the way they do business. A recent Capgemini research report concludes that poor information exploitation is costing the U.K. economy over $100BN/year in lost profits, representing an average 29% suppression of business performance per organization. Better information exploitation is one of the last bastions: an opportunity to both grow revenue and improve profitability.

10:00 am – 10:45 am

Listen to a Preview (3 MB)

Nobody wakes up in the morning thinking: “Oh Boy! Today I’m going to search my entire enterprise!” Instead, they arrive at your intranet, extranet, or enterprise site with a simple goal: to find what they’re seeking. In this entertaining presentation, Spool will share some of Search’s well-kept secrets. He’ll show you how users follow a “scent” that leads to the content they seek, as well as what site designers can do to make that scent stronger. He’ll reveal the tricks to getting the most out of an enterprise search implementation and show you the best way to keep tabs on how your users are succeeding on your sites. You’ll come away with solid insights based upon User Interface
Engineering’s up-to-date research.

Breakout A-1 — Alternatives to an Expensive New Engine

Listen to a Preview (5 MB)

Could your existing search engine be improved, or make it one more year? What types of searches are failing, and what can you do about it? And if you really do need a new search engine, what are lower cost options? This session will provide a checklist for reviewing a search configuration, existing content, and presentation of results to either find and fix problems with an existing engine, or which solutions to look at and what to ask.

Breakout A-2 — Ensuring Enterprise Search Success

Listen to a Preview (3 MB)

In large enterprises, content sources are not connected. It is hard to make search work when it must reach beyond a single room to encompass petabytes of data. Based on over 15 years of experience in IA, e-discovery, enterprise search, and more, Dawson will outline the reality of large-scale search deployments. He will discuss real issues faced in sensitive legal and insurance environments and deliver practical advice.

Breakout A-3 — Emergent Social Search Experiences

Listen to a Preview (7 MB)

The social web is all the rage, but does “social” work for the enterprise? Like water cooler wisdom, social search can help surface information workers might not know exists. Search inside and outside the enterprise is becoming social. This session will look at the evolution of social search, how crowd wisdom is evolving search interfaces, and what the enterprise can learn from these developments.

Breakout B-1 — Don’t Just Find Stuff–Answer My Question!

Listen to a Preview (7 MB)

This session will show how to use three techniques to make enterprise search more effective. Few people doubt that search tools are useful for any organization, but the standard search implementation is a blunt instrument. Clever functionality can be employed to narrow down the resulting hit list, (e.g., relevancy tuning, pseudonyms, guided navigation). Yet the effectiveness of the hit list remains compromised by the question’s limitation: It is still a list of “stuff out there.” This session will demonstrate effective ways to help users help themselves.

Breakout B-2 — Tuning a Complex Search Solution

Listen to a Preview (1 MB)

With dozens of high-traffic websites as well as a large quantity of internal assets to manage, the software development team at MTV Networks is always looking to reduce complexity. Sometimes the media business moves too fast to realize it takes time to develop solutions that will simplify work in the long run. Cohen will discuss MTV’s experience with search on its public site, as well as in internal content management tools and how the needs of users have swung from a complex, hard-to-maintain system to a simple, hard to customize system and then back toward the middle, where it is finding a happy medium leveraging both types of systems to satisfy different needs.

A social bookmarking system (based on open source tools) has been implemented at MITRE and proven successful through its adoption rate and ability to provide users with shared knowledge. One goal was the integration of shared bookmark content with enterprise
search via the indexing and searchability of individual social bookmarks. After the basic implementation, there have been “downstream” enhancements and benefits in user search interfaces and tools. These include the creation of a dedicated search interface for bookmarks, the promotion of bookmark search results via Google Onebox, and the use of social bookmark content and accompanying personal staff metadata to identify corporate expertise via a homegrown expertise finder system.

Breakout C-1 — What to Do Before You Replace Your Engine

Listen to a Preview (1014 KB)

A poorly implemented search solution can create an experience that causes users to abandon searching permanently, creating the impression that the company is not tech savvy enough to manage content and data. When search results are subpar, users and site owners assume the search engine itself is the culprit, which results in mandates to replace the search engine. The thought is that if the old search isn’t working, it must be broken and a new search engine will resolve everything. Typically, however, the technology is not the primary issue. Often, simply improving the quality, construct, and tagging of the content can provide remarkable improvements.

From cars to shoes to computers, it seems that everything these days is personalized. Why not search? Thompson will discuss how search can be personalized to the department or even individual level—providing users with the exact information they seek.

Unified Information Access: Combining Search and BI Capabilities for Better Business Decisions

Listen to a Preview (2 MB)

Search-Powered Compliance

Listen to a Preview (3 MB)

The Improvement and Development Agency (IDeA) for local government has integrated Exalead’s API online search functionality into its Communities of Practice (CoP) Platform. Many online communities have been established by local authority, regional and IDeA staff. Community facilitators can now generate a “favorite” list of websites for members that are used as the basis of their domain-specific online searches. This improves precision and relevance of results, since only CoP member-recommended websites are included in the search. The effect is that all members of the CoP—all 411 councils in local government in England and Wales—have access to a collectively agreed set of favorite websites for domain-specific websites as opposed to the entire web.

Breakout D-1(a) — Enhancing Findability and Usability: Taxonomies and Search

Listen to a Preview (5 MB)

The State of Minnesota uses several tools to integrate taxonomies, navigation, thesauri, and search on both agency sites and cross-agency “topic portals.” Information architect Eileen Quam describes the IA methodologies employed to organize websites and shows live examples of public-facing and back-office implementations. Hear how Data Harmony’s Thesaurus Master is utilized to enhance enterprise search, integrate with content management systems, and display taxonomies. Takeaways include tips for implementation and usability.

In 2003, Meredith Corp., publishers of Better Homes and Gardens and other magazines, moved its photographs, illustrations, layouts, etc., into Open Text’s Artesia program in what it calls the Creative Library. After a few years, users began complaining that they weren’t getting many results when they searched for photos. Delving into the metadata revealed there just weren’t enough search points. Meredith chose A2Z Keywording to add the additional search points/keywords. Now in its second year of the project, Meredith’s Creative Library has added more than 9 million keywords to more than a quarter-million of its digital assets. The Creative Library team now seeks to balance
delivering too many results with over-refining the metadata.

Breakout D-2(a) — Wyeth Finds a Fit for Pharmaceutical Search

Listen to a Preview (5 MB)

Wyeth recognized the shortcomings of its incumbent search technology: Declining use and growing complaints confirmed that employees could not find the information they needed. A reorganized Wyeth Enterprise Search team rose to the challenge and delivered improved internal search. They adopted powerful Pharma techniques—problem segmentation, learn & confirm, controlled clinical trials, tracking adverse events— and questioned many search conventions search strategy, federation, taxonomy and heavyweight projects. The new search solutions cover a broader range of repositories, give near-instant response with relevant results, and provide a platform for enhancements. Search usage grew nearly fourfold year-over-year while annual search technology costs dropped.

Large multinational corporations have invested in global security operations to detect and interdict threats ranging from fraud and intellectual property theft to the safety and well-being of the company’s physical assets and employees. Pharmaceutical counterfeiting
alone costs the industry billions of dollars per year. At Pfizer, the world’s largest research-based pharmaceutical company, search, real-time monitoring/alerting, and batch analysis are being used to discover and continuously monitor available information across a number of decentralized repositories and sources.

Breakout D-3(a) — Building the Foundations for Semantic Search

Listen to a Preview (1 MB)

Cutting-edge problems in search, KM portals, and enterprise information management alike have a pressing common need for in situ creation of context vocabularies, long term modeling of interest profiles, and personalized presentation of information within
the world view of the user. We argue that automated generation of vocabularies, taxonomies and ontologies can support the level of productivity expected from the next generation of knowledge management professionals and stands to power information delivery platforms. Beginning with a description of numerous use cases, this talk will demonstrate how cutting-edge research in mining large graphs powers the latest features in KM, conceptual search, data integration, and compliance.

Sentiment analysis is one exciting application of text technologies today. It may also be the most challenging. Sentiment analysis uses software to discern, extract, and process attitudinal information found in text. Correctly detecting sentiment—object, polarity, and intensity—is no easy task. Systems need to go beyond value terms to understand context, slang, idioms, and patterns that indicate negation and even sarcasm. Sentiment analysis is applied to a variety of web and enterprise sources, which capture both fact and opinion. Grimes will cover the business case for sentiment analysis, the technology, and applications, including reputation and brand management,
customer support, competitive intelligence, marketing, and quality control.

Breakout E-1 — Improving Findability Through Site Search Analytics

Listen to a Preview (3 MB)

Site search analytics is gaining recognition as a powerful method for understanding users, information needs, and information seeking behaviors. It can help diagnose and address problems with search, metadata, navigation, and content. Rosenfeld—coauthor of Information Architecture for the World Wide Web and the forthcoming Search Analytics for Your Site—will show you how you can analyze users’ queries to improve searching and browsing. Attendees will learn to balance the appropriate mix of qualitative and quantitative methods required to make truly informed decisions.

Quick and focused retrieval is important, but insufficient because “finding” isn’t the final answer, it’s a first step. Successful enterprise search projects must answer the questions, “After search technology found it, what is the user going to do with it? What problem are they trying to solve and how can we help?” We need to answer the question “what is the next step for the user when looking at the results page and position search to assist with that?” Only then will enterprise search really start to deliver the promised value. Denissen will provide real-world insights into how she’s helping Shearman & Sterling LLP go beyond finding to taking action with information.

Breakout E-3 — Searching Collaborative User-Generated Content

Listen to a Preview (983 KB)

Collaborative user-generated content (UGC) creation and publishing have made their way into the business practices of large companies. Ranging from unstructured wikis to topically structured forums, UGC is forming the institutional repositories of knowledge created, edited, and organized by the contributors. One variety of UGC, Community Question Answering (CQA), is starting to make inroads into the enterprise. This talk will focus on the challenges and new techniques in modeling, organizing, and searching user-generated content, with a particular focus on CQA paradigm of information sharing. Agichtein will describe new problems related to mining this content, including learning to rank for UGC search, estimating content quality, contributor authority, and asker satisfaction with the results.