Gartner advocates cost cutting for enterprise search projects

1 min read

Gartner says large manufacturers could slash $250,000 off the cost of enterprise search licenses by using federation or alternative delivery compromises.

The analyst says it's all about avoiding or reducing the costs of a full search platform installation. "Enterprise architects, information architects and IT managers are frequently instructed to make web sites, intranets, extranets and portals, as well as document libraries and mainframe data stores, as simple as possible. [But] extending such capability across unmanaged data and content stores is often expensive to implement and rarely includes extra setup funding," explains Whit Andrews, vice president and analyst at Gartner. He suggests that companies intending to acquire enterprise search technology, to roll out a universal platform for search, can postpone or eliminate their plans and save money with a tactical alternative – such as federation, application service providers (ASPs) or inexpensive software and appliances. Companies using existing, as well as new, tactical search engines as target federation points will save money by reducing redundant indices, which cuts storage costs, he asserts. The caveat: "While federation can help save money, there are some disadvantages. The costs of such a strategy are reduced success in the relevance of searched information and, in some cases, increased load on subordinate search engines," he warns. "Particular concepts or document locations dominate the results returned to many queries, which frustrates users and makes it difficult to discern the proper document in a results set. Additionally, unless queries are pre-processed before they are handed off to federation, they can overload their target engines," he adds. Gartner's recommendation: search delivery alternatives, such as ASPs and appliances, are viable ways of addressing the need for search capability in places such as web sites, intranets and file servers. "Vendors increasingly offer an ASP model for search engines, and enterprises requiring searchable external web sites will find such models more predictable in relation to capacity versus expenditure," suggests Andrews.