Thursday, April 24, 2014

Do-they-or-don’t-they? EBSCO Information Services; Part 2

With Ebsco’s new announcement about metadata usage with discovery systems from all vendors, it appears an important step forward has occurred for those libraries and their members/users where discovery products are utilized.  However, as with most announcements of this nature, a close examination of exactly what was said needs to be performed before a firm answer can be determined.

One of the most important things for librarians to realize about this recent Ebsco announcement is that market forces work when they’re utilized.   Ebsco was clearly rushing this announcement to completion in order to beat a deadline; I specifically suspect an order cancellation deadline.  I personally heard from three Ebsco team members in one day, concerning this pending announcement.  Given the public dismay expressed by the Orbis-Cascade Consortium towards Ebsco and Ex Libris, one might suspect that the consortium, or another large customer like them, made it clear that their Ebsco content renewal would not be happening unless progress was made on this front. 

The next most important thing to realize is that this deal does NOT fully meet the demands of librarians.  As I, and others, have openly and repeatedly said: What we want is for any EBSCO content we license to be searchable under the Discovery interface of our choice.  That is not what we got.  What we received was at best 70% of our request, somewhere around 120+ databases total.  There are some very important products in that list, but equally notable is the exclusion of ABC-CLIO, CINAHL, the Wilson Indexes and others where Ebsco has done a great deal of work in enhancing the searching of databases through the use of subject terms.  We did get the metadata, abstracts and (Ebsco’s words here) "full-text where possible". The reason I note the phrasing “where possible” is that many content suppliers often try to negotiate exclusive agreements with the publishers.  If they’ve done this and then cite the exclusivity as a reason they can’t supply us with the full-text to index, that would be self-serving at best.  I asked this question and was told we should compare what other providers, such as ProQuest are doing to see if in fact the offering is different.  If so, then EBSCO claims they will seek to modify their agreements with the publishers.  We shall see.  

In fairness to all concerned, Ebsco is a business and not unlike other businesses, must please owners/shareholders as well as customers.  Growing value by preserving the unique and differentiating features of products is a tried-and-proven business strategy.  Of course, Ex Libris and ProQuest are also businesses trying to do the same.  Consequently, and at the same time, all parties will be trying to negate the unique and differentiating value of their competitors while at the same time, maintaining their own.  It’s just business.  So, you have the classic compete/cooperate conundrum.  When dealing with a direct competitor, such as these companies will be doing under this metadata policy, it requires a great deal of trust to flow in both directions.  As you will know from reading my previous post on this subject, trust is something that is in exceedingly short supply between these companies.  To restore that will require small steps with plenty of stop-and-evaluate moments along the way.  The Ebsco representative I spoke with clearly stated that depending upon the usage of the Ebsco data under these discovery interfaces, they might revisit this policy in the future.  Note however this wording in the release: “As a content provider, we want to work with all discovery vendors in a way that encourages and requires mutual sharing. Because all major discovery vendors are also ILS vendors, content providers, or both – there exists a logical "two-way street" and opportunity for collaboration on behalf of our mutual customers.”   There are a LOT of exit ramps in that wording.  Caveat Emptor.  Plus, it’s important to note that Ebsco has the sole distinction in the field of being the only metadata supplier to have ever withdrawn their metadata from Ex Libris' Primo Central. And if it happened once….   

The policy Ebsco issued also indicates that the databases being withheld from this policy might be included at a later date.  From the policy:  “A&I resources are developed with specialized components and remain our industry’s most sophisticated data sets, but as such, require intricate, refined search algorithms and approaches to properly leverage their value. At present, not all discovery services are designed to leverage the specialized components of these individual collections, and as a result, are likely to inadvertently de-value and subsequently harm library research were these to be included without proper guidelines. As approaches for discovery service development around A&I resources are more thoroughly documented (with the help of Ebsco and possibly other subject index providers) and subsequently addressed, Ebsco will re-visit its policy for sharing these unique databases with other discovery service vendors.”  See my comments about product differentiation above in evaluating these statements.  What is somewhat frustrating about this however is that the profession and industry have largely come together under the umbrella of NISO and have formed a working group called the “Open Discovery Initiative” to deal with these very kinds of topics.  Ebsco is a member but appears to not totally buy into, nor support, the recommendations of the group.   For the profession of Librarianship, this is frustrating.  From a business perspective, it’s understandable in the short-term, but in the long-term these kinds of decisions tend to hurt the profession overall, thus diminishing the overall market size and thus the number of potential buyers (libraries).  In other words, this is shortsighted on Ebsco’s part in my opinion.

There are a number of answers that Librarians must continue to pursue to properly evaluate this development in the marketplace.  They include:


  1. Bias and Known Item Searching.   There was a very interesting article on Discovery systems in this weeks Chronicle of Higher Education that reinforces a concern I’ve expressed numerous times in this blog about bias in results.   This is a massively complex question and the Chronicle article will give you some flavor of the complexity surrounding the bias issue.  Known-item searching is frequently a point of frustration with many Discovery users because the single-search box approach is left to try and sort through data elements and properly recognize things like author name from words in the title and anything else the searcher might throw into the broad search.   EBSCO’s content and subject enhancements and products like Ex Libris’s Primo can, and frequently do, bump heads here.  Primo has had numerous enhancements to improve known item searching.  But it is likely those enhancements when combined with EBSCO’s subject enhanced metadata, might not produce the prioritization that EBSCO would like to see of their content.  This is in part why EBSCO should provide more compliance with the NISO ODI recommendations, but as I pointed out above, to do so would diminish a key product differentiator, so it seems unlikely to happen.   The result, as the subject of the Chronicle article noted, is that many discovery users just give up and use Google Scholar.  Clearly, that is very unfortunate for librarianship.
  2. Library-by-Library Analysis.  Each library will need to do an analysis of what EBSCO content they subscribe to and pay for and what will be included under this policy such that it will be discoverable with the library’s chosen discovery tool.   Overall utilization of the EBSCO content should also be factored into this model.  One needs to determine a per/use cost and keep careful track of the ratios of those discoverable within the discovery tools and those that require use of a native interface.  It will remain a library decision if that utilization cost is worthwhile for the library members/users.   Clearly, renewal decisions should utilize this analysis.

To wrap this up; was an important step forward achieved?  Possibly.  It’s a better policy than was previously used by Ebsco.  Still, there are a number of caveats in the wording, and in what was verbally said, which could result in this policy suddenly being retracted.  That’s a continuing cause for concern.

The bottom line is that as librarians, we must continue to use our purchasing power, combined when necessary, to collectively and where appropriate, exert pressure on content suppliers to provide what we need to make for meaningful experiences with our discovery tools.  As I said above, this looks like progress, but it is not all that we need. So we must continue our efforts, until such time as we can search all content we license, under the discovery system of our choice.