Google Scholar @ GSK: from discussion to implementation

Jennifer Whitaker, a member of the Published Information Group within the Information Management team at GlaxoSmithKline, gave a succinct presentation at her briefing session, which nevertheless provoked much discussion amongst the attendees.

Jennifer described the thinking behind the decision to promote Google Scholar to researchers as a means of providing a quick search across scientific information on the web. She went on to explain how Scholar was positioned within the company; the information given to researchers; how Scholar has been subsequently used at GSK; and the effects on usage the standard bibliographic databases subscribed to at the organisation.

The reasons behind the decision to promote Scholar were pragmatic:

• There was a wish to maximise the use of expensive subscriptions to full-text e-journals.
• There was a need from researchers to be able make "quick and dirty" searches for background information on topics which would still yield information from quality-controlled sources.
• The standard Google search was already a highly used tool, and as such there was high awareness of the Google brand.
• Google Scholar was easy to use, yet offered some of the standard search features of a bibliographic database (such as journal title search field).
• Google Scholar offered broad coverage of scientific information across disciplines.

There were several strands to the implementation:

• An evaluation of scientific search engines was carried out, with users being informed of progress and results via the Library webpages.
• Once it was decided that Google Scholar would be the preferred choice, the Google Scholar toolbar was added to the primary Library Resources webpage in a prominent position.
• Much thought was given to communication with users, both in person through roadshows to departments, and through FAQs mounted on Library website.

Information professionals at GSK were at pains to inform users that Google Scholar would not be the answer to all their information needs, and should not replace the use of bibliographic databases for in-depth searching. It does not offer comprehensive coverage of scientific literature, and does not necessarily pick up the most recent publications. The search engine is also still in beta, which means that functions could change, and that there is a possibility that it could be withdrawn or become a chargeable service at any time.

An additional factor at GSK is the commercial sensitivity of the searches that researchers carry out. All employees are trained to be aware that their searches of the open web are insecure and can be tracked, and that standard web search engines should not be used when there is a need for confidentiality. Researchers were reminded of the other online resources on offer, and that the Library staff were also on hand to offer advice on searching and locating full text.

Trends in usage statistics for Scholar at GSK showed a steady increase between the implementation in June 2005 and November 2006, with marked increases in usage at the times when the search toolbar was added to the Library webpages, and also when full text linking to GSK's own full text subscriptions was added. Interestingly, usage statistics for the top 6 bibliographic databases at the company showed static usage over the same period, demonstrating that Scholar was a complementary facility rather than a competing one. Unfortunately, no statistics on the usage of full text journals were offered, so we could not see if one of the objectives of the exercise, to make these resources more visible, was achieved.

Jennifer concluded by stating that the decision to promote the use of Google Scholar at GSK was a successful but pragmatic one, and that it will be kept under review for the forseeable future, particularly as new scientific search products become available.

There were many questions from the audience - I have paraphrased some of these below, with Jennifer's answers:

Q: Is there source/coverage list for Google Scholar that GSK users can consult?

A: No there is not.

Q: Does the fact that Google Scholar is still in beta concern you?

A: It is a concern, but users are informed of the beta status and what it implies through face-to-face discussion and online FAQs.

Q: Has there been any research on how researchers at GSK use Google Scholar?

A: Nothing formal, however anecdotal evidence suggests that many are using it as a quick way of locating a paper for which they already have the bibliographic details, and for basic background information on a topic, for example a particular disease.

Q: Have you considered using server log metrics to find out who is using Scholar and how much?

A: Not yet, although we will think about this.

There was some general discussion amongst the audience about the desirability of Google having data on Library holdings in order to provide the OpenURL linking service, although no definite conclusions were drawn.

Q: Would you consider the implementation of a cross-search or federated search engine which could search across your subscribed databases?

A: We are keeping all types of search options under review as they are developed.

Q: If your statistics had shown a drop in use of your bibliographic databases, would you have considered cancelling subscriptions?

A: We would have felt that this was a failure of communication on our part, as Google Scholar is not intended to replace these resources. We would be inclined to re-double our efforts to market the databases rather than consider cancelling them.

It was noted by some members of the audience that users trust (or at least are highly aware of) Google, whilst some information professionals show a marked distrust of Google Scholar. It was also noted that users conflate brand awareness with trust - for example unpublished research has shown that if a set of search results is branded with the Google logo, users will trust the results, even if they have actually been drawn from other search engines.

Altogether this was an interesting session which once again highlighted that speed and ease of use is incredibly important to searchers, even those in the pharmaceutical industry.