How to Improve SharePoint Search with Best Bets in SharePoint 2010

Summary

SharePoint has become an integral part of many enterprise content management solutions, with a reported adoption rate of 78% percent of Fortune 500 companies as of this writing. The software is attractive due to its clean UI, collaboration tools, seamless integration with Microsoft Office, and the ability it gives IT administrators to deploy and securely manage intranet, extranet and Internet sites from one centralized platform.

The vast amount of content a typical SharePoint site accumulates necessitates a robust search capability, which SharePoint provides. The expectations of users performing searches in an enterprise context, where relevant documents are expected to populate the top of results every time, rather than merely hoped for as in a general web search, can result in a poor user experience and abandonment of search when those expectations are not met. Improving the search results using metadata and Best Bets can reverse this and allow relevant content to be displayed and, therefore, used.

This white paper provides an overview of Best Bets and how SharePoint administrators might increase user adoption and organizational efficiency by implementing them to improve the search experience. It will also discuss how augmenting Best Bets with Ontolica from www.surfray.com can further boost user engagement and search performance through advanced search functionality, such as Visual Best Bets, Audience-Targeted Best Bets, and Time-Managed Best Bets.

Best Bets Defined

Best Bets can be defined simply as promoted search results within SharePoint. Administrators are able to select certain items to appear at the top of the search results page, marked with a yellow star, whenever certain queries are performed by users. Best Bets, in effect, bypass the normal search engine algorithm’s assignment of relevance and value to a particular piece of content in order to make it rank highest for the selected keywords. The value in using Best Bets is that they allow the appropriate content to be displayed even when that content has characteristics that would normally bury it in search results.

The Four Behaviors of Search Users

Understanding what typical users do when presented with search results helps IT managers understand the benefits of Best Bets and how to work with them effectively. This behavior can be broken down into four general categories:

Success. The user performs a query, gets results and selects a page that contains the information they want.

Partial Success. The user performs a query, gets results makes a selection, but the information does not quite match what they are looking for. They click “back” and attempt another query.

Unsuccessful with Results. The user searches, but sees nothing they like on the results page so they don’t click on anything. They search again or leave.

No Results. The user searches, gets zero results delivered by the search engine, and either tries another query or gives up and leaves.

The bottom two behaviors are what should concern IT managers, assuming the search queries are not misspelled or have zero relevance to the content hosted on the platform. When analytics reveal these user behaviors, they indicate the search engine is not delivering the expected results because they are either irrelevant or completely missing. If there is content that should match the queries, but this content is not being found as it would require several paginations to appear, Best Bets can solve this and reduce these unwanted user behaviors.

How to Use Best Bets to Make Search More Relevant and Rewarding

To create a Best Bet, site managers first enter the keywords they want to target. Then, the URL the Best Bet will to point to is input, along with a title and description if desired. Keyword synonyms which will also trigger the Best Bet are another option-one that is highly recommended as this can catch all permutations of a search query, and therefore deliver better search results.

How many Best Bets should be made for each site initially? The answer is: it depends. A range of 20 to 50 seems reasonable depending on the size of the organization and the amount of content it hosts. Site managers should consult with stakeholders to identify mission-critical documents that would best serve the majority of searchers with roles along specific business tasks. To that end, the first implementation of Best Bets should probably not focus on long-tail subjects but on documents that might not be inherently optimized to show up in results, but that would solve business problems for a substantial number of employees.

Let’s take an example from a fictional company that produces widgets. A new hire in the sales department seeks to learn some basics as to how they are produced in order to understand and sell his product a bit better and, to do this, turns to his company’s intranet that is run on a SharePoint installation. He types in his query, and is inundated with results about the history of the widget company, widget marketing plans, technical widget manufacturing specifications, and so forth. The information either just doesn’t address how they are made, or can be understood only by an engineer.

Rewording the query in several ways doesn’t improve results. Growing frustrated, he abandons search after some time and turns to one of his co-workers to ask, “So how do we make these things, anyway?” This pulls another employee off of production, and the information provided might not even be accurate or conflicts with company policy regarding dissemination of its internal procedures according to specific roles. The new hire would have been better served had he simply examined a flowchart called “How We Make our Widgets: A Basic Primer” drawn up some time ago that was scanned and stored on his company’s server as a JPEG.

The problem here is that the document might have shown up on page 10 of the search results for the new hire, as even with metadata, the JPEG simply contains no writing able to be parsed by the search engine, and therefore, is not judged as very relevant. Any documents that even casually mention widgets trump the JPEG in importance in the eyes of the search engine, even though it has the correct information that any user should have if they want to know how make widgets.

Best Bets would allow an administrator to solve the case above by forcing the flowchart JPEG to show at the top of the results page whenever someone types “how to make widgets” or similar search terms that are entered as synonyms in the Best Bet setup. This sort of customization in search results will improve the user experience and increase employee confidence that the system is a valuable resource to find the information they need to do their jobs, increasing adoption and allowing managers to better control the flow of data to both internal and external publics.

Using SharePoint Site Collection Web Analytics and Benchmarking to Improve Search

As mentioned earlier, search user behavior is driven by the results delivered from their queries. This behavior, particularly the ones resulting from the “Unsuccessful with Results” and “No Results” categories, can be used to refine and improve enterprise search functions through SharePoint’s Site Collection Web Analytics, which are accessed through the Site Settings menu item.

For our purposes, the metric that should be focused on is “Failed Queries.” This will illuminate exactly where the biggest search problems occur by displaying a list of up to 2,000 failed query texts, sorted by quantity, and the percentage of search results that are abandoned for each, or whether no results were displayed for them at all.

A value of 90% in the “Percentage Abandoned” column indicates that only 10% of queries resulted in a click-through to one of the displayed results. A “100%” in the column would mean that no click-throughs occurred on the options presented in the results page, and “No results” in the column means that apparently no content exists anywhere on the site that matches the search text.

These metric results are abysmal, assuming the content actually exists on your farm somewhere-and if the query is being executed by many users, it probably does (or should be created). The high failure rate reflects a disconnect between what users expect and what the search engine is delivering, so this search string is a prime candidate for a Best Bet to deliver the most suitable document to future searchers.

Once created, Best Bets results should be monitored regularly and performance benchmarked against earlier time periods to ensure they are being used and still relevant based on the enterprise’s current strategy. The Best Bets Usage report will spotlight successful ones and also indicate which ones should be refined through non-use. The report can be generated and automatically e-mailed at pre-defined intervals, such as once per week.

Advanced Search Functionality and Tools

Site managers can further improve the search experience using tools such as FAST Search Server, an add-on produced by Microsoft to extend SharePoint’s search capabilities. Some of the enhancements include:

• Auto-suggestions as users type their queries

• Page thumbnails in search results

• Related search suggestions link

• Advanced filtering tools

Other advanced search functionalities include:

Visual Best Bets. Third-party solutions like Surfray’s Ontolica allow you to expand the functionality of Best Bets by adding a picture, a much larger description that can use html tags to improve formatting, and produce a thumbnail for all the search results, not just the Best Bet. This makes the results page more engaging and helps the user pick the content they are looking for.

Audience-Targeted Best Bets. These allow you to restrict search results to certain groups previously set up in SharePoint’s User Profile Service in order to provide a contextual search experience based on position, geographical location, or other customized groups. You can set up Best Bets to show different content for sales personnel versus support personnel, for example.

Time-Managed Best Bests. You can set Best Bets to run from a specific start date to end date. This will help keep them relevant and able to tie in with certain strategic events, such as a specific marketing push.

Conclusion

In this white paper, we have provided a high-level overview of some of the ways Best Bets can be used by SharePoint administrators to enhance the search experience for enterprise users. Sites should implement at least 20 to 50 Best Bets, concentrated in most mission-critical areas, which use complete titles, description, and keywords to maximize performance. Implementing advanced tools such as Visual, Audience-Targeted, and Time-Managed Best Bets can further improve search efficiency and user satisfaction.

Ontolica is a third-party solution suite that provides enterprise site administrators with robust customization options for SharePoint, as well as out-of-the-box enhancements that improve the search experience without extensive coding. For videos with tips and tutorials on improving SharePoint search, visit http://www.surfray.com/videos-and-demos.html or visit the blog at http://www.surfray.com/resources/tech-blog.html.

Surfray were signed up as Silver Exhibitors for the ESP Conference 2013. Contact aoife@sharepointeurope.com or call +353 91 514501 to learn more about sponsorship & exhibition opportunities.

Stay tuned for more SharePoint content by joining our community or by following us on twitter or facebook.

Share this on...

Rate this Post:

Share: