SharePoint Search
Relevance, Refinement, and People

New and Improved

• Relevance
• Refinement
• People Search
• Enterprise scale
• FAST Search Server 2010 for SharePoint

SharePoint 2010 Search helps people find the information they need to get their jobs done. It provides intranet search, people search, and a platform to build search-driven applications all on a single, costeffective infrastructure. But what's so unique about SharePoint 2010 Search is its combination of relevance, refinement, and people. This new approach to search provides an experience that is highly personalized, efficient, and effective.

• The SharePoint Search capabilities are delivered through two server products: SharePoint Server 2010 delivers out-of-the box intranet and people search. SharePoint makes it easy to find content by combining relevance, refinement, and social behavior in the search experience.
• FAST Search Server 2010 for SharePoint offers a new choice in enterprise search. Combining the power of FAST with the simplicity of SharePoint, FAST Search Server 2010 for SharePoint delivers an exceptional intranet and people search experience and a platform for building custom, search-driven applications.

SharePoint Server 2010
Relevance

• Improved core relevance ranks the best search results more frequently in the top three positions. New ranking ingredients such as URL fuzzy matching, social tags, phrase matching, and extracted metadata (such as author and title) improve result ranking. For example, a PowerPoint presentation with a vague file name like "slides.pptx" could instead use the title on the first slide (for example, "Analysis of Quarterly Results") if it were more relevant.
• Social behavior improves search relevance. As users interact with results, search learns which results work best for your organization. Over time, these popular results will bubble to the top, hence leveraging the collective knowledge of your organization.
• Better result summaries allow users to identify whether they have the correct result more quickly than ever. Summary quality has been improved by providing longer, higher quality textual snippets from the best parts of the document.
• Improved multilingual support for 51 languages, helping people find information across the company in almost any language. Improvements include better coverage and handling of compounds, numbers, and dates. Additionally, an improved multilingual relevance model retrieves and ranks documents more accurately when in a mixed-language environment. Figure 17 SharePoint 2010 Search Results page

Figure 17 SharePoint 2010 Search Results page

Refinement

SharePoint 2010 provides a new way to explore information through search refinements. These refinements are displayed on the left side of the page, to the side of the core search results, and provide people with self-service capabilities to finely filter the search results. SharePoint 2010 automatically determines refinements by using tags and metadata entered and indexed on the search results return set. These refinements include searching by type of content (for example, Web page, document, spreadsheet, presentation), location, author, last modified date, metadata tags, and more.

In addition to refinements, SharePoint 2010 includes several other features that help people explore the result set. Search results include short excerpts, and definitions at the bottom of the list are automatically generated based on the context of the query words. "Did you mean¡¨ suggestions appear to help people specify misspelled and ambiguous queries, and acronyms are expanded to help people find the right content. For example, if a person searches for "ECM" the search results will show "enterprise content management" Results are quickly returned and are available in a browser, on a smart phone device, or through desktop search integration with Windows 7.

People Search

Businesses thrive when people work together, and finding the right people who have specific skills and talents can be challenging in a large company.

Finding People

The People Search experience in SharePoint 2010 is similar to an address book with better name matching, making it easier to find people by name, title, and organizational structure. The name search includes phonetic name matching that will return names similar to the query, and will return variations of common names, including nicknames. For example, if someone searches for "Robert" the results will also show people with the names "Rob" "Bert" "Bob" and "Robby" People can also use wildcards in searches, helping them find the right person.

People Search results feature the same refiners as the core search results. By exploring results through name, title, and expertise, people can quickly browse and select people and expertise. People Search results also include real-time presence through Microsoft Office Communications Server, making it easy to immediately connect with people from search results.

Figure 18 People Search results page

Mining and Discovering Expertise

SharePoint 2010 builds a model of people's regular interactions based on e-mail conversations and group memberships, which it uses to organize people search results based on social distance. People can manually submit or automatically generate a list of colleagues mined from Microsoft Outlook. Automatically generated lists of colleagues are a way of rapidly inferring social relationships throughout the organization, speeding the adoption and usefulness of people search results. SharePoint 2010 also infers expertise¡Xautomatically suggesting topics mined from people's activities (for example, tags, keywords) and Outlook inboxes, and suggesting additions to their expertise profile in their My Site. This makes it easier for people to complete My Site profiles, leading to more people with informative profiles. With complete, visible expertise, companies benefit in both search and communities capabilities.

Improving Search Based on Social Behavior

For many companies, SharePoint sites have become the gathering place where people create, share, and interact with information. Social behavior improves search results in several ways. First, the relevance ranking for people search accounts for social distance. For example, a direct colleague will appear in a search before someone removed by three degrees. Second, SharePoint 2010 supports social tagging and content rating, and this social feedback on content can influence the relevance of content in search results. People's daily use of information in SharePoint and Office applications can have a measurable impact on search relevance, helping companies capitalize on the collective wisdom of people. Finally, search solutions in SharePoint 2010 use click-through on search results for relevance ranking; that is, by identifying that the user clicked a link in the last similar search, that link is more likely to be the best one this time.

Enterprise Scale

SharePoint 2010 helps people add multiple instances of the crawler, indexing, and query components. This level of flexibility means that people can scale SharePoint farms. (Previous versions of SharePoint Server did not allow companies to scale the indexing components).

By scaling components of the search architecture with SharePoint 2010, companies can gain sub-second query times for all queries, regardless of the size of the farm. SharePoint 2010 can also remove bottlenecks that were present in previous versions of SharePoint Server.

Figure 19 Scaled architecture for query servers and indexers

Companies can add multiple indexers to farms to provide availability and to scale to achieve high performance for the indexing process. In a new capability for SharePoint 2010, each indexer can crawl a discrete set of content repositories, so not all indexers need to index the entire corpus. Furthermore, indexers no longer store full copies of the index; they simply crawl content repositories and submit the indexes to query servers.

People can also add multiple query servers to provide availability and to achieve high query performance. Adding multiple query servers is actually implementing index partitioning; each query server maintains a subset of the entire logical index, and therefore does not need to query the entire index (which could be a very large file) for every query. SharePoint 2010 automatically maintains the partitions, using a hash of each crawled document's ID to determine in which partition a document belongs. The indexed data is then propagated to the appropriate query server.

FAST Search Server 2010 for SharePoint

FAST Search Server 2010 for SharePoint offers a new choice in enterprise search. Combining the power of FAST with the simplicity of SharePoint, FAST Search Server 2010 for SharePoint delivers an exceptional intranet and people search experience and a platform for building search-driven applications.

Every organization has unique search requirements. FAST Search Server 2010 for SharePoint can customize the search experience so that it fits the company and its people.

Advanced Content Processing

FAST Search Server 2010 for SharePoint includes extremely powerful content processing based on deep linguistics and text analytics technology. This processing adds tags and structure to unstructured information, creating metadata directly from the content. This gives users the ability to search using terms and language that are familiar and unique to the business.

Most companies frequently use a set of internal names, acronyms, or code words. These words can be confusing to different groups, outsiders, or new members of the company. People will be able to use their own terminology to sort, refine, and query content. Furthermore, advanced language support helps people find content written in its native language.

Figure 20 FAST Search for SharePoint Results Page

Contextual Relevance

FAST Search Server 2010 for SharePoint can give people results that are meaningful and dynamically tailored to their jobs, roles, and functions within the company. For example, the sales teams will be able to quickly find product information, collateral, and answers to RFP questions, while the engineering teams will see specifications and requirements documents at the top of their results sets. Site administrators can tailor search quickly and easily to deliver contextually relevant results the first time.

Tunable Relevance to Meet Diverse Needs

Relevance is central to search, but one size doesn't always fit all. FAST Search Server 2010 for SharePoint provides a highly tunable relevance model to meet the needs of a diverse set of applications. Multiple ranking profiles can be available simultaneously¡Xeven presented as options for people to choose. By extending the default relevance algorithm, site administrators can quickly and easily create and deploy new custom ranking algorithms that are tuned to meet multiple business demands.

Visual Search Experience

With a more visual search experience, people can more easily find content. Most people identify better with information when it's presented visually. FAST Search Server 2010 for SharePoint includes document thumbnails and scrolling previews of PowerPoint presentations, so people can find what they are looking for quickly, without opening the documents. With Visual Best Bets, site administrators can easily add visual banners and applications to highlight essential information.

Figure 21 Highlight essential information with Visual Best Bets

Search-Driven Applications

Companies can drive measurable return on investment with search-driven applications. Search-driven applications help a well-defined set of people accomplish specific business tasks more efficiently. They aggregate information from a defined set of content sources, add structure to unstructured information, and provide a contextual, interactive, and usable experience. Search-driven applications are valuable to all roles, including sales, marketing, research and development (R&D), consulting, support, operations, legal, human resources (HR), IT, and finance. With FAST Search Server 2010 for SharePoint, companies can meet all search application needs across the business.

Figure 22 Example of a sales search-driven application