Last week the UT Arlington Libraries Access & Discovery staff received an email from SerialsSolutions, relaying the news that Web of Science would no longer be indexed in Summon, our discovery service. This post provides a bit of background to and opinions about that announcement.
During the Charleston Conference, an annual conference aimed at those in the collection development and acquisitions industry, Thomson Reuters and Google Scholar announced that they had made a deal to link their content: Web of Science citations would show up in Google Scholar with links directly to Web of Science, and researchers could click on a link from Web of Science to see a search on the item from within Google Scholar. This news came a few days before Google’s launch of Scholar Library, their foray into citation management.
Initially this seemed like a great development, in that it increases access to Web of Science citations. Also, researchers conducting literature reviews or research impact analysis could take advantage of Google’s speed and ease of use along with Web of Science’s advanced search, filtration, and analysis options to, for instance, more efficiently filter out non-peer reviewed items and obtain an in-depth knowledge of an article’s citation counts.
However, after a few days, many libraries (including UT Arlington) received an email from SerialsSolutions (the company that runs Summon), stating:
Last week we were notified by Thomson that they will no longer be participating partners with any discovery service. This means that after December 31st, Web of Science content will no longer be discoverable in the Summon service – nor will we be displaying WoS citation counts.
Libraries with ExLibris’s discovery service, Primo Central, received the same message. This blog post directly reproduced the letter from Thomson Reuters that was sent to ExLibris:
“Thomson Reuters has decided to focus on enabling customers and end users to use the Web of Science research discovery environment as the primary interface for authoritative search and evaluation of citation connected research. For this reason Thomson Reuters will no longer make Web of Science content available for indexing within EBSCO, Summon, or Primo Central. Thomson Reuters will, however, continue to support Web of Science accessibility via integrated federated search tools that are available in Primo or other systems.
“The impact of this decision on your end users will be limited because the vast majority of the Web of Science records are available in Primo Central via Elsevier Scopus and other resources of similar quality. The Scopus collection is now fully indexed in the Primo Central Index and is searchable by mutual customers of Scopus and Primo Central.
Another letter from Thomson Reuters is reproduced here. Of interest:
Increasingly we have seen that while Discovery Services are a valuable tool for accessing full text content, they are not a replacement for the unique research discovery experience that the Web of Science offers.
As enabling research discovery based on citation connections is the primary objective of the Web of Science, and the foundation upon which our evaluation, selection and indexing process is based, we believe it is in the best interest of our customers and end users to use the Web of Science research discovery environment as the primary interface for authoritative search and evaluation of citation connected research.
For this reason we will no longer make Web of Science content available for indexing within EBSCO, Summon or Primo Central. We will, however, continue to support Web of Science Accessibility via linking and federated search capabilities in each of these discovery layers.
Notwithstanding their remarks on “enabling” users to rely on the Web of Science interface and that using their interface is “in the best interests of [their] customers,” they went through with the deal with Google to facilitate the use of their citations on Google Scholar. This implied to us that Thomson Reuters deal with Google Scholar was exclusive, and that Thomson Reuters was decisively turning its back on library discovery services, banking on Google Scholar as the discovery service of the future.
After what we assume to be quite a bit of push back from various libraries, Thomson made another announcement, that they “…have been in the process of evaluating our relationships with the three major discovery service providers [EBSCO/Proquest/Ex-Libris] and have decided to continue the indexing of Web of Science in them.”
In our opinion, this is not an issue of open access, since these links will only appear to Google Scholar users who are logged in as affiliates of universities that already subscribe to Web of Science. What this seems to be is a push by Google Scholar to enter the Discovery Service market as a closed service, in competition with services like Summon, Primo Central, etc. In short, “Google does not make it easy for others to build applications on top of Google Scholar, i.e. by providing APIs to allow third party access — the way they do for Google search or Google Maps.” It is possible that Google Scholar’s agreements with some journal publishers forbid them from opening their citation data by releasing their API. This explains the deal with Thomson Reuters. Thomson Reuters is positioned to increase the traffic to Web of Science. Crucially, given universities’ increasing reliance on metrics and impact factors, they will maintain their primacy in determining impact factor. In the end, librarians may get a discovery platform that is user friendly, but is still a closed system.
Creating closed systems such as this is damaging to libraries and to end users because libraries
want to try and take our content and services — purchased from a variety of different vendors — and present them to our users as if it were one single ‘app’, one single environment, as if the library were one single business. This makes matters more complicated, but it also makes this ‘de-coupling’ of UX [user experience] layer from underlying content and services — even more important. Because if the content and services we purchase from various vendors are tied only to those vendors own custom interfaces and platforms, there’s no way to present it to users as a unified integrated whole. (How would you feel about Amazon or Netflix if they made you use one website for Science Fiction, and a completely different website that looked and behaved completely differently for History?).
Whether the citation indexing system is Google Scholar or Web of Science, it is closed either way, in that it does not allow libraries to use the data to create customized user interface platforms, as the above post argues. Furthermore, there is always the danger of a for-profit service ending (like Google Reader) or going out of business. Also, such systems may be priced out of the range of many educational institutions, even though the data they gather is ultimately produced by those educational institutions. While Thomson Reuters reversed their decision in this case, it is important going forward that there are conversations about the need for an open system, where scholars can quantitatively analyze and reuse citation data for research, as well as evaluate the algorithm the citation system uses to rank results.