Recently, I asked a list of colleagues for their thoughts on building the business case for search, since a client was trying to get funding for a project. The basic response was, "Business case? We scoff at such silly questions. Search is infrastructure. How would you make a business case for email these days?" I thought to myself, "Am I miss-ing something?"

In my opinion, the infrastructure response reflects the fundamental problem with how people are conceptualizing enterprise search and how it needs to be designed, built, and deployed. Search can certainly be considered infrastructure, but it still requires development and configuration to become truly useful. Content management systems and document management tools are other technologies that can be considered infrastructure, but almost nobody would just install them out of the box and expect magic. It is the appli­cation of search infrastructure to business problems that requires the same rigor that any application requires.

A successful approach to search needs to factor inintegration of content tagging, best bets, faceted search, investment in federated search, and so on. You can have the basic infrastructure, but to get real value, additional investments need to be made in operations and technology.

What is "Enterprise Search" Anyway?

There are misconceptions about the nature of enterprise search and its impact in the organization. As enterprise information has grown more complex and information sources more varied, users have grown to expect a similar experience to searching on the internet. In essence, users are looking for a "Google experience," which is not a realistic expectation, as web search functions are fundamentally differ­ently than enterprise search.

At their core, search systems try to make sense of a user’s intent—using only those short, ambiguous search terms that people put into the search boxes—and present the content that represents something meaningful to the user. Enterprise search adds to this scenario a variety of information sources and contexts. Imagine I asked you for information and only gave you one- or two-word clues,say "deliverable" or "proposal," and expected you to come back with something meaningful. You’d probably want a little more of a clue: "Tell me more about the types of things you need. What are you doing? Who is this for? Who are you,for that matter?"

If we know something about our users and what they are trying to accomplish, we can do more to present more meaningful search results. Through planning and consid­eration of user needs, we can give people the ability to discover more specific, more relevant, and more valued content than if we just plug search in and consider it infrastructure.

Add to the mix the current trend toward using search as a mechanism for surfacing structured data from business intelligence and transactional systems. In these emerging applications, unstructured queries are being mapped to data from ERP systems, sales management tools, accounting systems, and other nontraditional "search" resources. Therefore, the value of enterprise search lies in its ability to provide simple, intuitive ways to consolidate information sources, whether structured or unstructured, and surface information assets in the correct context and at the right time.