Setting Expectations for a Search & Discovery Solution

Most of us have used Google search on the Internet. But many have not yet experienced a “Search & Discovery Solution.” I recently described SDS as a special version of semantic search technology focused on supporting product development in engineering and manufacturing enterprises. An SDS helps users find product information scattered throughout an enterprise's Intranet-based information systems.

Consider this column an alert: Don't be misled by thinking about Google methods as a point of departure to understanding, evaluating and selecting an SDS.

Rethinking expectations is one of the first steps to selecting an SDS that meets your company's specific needs. Yes, entry of search criteria into Google is similar. With Google's new Hummingbird algorithm, semantic capabilities also are used to enable searching on the intent of the entire search phrase, not just individual keywords. And users still should expect relevant results to be returned immediately.

Yet many other capabilities of an SDS are different; they must be to adequately support product development processes. I'll focus on just one difference, for now: Relevancy. It determines the desired search result, leading to the “a ha” moment. Arguably, it should be a highly rated selection criteria.

In SDS software, two dynamics help to determine relevancy: 1) SDS algorithms and 2) user actions.

SDS algorithms determine the ranking of the results presented to the user by logic, such as the frequency and proximity of the criteria to other words. Relevant ranking methods can range from simple to sophisticated logic.

The second dynamic is the ability of a user to determine the appropriate choices for further discovery, and ultimately, timely decision-making. In addition to the displayed results, a user will see filtering choices to narrow subsequent searches, including dates, classifications and file types. Further exploration of the initial results is very likely; a typical initial search may reveal hundreds of choices.

Certainly, there are additional capabilities of an SDS that must be understood, e.g., indexing methods that provide speedy search and discovery of desired data, connectors and security provisions, among many.

But, above all, bear in mind that Google may have swayed some user expectations with its methods. In contrast, expectations for an SDS must be set to meet significant enterprise needs. For instance, availability of connectors to crack open and look inside CAD files.

Reference
Search . . .and find from the September issue of Mechanical Engineering , the magazine of the American Society of Mechanical Engineers.