click to search this site
Google+
 
 
 
Google Search Appliance
Google Search Appliance

Google Search Appliance Case Study


This global manufacturer required an Enterprise Search Solution for their internal Sales & Marketing Portal.

BACKGROUND AND BUSINESS OBJECTIVES
Like all Fortune 500 companies, this customer has a vast store of corporate knowledge housed in multiple repositories, and accessed through a variety of intranet portals.

The Sales and Marketing organization needed easier access to research developed by an internal business intelligence group, as well as to other sales support materials scattered around several content repositories. As part of an overall corporate initiative to leverage research and other intellectual capital to improve corporate performance, the customer chose Search Technologies to help implement the first phase of an enterprise-wide search system.  

The primary project goal was to provide Sales and Marketing employees with an easier way to find consumer and competitive research, information on sales trends including seasonal adjustments, plus templates and worksheets used in campaigns. A new sales portal would enable these employees to easily find all of these materials from a single search box, regardless of where the information was stored.

The project would also demonstrate the value of enterprise search as a cornerstone of an overall corporate knowledge management strategy.

CHALLENGES AND REQUIREMENTS
The customer faced several challenges in the development of a sales portal:
  • Search was being underused across the company, in part because search results were perceived as being mediocre. Search had a poor reputation.
  • Sales personnel needed to go to 10 different sites or portals to search for the information they needed. And, they had no access at all to the important intellectual capital created by the Research Group.

A lack of trustworthy document-level security was a key factor. In such a competitive industry, preventing the leakage of intellectual property is a priority, and this was getting in the way of enterprise search.

In addition to these overall search challenges, the customer had some specific requirements that a new search capability should fulfil:
  • Content needed to be extracted from several different content sources, including WordPress, file systems, and a new content management system (Alfresco, with a Liferay portal)
  • Search navigators were highly desired by users, although very little metadata existed to drive them
  • Many documents were very large, including PowerPoint presentations and highly technical PDF files.
  • Support for a search spell-checker and a “did you mean” function were seen as important
The customer previously facilitated internal research and stakeholder surveys to compile their list of requirements. A comprehensive set of wire-frame designs had also been developed. These helped to articulate the company’s vision for its "ideal search system".

SEARCH TECHNOLOGIES’ SOLUTION

Armed with these requirements, and aided by further, more detailed requirements gathered through a Search Assessment, the Search Technologies team helped this customer to develop an implementation strategy.

The content sources to be indexed were approached individually. This included taking full account of document-level security requirements. These varied depending on the data sources, and the variations introduced complexities.

End user features were discussed. For example:  query auto-completion; "best bets"; stemming; results biasing; and search navigators.

Taxonomy development was also discussed, and a plan was made. Metadata was lacking in the content sources. Auto-generation of new metadata, to a consistently high standard, was recognized as a foundation for better search.

In discussions with the customer, it was decided that the Google Search Appliance, working together with Search Technologies' Aspire Content Processing platform, provided a full set of functionality for the project.

Using the Aspire, Search Technologies developed custom connectors to crawl documents from Alfresco, Liferay, and WordPress. The crawled documents are filtered to determine which ones are new, changed, or deleted. Updates are then passed to the Google Connector Framework. This combination is much more efficient than repository crawling. It ensures that the GSA's search indexes are updated as content changes, rather than having to wait for a daily crawl of the repository.  This approach mitigated a major security concern, and helped the customer's search team to build a close relationship with important repository owners.

DEPLOYMENT
The system was delivered on time, despite a number of unexpected challenges. The solution included:
  • Aspire-based connectors for Alfresco, Liferay, and Wordpress
  • Normalizing LDAP Proxy/ LDAP Group Expansion capabilities into a single security resource for search purposes. This addressed a specific and problematic feature of the customer's overall security infrastructure

CURRENT STATUS

The customer’s Sales and Marketing team now has easy access to all of the research, trend data, and supporting materials they need, through a simple query interface that looks just like Google.com. The underlying complexities involving multiple data repositories and heavy document-level security requirements are taken care of automatically. Users experience a simple yet highly productive search environment.

The information they find is up-to-date and relevant. Users are also provided with contextual navigation options, enabling them to browse the available information through Dynamic Navigation.This fits a common scenario, in which the user is not exactly sure what they are looking for. Now they have the entire content set available to explore (subject to their permissions) through a combination of search and browse.

Search Technologies also provided training, knowledge transfer, and general guidance to the customer's staff who today, are managing the implementation efficiently, and maintaining search excellence for the benefit of their users.

COMMENTARY
This customer engagement exemplifies a number of common themes, which can be summarized as follows:
  • A key to the project was getting repository owners, who were under huge pressure to prevent IP leakage and protect sensitive data, to get to know and trust both the search team, and the technologies that were being proposed. Security was complex, but this can be fully addressed through the application of best practices and customization
  • There is an emotional aspect to security too, which this project was able to address
  • Diligent content processing, to prepare data for indexing, through enrichment and normalization, and to address security issues, helps the search engine to perform at its best. Combine this content preparation with the core relevancy and tuneability of the GSA, and the result is a highly efficient search experience. "Just like Google."
  • The GSA's built-in ease of administration, together with an overall approach that identified and mitigated potential complexities, ensures that this customer runs a great search system at a low overall cost-of-ownership

GSA Data Connectors
Content Processing