A few days before the ALA (American Libraries Association) conference was scheduled to start, ProQuest (www.proquest.com) announced an "all-new platform" that would "redefine the search experience for library users around the world." The platform, which is as yet unnamed, will be introduced sometime in the first half of 2010; no prototype is currently available to demonstrate its capabilities. The platform will provide single-point access to ProQuest content. At launch, that content will include information now residing on the ProQuest platform and the CSA Illumina platform, plus selected Chadwyck-Healey products. All ProQuest products, including Dialog, will migrate to the new platform, probably later in 2010.
According to ProQuest's CIO, Bipin Patel, ProQuest's realization that it needed a new platform stemmed from its research into user behavior. Users had access to a tremendous amount of information, but learning the different searching systems was a barrier. How to effectively gain access to all this information was a challenge that ProQuest was not effectively meeting. End users were confused about how to search; librarians were concerned about gaining access to the databases for which they had paid. The legacy search systems-and there were several due to ProQuest's growth by acquiring other companies-needed to be revamped.
Total Platform Redo
Patel explained that, in thinking about the architecture, ProQuest decided to completely redo the platform, build it from the ground up rather than tweak existing platforms and bring them together. "We want common services," he said, "but at the same time, we need to offer unique access based on differing user types." There are two separate markets that must be satisfied: the customers who make the decision to subscribe (the library community) and the people who do the searching (end users).
ProQuest's research included extensive student observations, surveying more than 6,000 end users and conducting focus groups and individual interviews. Running alongside this was the constant interaction that ProQuest staff has with end users, librarians, and faculty. Using formal usability laboratories at Michigan State University and City University in London, researchers saw how differently individuals approach search. One thing the web has taught people of all ages, explained Patel, is that "It's OK to click. People are more inquisitive, more willing to learn." ProQuest research into user behavior has a 10-year history, so developers have the opportunity to quantify how people's interactions with search engines have changed over time.
The new platform will use Microsoft's FAST search engine to power what Patel refers to as a "common store." He believes FAST is sufficiently scalable to handle the vast quantity of data spread across the ProQuest stable. From a technology perspective, he wants to bring uniqueness to the end user, get the right content to the user, and "do more with the same information."
Agile Development Process
Librarians have been clear in telling ProQuest "don't screw up what works well." They like thesauri, indexing, SmartSearch, and many of the Illumina advanced search features such as hyperlinking. To create the new platform, Patel says, ProQuest has created storyboards, wireframes, and personas. He's also keen on the agile development that can be employed at an information company, where feedback is quick and changes can be made incrementally to the design, a significant change from his previous experience in the automotive industry, where designing a car takes years.
At the ALA conference, Mike Visser, director, platform manager, ProQuest, showed eight typical personas ProQuest developed. The personas represent an undergraduate, a reference librarian, a corporate/government librarian, a public librarian, a postgrad student, a faculty member, and a systems librarian. Visser said that as the ProQuest team worked on the platform, they continually asked, "Now, how would Gary (the undergrad) benefit from this? What about the librarians?" Personas make the development process more real.
Visser identified the primary drivers for the new platform as search, find, and use. ProQuest wants to put the user in control, provide context, and let existing users inform the product. Flexibility is a key element; the platform should adjust to the product line yet retain the single search elements that brand the platform as being from ProQuest. Users of ProQuest products range from elementary school to university students and faculty to corporate and government researchers. Children certainly need a different approach, such as SIRS' graphical interface, than does a serious adult researcher.
Singling Out Summon
ALA, ProQuest's Serials Solutions announced the first commercial adoption of its single interface library service, Summon, by Michigan's Grand Valley State University (GVSU) Library. Summon has been in beta at Dartmouth, Oklahoma State, the University of Sydney, Claremont Colleges, University of Calgary, Western Michigan, and the University of Liverpool. It is designed to provide a single search box that enables libraries to offer a single search box interface to its entire collection-books, documents, videos, e-resources at the article level, and anything else a library might include in its collection. ProQuest and Gale, a part of Cengage Learning, content were in the original beta version. They have now been joined by Springer, Taylor & Francis, SAGE, IEEE, Emerald, Scitation, the Royal Society, LexisNexis Ingenta Connect, and ISI Web of Science.
Much of the same user-behavior research that fueled the Summon development team is being used to develop the new platform. So what's the difference? Patel says that Summon is about discovery for the library. With Summon, library users can find all the resources, the physical assets of the library. The new platform brings discovery to library databases. It maximizes discovery of the library's subscription items. Visser added, "Summon is search and find. The ProQuest platform adds use."
Patel cited focusing on the end user; responding quickly to market needs; and any time, any place, any language, any device use as ProQuest's aspirational goals. He mentioned "purpose-driven" design, which means that ProQuest has seen other companies go through this process. The development team wants to pick and choose cool features from the open web, put them together, and use them if they fit the user's purpose in searching. It's the opposite of arbitrarily adding in features because they appeal to the "techies."
Somewhat less long term is the plan to bring the new ProQuest platform to fruition. According to Visser, ProQuest will bring its first test partners on board later this summer. It intends to bring others on board in stages. It will continue to present webinars to librarians as customer feedback affects product development. It should first roll out in the first half of 2010 with ProQuest, CSA Illumina, and selected Chadwick-Healey data. "We will continue to evolve the platform as we go," said Visser, "building in new features and product in subsequent stages."
ProQuest is not the only information company focused on single search, on building a new platform, or on going beyond federated search for library collections. Both ProQuest's new platform and its Summon service face competition from EBSCO's Discovery Service, which will allow for single search access to EBSCO databases, plus additional resources from OCLC, NewsBank, Readex, Alexander Street Press, and LexisNexis. EBSCO's just launched Integrated Search aims to go beyond federated search for library collections.
Web search engines, particularly Google, have raised searcher's expectations. As they gained ascendancy as the primary place people go when they want an answer to a question as well as when they're doing serious research, web search engines have brought Boolean logic from librarian obscurity into the light. They have trained people to expect all information to flow from a single search box.
In some ways, all the library-oriented search engines need an interface that both accommodates user expectations and effectively surfaces the enormous amounts of information resident within owned library collections and subscription-based research sources. It is not a trivial endeavor.