The primary feature of Universal Search is user convenience in the elimination of redundant search queries of content holdings on multiple platforms. Users can simultaneously search all of their institutions’ subscribed content wherever it resides, whether it’s available on OvidSP, free sites such as PubMed, or other subscription platforms. The system uses one search query to search OvidSP databases and up to 10 web-based databases. Through conversations with users Ovid determined that 10 external sources would satisfy most users’ needs. Ovid’s customers are in medical and scientific fields and can accommodate most of their needs with Ovid’s databases and some web sources.

The OvidSP results are integrated, deduped, and ranked according to a menu of preferences. Results of external searches are returned one at a time in tabbed form with one tab for each source. The results are ranked by the source provider.

Users can employ OvidSP’s search and workflow tools to refine, manage, and save results retrieved through Ovid Universal Search. These tools include natural language searching, Boolean search terms, and searching across content types, such as books, journals, and databases, at the same time. The ability to search different content types with one search represents significant value added.

I asked Carl Gustafson, vice president of technology, medical research, why results from external searches are not federated with OvidSP results. He said, "We’re aware that one of the real limits of federated search is that you are completely dependent upon how your query is being serviced by the external provider. For true federated search results you are at the mercy of the slowest source." Gustafson added, "We are aware of the need for a good, responsive federated search tool." There are other problems as well in federating search results. (For more information, see my article, "Federated Search: One Simple Query or Simply Wishful Thinking?" Searcher, July/August 2008.)

Universal Search is a collaborative effort with MuseGlobal (www.museglobal.com). In order to learn more about how the system works, I talked with Kate Noerr, chairman and CEO, and Jed Gilmore, senior vice president, of MuseGlobal. Gilmore explained how queries to external sources are processed. "As the query comes in, we first normalize it such that it is understood equivalently by all the selected targets. We may include other attributes, limiters, or commands such as a ranking order. Most typically, clients will not specify a ranking order in a default query and simply take the order as ranked by the targets. We normalize the records as they come back so that they will match a display format." The interface and format are specified by Ovid in accordance with its customers’ requirements.

Gustafson indicated that Ovid is committed to meeting its client needs in terms of contextual inquiry and design. This is an effort to tailor the system to the needs of Ovid’s users. "We actively watch clients work, watch the workflow, and watch the workarounds they use to see the intent of what they are doing. We’ve done several hundred of these observations and interviews in the last year and a half."

Universal Search was introduced on Oct. 2, 2008. Ovid is in the process of setting up trials for the product, so there were no users who could be interviewed.

Regarding future enhancements, Gustafson points out that Ovid views the current system as a foundation. It is building on this foundation and will release more features over time. He added that Ovid is committed to providing the most accurate set and suite of results they can. Developments at Ovid are worth watching. The company’s client context-based approach, attention to workflow, and presentation of results can result in systems that contribute to user convenience and productivity.