Which Search System?
Selecting a search system is not a stress-free job. Most professionals know that entering a word or phrase in Google’s public search engine delivers on-point results almost every time.
Most professionals see search as a “no brainer.” The reasoning takes Google’s success and jumps to the hunch that search is “not a problem” or “search works now.”
The fact is that search is one of the most complex problems facing computer scientists and information retrieval experts. Google works because it looks at what’s popular using about 100 criteria in a very large user community.
In organizations, there are different considerations. Consider the type of search system chemical engineers require. Chemists search by specifying structures. Words don’t work too well for chemical structures. (See a sample of typical search requirement differences in small, mid-sized, and large organizations with representative search systems identified for a client.)
Consider the type of search needed by a customer service representative. A mix of information about the customer and information germane to the problem the customer has must be searchable. In most customer service applications, search is one of the most used and least liked subsystems. When a person without access to the customer service data needs access, most search systems struggle to federate information efficiently.
Two years ago, licensing a search system from one of the Big Four in enterprise search was a safe bet. (The Big Four are the publicly-traded companies offering enterprise search systems. The Big Four includes as of January 2005 Autonomy, Convera, FAST Search & Transfer, and Verity. At this time, less than two percent of Google’s revenues come from its enterprise search “appliance.” Companies such as Microsoft and Oracle offer search tools, but these are “baked into” their products. Other companies such as Hummingbird PC Docs and OpenText embed search in other enterprise systems, having learned that licensing a stand alone search system is a difficult process for the company and the customer.
Arnold IT can help. Our professionals can assist organizations in identifying the requirements for a search system. There are search systems that handle certain problems better than others. Need a better search system for a large Web site used by employees and customers? Consider Thunderstone’s “appliance.” It costs less than Google and provides security and operational controls better suited to an enterprise environment than Google Appliance’s system.
Do you need to retrieve information from an Oracle, SQLServer, DB2, or Informix database without buying more hardware and adding external software? Consider Speed of Mind’s snap-in technology. In a matter of hours, performance can improve by several orders of magnitude or more. More importantly, no additional hardware is needed to revivify a baked in database search engine.
Want to operate an enterprise search system without running it locally? Consider Blossom Software’s managed service system. Developed by a former Bell Laboratories engineer, Blossom provides most of the search functionality at a fraction of the cost of most enterprise search systems with bullet-proof reliability.
In addition to a thorough knowledge of enterprise search systems, Arnold IT can collect information about the types of information that must be indexed and made searchable, the interface and functionality needed by users, and the technical environment at an organization in order to suggest specific systems best suited to that organization and its system professionals.
Avoid selecting a search engine without having dependable requirements or an informed third party validate the requirements. Making an error in a search system can cost an organization six figures or more to rectify. Arnold IT does not sell search systems; its professionals provide objective counsel about search systems.
We know that these questions can be difficult to ask and to answer objectively. Arnold IT provides an unbiased, objective resource to assist our clients in framing a search project in a pragmatic way.