The Summon discovery service is used by more than 500 libraries in more than 40 countries. It is built on a unified index of content, leveraging “match and merge” technology to combine rich metadata and full text from multiple sources. In the past 3 years, the Summon unified index has grown from 200 million items to more than 1 billion—with the vast majority of article and book content full-text searchable. Serials Solutions markets Summon to academic and large public libraries, where it is tailored to individual library electronic resources.

The new Summon Suggestions suite of features builds on librarian expertise, community collaboration, and global usage data to provide and improve automated and contextual research assistance to lead users to better research outcomes.

Summon librarians can use Summon Suggestions to integrate localized recommendations, automated search guidance, and live reference help. Automated topical refinement suggestions embedded in the search results provide users with options to intelligently expand queries to improve results.

“Our true Software as a Service model enables us to offer libraries unprecedented levels of local customization and librarian engagement via easy-to-use administrative tools that don’t require complex local programming skills,” said John Law, vice president for discovery services at Serials Solutions. “By leveraging Summon usage data from our global user community to improve discovery we are just now beginning to scratch the surface of possibilities for globally-informed data-driven features—previously the sole domain of open Web search engines.”

The new features include the following:

• A new version of the Database Recommender that provides libraries with local control over database recommendations to ensure users are directed to resources pertinent to their research and in the context of their queries. Librarians can locally curate recommendations as well as take advantage of global “community-sourced” recommendations to guide users to valuable, targeted research resources. According to Law, the Recommender will only name databases in the user’s library’s e-collection.

• “Best Bets” are locally controlled, library-generated recommendations that expose users to resources such as research guides, specialized collections, library web pages, course reserves, search tips, help tools, and more. Again, the Best Bets highlights will tie to the context of a user’s query. According to Law, if the guides, web pages, etc., that go into Summon are available to the open web, the resources will become available to all Summon users, not just the individual library serving the individual user. Law stated that he was “kind of frustrated with the way other discovery services work, the walled garden approach, where libraries were not able to stay in the research process.”

• Related Search Suggestions provide users related concepts and expanded queries that can lead to better search results. Leveraging real-time, global Summon usage data, related searches offer users scholarly and multilingual suggestions for query refinement. Topical suggestions for new queries, similar to what users experience in open web search engines, are available at the time of query entry (Autocomplete) as well as embedded within search results. For example, “Global Warming” will conjure up “Climate Change.” The feature seems designed to expand the size of search results, which some searchers may not find the main problem. However, Law remarked that the faceted and advance search features can, for example, restrict a search to certain types of documents, and work to narrow search results.

• Embedded chat and reference widgets allow users to receive real-time help from reference staff without leaving the Summon user interface. Libraries can integrate and customize chat and reference widgets in the Summon environment with easy to use administration tools. However, this service will probably require libraries to deal with third parties. Law specifically mentioned Springshare. Clearly, the advice services would work best if they worked on the same time schedule as Summon, but that schedule is 24/7 and most librarians are not quite that available. Law said that they have no plans to offer live service support from Summon on-site.

• A custom text editor allows libraries to locally modify text in the Summon user interface to ensure their users see textual information preferred by the library. For example, libraries can locally control and edit labels and text—including error and informational messages, facets labels, and limiters—in any of the 34 interface languages and dialects available in the Summon service.

To see details of how features in Summon Suggestions look, go to www.serialssolutions.com/en/services/summon/features-summon-suggestions/.

In creating Summon Suggestions, Serials Solutions relied on the process of “agile development.” [Those interested in the process and its practitioners might try www.activedevelopment.org.] According to Law, the agile software development process helps them make quick changes, even for major features. He remarked, “We usually make iterations of Summon every three weeks. Sometimes the changes involve internal, back-end tools, but nine out of ten times they are new enhancements.”

But that’s not the only procedure driving changes in Summon. The system relies heavily on user feedback. In fact, every search result carries a request for feedback and Law says he gets piles of them on his desk each day. Librarians probably give the most extensive input to the development processes. Summon has a real-time listserve and a major wiki.

“Summon uses the stretch expertise of multiple libraries, e.g., in the Database Recommender," said Law. "We’re harnessing librarian expertise into our product and allowing librarians to customize it, for example by choosing a specialty like history or the humanities, to serve their own constituents. But, it also looks across libraries and matchpoints quantitative analysis with recommended scenarios from multiple librarians That enhances it for all libraries. Now we are engaging librarians in the process the platform enables. Librarians will learn from their engagement with the platform how to optimize it. The metrics on different uses of the Database Recommender are part of the feedback. All of our excellent feedback services provide ways to hear from users and librarians and that is powering lots of innovations.”

So what does the future hold for Summon? Lots, according to Law, but he could not share many specifics. When I asked him whether Summon could build reading lists for users, I was told to talk to him next year. When I asked about current awareness services, Law indicated that they currently use RSS, but that they might offer such a service to future user accounts. When I asked whether the chat and live reference support experts got copies of searches to use in generating their advice, he said no, but it was a great idea. When I asked whether they had any plans to integrate Summon with the end-user Udini service from ProQuest, he hedged his answer. It sounded like he might be preparing an alternative for graduates of academic settings from Summon.

Clearly, Law has a keen awareness of the expansion in alternatives for scholarly communication. He commented, “Right now there are over 300 institutional repositories offering openly available content with a, growing wealth of scholarly open access as well as a wealth of information in library catalogs.” For right now, Law wants to wait until librarians get used to the Summon Suggestions changes. But, he advised people to “look at what we have accomplished over the last few years and the level of investment will only be increasing over time.”