Printer Friendly Version

Microsoft's SharePoint Portal Server 2003 was sold into a large number of organizations based solely on the strength of the search tool. Organizations hungered for a way to find the data they had generated.

Structured data such as invoices, products, and shipments may have been easy to find in the applications designed for that data, but the growing mountain of documents seemed to make the unstructured information that you were looking for perpetually out of reach.

Search in SharePoint made significant progress in its ability to connect users with the unstructured information that they were seeking. But the effectiveness of searches depended upon the skill of the searcher and the alignment of the terms that the searcher used to the terms in the documents. The world of search analytics was still very foreign to most organizations. Thankfully, the next version of SharePoint Search with its focus on relevancy will also include reports that allow you to see the effectiveness of the searches users are executing.

Microsoft Office SharePoint Server 2007 is a part of the Office System and is set to debut sometime in late 2006 or early 2007. Microsoft Office SharePoint Server 2007 includes numerous enhancements designed to improve search relevance, Internet usage, content management scenarios, and many other features which were shared this week at the SharePoint Conference in Bellevue, Wash., this week.

In this article you'll learn about the basics of search analytics, what you can do today to improve your search results, and what to expect in Microsoft Office SharePoint Server 2007.

Defining Search Analytics

When taken one at a time, we all know that we get different results to a search. One possible outcome is that the search will yield no results -- a condition that almost universally means that the search was not successful. On the other end of the spectrum is a search that returns a large number of links -- perhaps as many links as there are within the organization. In a more relevant search there may be a handful of results where one of those documents does indeed include the information that the user was searching for.

While there are different experiences for the user for each of these results, in SharePoint 2003 there was no way for the administrator to aggregate these experiences and see trends in what people are searching for, what searches appeared to be successful, and how many searches were not successful. That is the role of search analytics, to capture and report on the results of searches in aggregate so that changes can be made to the way that search operates, or what content is on the site to improve the relevancy of the content for the user.

Changing the Results

In SharePoint, there are three primary ways that you can control the results that the users see in the search results. These are the three items that you'll be manipulating when you look at the results of the analytics:

• Noise Words -- The noise words are those words that are eliminated from the index. Words like the name of our organization, your primary industry, and commonly used terms should be included here so that they are not included in your index or used to prioritize the documents that are seen.

• Thesaurus -- The job of the thesaurus is to convert one word into another word. So for instance, "seat" may be converted to "chair." They are similar conceptually. However, the search engine has no way of knowing that seat may also mean chair. Conversions like this allow the search to go beyond the specific terms that were entered and include words that mean the same, or nearly the same, thing. The thesaurus can also be used to correct common misspellings so that even if the term is incorrectly entered into a search text box it can find useful content. The thesaurus is most helpful in reducing the number of queries that return zero results. It broadens the search results.

• Best Bets -- SharePoint allows you to specify what documents are the most likely documents to be needed based on the searched-for words. So for instance, you could map the company's employee handbook to the word "handbook" in the search box. This allows you to highlight important and likely search results to reduce the impact of searches that contain many results.

Analytics Today and Tomorrow

If you want to do search analytics with SharePoint 2003 you have to purchase a package like MondoSoft's Behavior Tracker product. This tool allows you to see the most frequent searches, the searches with no results, top not-clicked items, and other reports. In addition, Behavior Tracker also offers higher-level analytics like a performance and expectation map and recommendations on optimizations.

Microsoft Office SharePoint Server 2007 will offer a small subset of the most frequently used reports including: top searches with no results and the most frequently searched terms. This will allow every customer to start to improve the search results on their portal.

Printer Friendly Version

Other Resources

from Intranet Journal

Intranet Journal Discussion Board An Introduction to Microsoft SharePoint Portal Server How SharePoint and SOA Fit Together

from JupiterWeb

Oracle's Collaboration Suite Aims at Sharepoint (InternetNews.com )

from the Web

Microsoft SharePoint