Only six months ago, the need for next-generation discovery and delivery tools was still being discussed. Questions such as “So, is it an OPAC keyword search?” were still in the air. Now, with this era clearly behind us, libraries seem to be split into three groups—those that already have a modern discovery and delivery solution, those that are in the process of choosing one, and those that are trying to obtain the budget for one.

No matter where your institution stands in its decision life cycle, the bottom line remains that users expect a modern discovery and delivery mechanism to help them find library resources effectively. If a library does not implement a solution that meets users’ needs and expectations, those users will drift to other information portals, turning to the library only when there is no alternative—a rare occurrence in today’s easy-to-access world of information.

Discovery and delivery tools and what we expect from them

Next-generation discovery and delivery tools offer the following core functionalities:

• Fast response time for searches
• High-quality relevance ranking
• A single point of discovery for all resources
• Collaboration capabilities that encourage patrons to add content such as reviews, ratings, and tags, and to view content contributed by other users
• A simple, intuitive user interface

Whereas many of the solutions on the market today boast all of these features to some extent, the quality of support for these features differs vastly from product to product. Let’s look at a few examples.

Search-Response Time

Most people do not measure how long they wait for their search results—they don’t know if the average response time is 150 milliseconds (ms) or 3000 ms. However, as is evident from usage statistics, slow response times do affect users. The response time of most sites running Primo®—the Ex Libris discovery and delivery solution—is below 500 ms for an average search. In some cases, the time is less than 150 ms.

Relevance Ranking

Studies have shown that internet users, including academic researchers, generally examine the items that are presented on the first page of search results—but rarely go beyond that page.[1] Thus, the order, or ranking, in which search results are presented on the screen is of major importance, leading users to choose a search engine according to the relevance of its result sets. When users search for a specific item, they can easily gauge the quality or relevance of the result set by checking where in the result list the item that they are seeking appears—is it close to the top or not? However, users who are looking for material on a particular subject—as opposed to looking for a particular item—have a harder time identifying the quality of the result set because they are not always aware of material that is not presented to them.

Today’s users’ are very sensitive to the appearance of false positive—irrelevant-- results high up in the result list, To combat the dissatisfaction that users may feel with false positive results, some search engines go overboard in limiting the result sets so that users do not feel they are being inundated with irrelevant information. Unfortunately, the flip side of this measure is that users may be missing valuable academic data.

We understand the importance of not missing relevant information. Thus, in Primo, we have invested a great deal of effort in striking the appropriate balance between presenting all results and only relevant results, with the goal of ensuring that users will not miss out on the real information that they require. To optimize the way in which Primo ranks results, librarians and Ex Libris software developers teamed up with a company specializing in relevance ranking algorithms and tailored the Primo algorithm to searches in the library environment.

Single Point of Discovery for All Resources

Many next-generation discovery and delivery solutions claim that “anything can be searched.” In reality, the only information that can be searched is the integrated library system and that residing in other systems that the discovery and delivery solution is tailored to integrate with. Ex Libris believes that searching all of a library’s resources touches on the most complex part of relevance ranking—blending multiple item types.

The question of how multiple items are blended and displayed logically in a single results set is a challenging one. It will require the library to answer questions such as: Which is more relevant to a user’s search query—a musical score or a book? This is the type of question that librarians can address by leveraging their deep knowledge of metadata. Keeping this fact in mind, Ex Libris has designed the extensive Primo publishing platform so that librarians—not programmers—can define the manner in which various metadata formats are converged into a single, all-inclusive format that compares different material types—apples to apples—for ranking. Entrusting this task to librarians—the metadata experts—ensures that the search results that end users receive are not only accurate but also present an appropriate ranking of items from a variety of resource types.

User-Generated Content

Today’s users, accustomed to Web 2.0 platforms, expect to contribute content about the materials they are accessing. Primo thus features three user-generated content options: tags, reviews, and ranking.

Ratings and reviews are important to users wishing to comment on a specific item or to receive additional information on it. For this reason, ratings and reviews appear in Primo’s full-item view. Tags are the most common form of user-generated content, enabling people to classify items. In Primo, tags are presented in the full-item view and on the tags page for all items.

Simple, Intuitive User Interface

According to user interface experts, the interface of a product should resemble something that the user is already familiar with. Users that have to learn a completely new interface will typically opt for a more intuitive option elsewhere. Whereas many search tools have beautiful graphics, those search tools that people actually use contain little visual distraction and focus on presenting results, as evidenced by the success of Google.

The Primo interface is constantly being tested by customers that carry out independent usability tests in their own usability labs. The results that they share with us serve as the basis for our regular interface updates and enhancements. The now mature Primo user interface serves as an excellent entry point to library collections. We also offer customers the option of customizing the interface should they wish to do so.

Primo–all this and more

There are two fundamental architectural elements that differentiate Primo from other discovery tools: Primo’s open-platform architecture and deep-search architecture.

In adherence to the Ex Libris open-platform strategy, Primo now includes an API for 50 services. This API facilitates the customization of Primo and enables customers to develop code extensions that can also be shared with other members of the Primo community.

The strategic commitment to enable Primo to provide quality access to all library research resources is backed by an architecture that facilitates deep searching; that is, libraries can run Primo on top of external search engines, thus allowing users to search other resources, through Primo, with minimal overhead. The results obtained through the deep-search architecture are ranked by relevance and displayed using faceted categorization. 

For additional information on Primo and its unique functionalities, see the Ex Libris Web site.

Contributed by Gilad Gal, Primo product manager

[1] Kules, B and Shneiderman, B. 2005. Using meaningful and stable categories to support exploratory web search: Two formative studies. Read more