Beyond Federated Searching: "Discovery Searching" and Its Application to College Libraries

Gentle readers,

As noted in my post about SCELC Vendor Day, I attended a presentation by Serials Solutions on their new product Summon.  The following day (March 4), I saw a demonstration of the Voyager integrated library system (ILS) sold by ExLibris Group. The Voyager package includes Primo, another "all-in-one" library search product.  Both packages (and their competitors) allow library users to search book catalog records, serials databases (such as EBSCO and ProQuest) and electronic book subscriptions with simple, Google-like searches. 

At first, I was inclined to say, ho-hum.  What's the difference between this and various federated search products offered by EBSCO, Reference Universe, LibRef and others?  An example of federated searching can be found at this page    Notice the detailed, confusing explanation of how a patron can search five databases at once, integrate results, etc. (This is not meant to be disrespectful of a particular library, only as an illustration).  Federated searching is geeky and confusing.  Little wonder that one librarian after the Voyager demonstration told me that a previous employer had pulled the plug on a federated search product shortly after its inauguration.  User reaction was a mixture of "huh?" and "so what?"

Vendor claims and cant aside, Summon, Primo and similar products offer something new:  the ability to enter one simple Google-ish search, access an entire collection and output meaningful results.  The format of what's found is clearly labeled:  book, e-book, scholarly article, newspaper article, etc., as well as its publication provenance.  Which database it's taken from is not shown...and the user doesn't care.  Students are typically told to find a certain number of items from each format and to suss out information from reputable organizations.  The fact that a particular article came from a particular aggregator is completely irrelevant to the research process.  The "payload" is what matters, not the "delivery system".  I, for one, was impressed with the results I saw from real-life libraries using "discovery" search systems (ASU, for example).

As with all new technologies, this one will have glitches.  Integrating all these resources into one search engine-like box and getting consistent, high-quality results is a daunting task.  However...

I submit, fellow librarians and researchers, that something like these "discovery" products must happen in Libraryland...and soon.   For centuries, we have been clinging to a search model that still depends largely on author, title and subject.  Keyword searching, as useful as it can be, is merely a refinement of "ATS".  All around us, the world has changed.  Students can, and will, Google their way to what they think they need.  Behavioral analysis shows that most students will go for a maximum of three clicks.  If they don't find what they need, they move on to another source.  (This assumes that they bother looking at library resources to begin with, or at all.)  By using modern search engines and algorithms for library collections, we can provide something new, different and useful in our services.  If we don't, we will surely become increasingly irrelevant to our student users.

After hearing the speaker from Serials Solutions, I (rhetorically) asked him: "so, you mean students are never going to learn Boolean searching, are they?"  We shared a good laugh about that.  I didn't tell him this, but when I've seen the much-hyped "new" ProQuest search engine, my reaction was: "so, what?"  It's still based on Boolean searching, for pity's sake.  News flash:  students don't care about ANDs, ORs and NOTs and never will.  Here's another:  in the 21st Century, there's no reason why they should.

As my readers know, this blog has heretofore mostly been my "travelogue" for a sabbatical.  Now that I've opened this can of worms, I hope I may actually get some comments back.  Let's hear from you.  Best/Tom