TLA 2014 Conference Program Abstract
You implemented a discovery tool, so now what? We discovered there was much more to be done at Tennessee Tech after we implemented Summon in Fall 2012. Since then we have made a number of improvements related to the discovery tool both on the public side and behind the scenes. These changes have greatly affected the website and related services including interlibrary loan, Get It Now, RefWorks, and Google Scholar. We will also share how improving the relationships between the discovery tool and other library services and products create a better search experience for users.
When a library gets a shiny new toy that will benefit the users, like a discovery tool, the desire is to release it as soon as possible. That’s what Tennessee Technological University’s Volpe Library did when they purchased their discovery tool in 2012. However, after two academic years of experience, the library’s hindsight can offer insight to other libraries on the more subtle nuances of discovery tool implementation and how the to-do list continues well after implementation.
For the non-techies in libraries, a little background explanation of discovery tools in laypeople's terms might be helpful. A discovery tool searches multiple library resources at once, typically multiple databases and the library catalog, within a single search box. Users often equate it to Google for the library. When a user performs a search, the discovery tool searches the KnowledgeBase for matching results. The KnowledgeBase is a database the vendor maintains with millions of information sources. It is also the place where the library has marked what they have access to like library catalog records and subscribed databases and journals. Users can tailor their search to include all the millions of sources in the KnowledgeBase or just those to which the library subscribes. Once the search goes through the KnowledgeBase, it then goes through the link resolver. The link resolver matches up the results of the KnowledgeBase to the access links so users can click and find the information. Once the results are matched to the links, the user sees the results page with links. Since this process minimizes the need for users to perform multiple searches in individual databases, it makes library searching easier to understand and faster. However, there are multiple facets to consider and in-depth planning that should happen before actual implementation. Addressing these specific matters before “going live” can save library staff and users many headaches.
There are four main populations that need training: the library geek, technical services, public services, and users.
The Library Geek: This is the person who handles the technical side of the discovery tool and logs into the administration portion of the tool to accomplish various tasks. After purchasing the discovery tool, this person receives training first. The amount of training received will depend on the level of autonomy the library geek has. Vendors may offer different levels of control, one of which could be where the library submits troubleshooting requests, and then all technical issues are handled by the vendor. This option typically allows the library less control over the discovery tool, but it can save the library time and effort. There may also be an option that allows the library more direct control and access to the administration, or back end, of the product. Having more control will work only if a library has someone with the time and ability to manage the necessary technical tasks to manage the troubleshooting directly. However, this option will require more training for implementation and maintenance. In general the training is often online and includes communicating which subscriptions the library has for various databases, eJournals, and eBooks.This is also where the library learns how to get catalog records into the discovery tool, and that can be coordinated with Technical Services as needed. Library geek training should include how to pull usage statistics from the discovery tool. There should also be training available for how to link the discovery tool to external services such as RefWorks, Syndetics, and ILLiad. There are often updates to the discovery tool, which means there is ongoing training in the form of reading documentation on the updates or watching short, online training videos.
Managing the KnowledgeBase is probably the most time-consuming task for the library geek. The KnowledgeBase provides information on what full-text content users can access. Unfortunately, the Volpe Library had to purchase an additional link resolver during the first year of implementation, and the new link resolver was from a different vendor than the discovery tool. This was done in order to offer users the ability to access Get It Now, which allows them to request articles on demand while the library pays a per-article fee. This means the library geek now has to manage two KnowledgeBases because there is one KnowledgeBase attached to the discovery tool and a different KnowledgeBase attached to the new link resolver since they are from different vendors. Every time a subscription changes, both KnowledgeBases must be updated. The two vendors also use different terminology and procedures for accomplishing the same task, which makes this more difficult for the library geek. It is highly recommended to use the same vendor for the discovery tool and link resolver if possible.
A valuable part of training is learning how to customize the public interface and users’ search experiences. Many things may be customizable including the default search parameters, appearance of results, site branding with logos and colors, and default text and images. Consider all the possible customizations to determine what is best for users. At the Volpe Library, the default search was expanded to search everything in the discovery tool KnowledgeBase and not just material to which the library subscribes. Although searching everything yields results that are not available full text, this was the only way for users to access materials through Get It Now. Research guides were also excluded from the initial search results because the results were returning LibGuides from other universities, which included links only their users can access. Users can easily modify their results based on these defaults like if they want to include only full-text items and not everything in the KnowledgeBase. If the training does not mention customization for a specific area of the discovery tool, be sure to ask if it is possible. It took months to realize the phrases used for filtering could be replaced with more user-friendly language. For example, “Items in the library catalog” was changed to “Only items at TTU,” and “Items with full text online” was shortened to “Full-text items.”
One of the most important tasks for the library geek is to test the link resolver, which means doing sample searches and testing the accuracy of the links in the results. The Volpe Library was so excited to offer a single search to users, they made EagleSearch live on the website immediately after implementation and before testing any links. This was a mistake because it was later discovered that many links were not functioning properly, and that meant users could experience these broken links, which translates to poor service (see Table 1). Ideally, libraries will set up systematic link testing prior to implementation, most likely coordinated by the library geek. The Volpe Library tested 1000 links every other month for a while after implementation. Once broken links were discovered, the underlying problems were identified and corrected. Over time, the testing helped significantly increase the amount of successful links, so it is definitely a time-worthy project.
Table 1. Successful links in the discovery service before, during, and after testing.
MONTH OF TESTING
PERCENT OF SUCCESSFUL LINKS
A successful link means clicking on a link appropriately leads to the full-text or to the menu for no full-text request options like interlibrary loan. The broken links revealed more underlying problems than single broken links. For example, some databases do not handle openURL linking well. Therefore, the decision was made to “unsubscribe” to these databases in the KnowledgeBase, which means those databases are not considered part of the library’s online, full-access collection. Another problem was that some databases do not offer article-level linking. So clicking on the link to an article in the results does not take the user directly to the article but instead to the journal title as a whole, which then requires browsing by issue to find the article again. These databases were also set as “unsubscribed” in the KnowledgeBase. Databases marked as “unsubscribed” in the KnowledgeBase can still be searched in the discovery tool but do not lead to the full text. The databases can always be searched individually to access the full text. Lastly, the other major issue was that metadata for weekly newspapers was frequently wrong, which meant most newspaper results had faulty links. This metadata issue helped guide the decision to exclude newspapers results in the default search from the library website. Again, users can always opt to include newspaper results from the search results page if they so desire.
Technical Services: These are the people behind the scenes who manage uploading the library catalog records into the discovery tool. Since the library catalog is included in the discovery tool, technical services staff need to be trained in how to upload the catalog record into the KnowledgeBase. In addition to the logistical training of how to upload them, decisions should be made prior to implementation regarding how often to upload the records. Some libraries write programming code to send catalog records automatically to the discovery tool every night. However, this code is not typically available from the vendor, so the library must have someone who can do this in-house, which is not always an option. Another option is to upload records on two timelines. First, the total upload of all catalog records can be done periodically, which is every three months for Volpe Library. Then the catalog additions and deletions (i.e., purchases and withdrawals) can be uploaded on a more frequent basis, which the Volpe Library does twice a week. Changes to catalog records like relocations do not need to be uploaded because the discovery tool links to the catalog record. Since technical services personnel spend so much time and effort maintaining the catalog, it can be difficult to determine what the users’ preferences might be. Therefore, they should work closely with public services personnel to determine the upload frequencies that best serve the users.
Problems that arose with catalog record uploading were related to suppressed records, bib types, and eBooks. Originally the catalog records that were suppressed were uploaded to the discovery tool. However, clicking on a suppressed record in the search result yielded a cryptic error. After unsuccessfully trying to get the vendor to hide these records in the discovery tool, it was finally resolved by not including any suppressed records in the total uploads every three months. The bib types in the catalog are extremely important because they are used to map the content types in the discovery tool. This means that when a user narrows a search by content type and searches for only videos, that narrowing happens by the bib type within the catalog records. To make this mapping successful, technical services staff should first ensure that all bib types are correct and consistent in the catalog. Then they should ensure that each bib type is mapped to the correct content type in the discovery tool. An example of where this can be problematic is a catalog record for streaming video. Should that bib type specify it as a video or a streaming video? Is that bib type then mapped to the discovery tool as a video, internet resource, or some other content type? The last potential issue with catalog records involves eBooks. If a library subscribes to an eBook collection, and that collection is in the KnowledgeBase and the MARC records are in the catalog, the library has to decide whether to offer users access to the eBooks via the KnowledgeBase or the catalog. Access via the catalog records typically requires one more click than via the KnowledgeBase, which may influence the decision. If the library chooses to offer access through both, meaning they “subscribe” to it in the KnowledgeBase and leave the MARC records in the catalog, there will be duplicate records in the results and can cause confusion. The good news is that many issues, questions, and problems in the discovery tool related to catalog records can be solved in-house by technical services instead of relying on the vendor.
Public Services: These are the people working directly with the users through reference, instruction, information desks, interlibrary loan, and other related services. Often there is a resistance by public services personnel to learn a discovery tool because they are used to the individual databases and catalogs, and they are extremely adept as using them. They may dislike the discovery tool because it is less targeted and has fewer perks and functions than individual databases. Although that is true, they need to become adept at the discovery tool as well because they serve the users, most of whom will be using the discovery tool heavily. Public services personnel also need to use the discovery tool because if a library does not have eBook collections in their catalog, but do have have them in their discovery tool, they are not searching all of the library’s book resources by just searching the catalog and may be giving incorrect answers to patron queries.
In addition to learning the discovery tool as a user, public services personnel must also learn it well enough to teach it, whether it be in a group instruction session or individual reference assistance. Using a product does not require as much skill as teaching it, because teaching it requires deeper understanding of the advanced features, common pitfalls, and subtle nuances. Learning occurs on the fly as the public services personnel use it, plus the discovery tool most likely offers webinars and training, so they should take advantage of those to understand how to best use the tool.
Users: These are the end-users, most commonly the patrons. The great news is that users love a discovery tool because it speaks to them in non-library jargon and mimics their typical internet search box environment. They do not have to know what a database is or how Boolean operators function. Realistically, most users are able to use the discovery tool with no training. However, there is definitely value to training users on the features of the discovery tool to make their searching easier. This training comes in a multitude of venues, so the library must remember to update instructions and training in all aspects. There should be self-help instructions on the website so that users can best use the discovery tool even if they are not located in the library or prefer to learn on their own. The approach to training users' search skills in both reference interactions and instruction needs to be updated as well. A common approach is to start with the discovery tool for the initial search, and then branch out to other venues like individual databases as needed. If that is the library’s approach, that method should be consistent across all interactions with users so they are being trained in the same way. If the library’s instruction sessions also use assignments or activity sheets, those will need updated as well to reflect this training approach. Lastly, academic and school libraries need to ensure that the school’s instructors are aware of the discovery tool and how it works. Not only do they need this training as users of the tool, but they also need training so they can give research assignments in a logical manner. For example, prior to implementing a discovery tool, they may have been assigning their students to perform searches within a specific database. That type of assignment may no longer be relevant once they understand how to use the discovery tool.
How a Discovery Tool Affects Other Services
When a library purchases a discovery tool, changes to other library services and technologies are necessary for successfully incorporating the new tool. Additionally, although it has a lot to offer as a search product, a discovery tool may also benefit other library services.
Library Website: Altering the library website becomes mandatory as part of implementing a discovery tool so users can find and use the tool. The more subtle but equally important thing to consider when you are changing your website is how you direct users to your discovery tool. Most libraries place the search box on the home page, but related labels and terminology are very important to users. For example, when Volpe Library first implemented a discovery tool, the search box section of the home page looked like this (see Figure 1):
Figure 1. The Volpe Library search box with multiple tabs.
Over time, users revealed that the tabs above the search box were confusing, and they thought they were maybe supposed to search each tab separately. That is most likely due to the fact that users do not understand the differences in library jargon like eJournals, databases, and catalog. Therefore, the library changed this portion of the website to this (see Figure 2):
Figure 2. The Volpe Library search box with simplified tabs.
The tab for the discovery tool has more than just the word “Search” and the other tabs were condensed into one tab. Although this language and design option is subtle, these changes are meaningful to users in how they direct usage of the discovery tool. Since this change there have been fewer questions from users about searching duplicate entities.
ILL/Document Delivery: The ILL personnel should be trained on using the discovery tool as part of their work flow. However, the discovery tool can also change how users request items through ILL. Previously, the Volpe Library directed users to the ILLiad website to manually type in interlibrary loan requests. Now the discovery tool allows users to click on a "request" button that leads to the ILL service. Not only does it not require opening a separate window for ILLiad, but requesting something through the discovery tool automatically populates the ILLiad request fields like title, author, etc. It saves the user time and typing, which is very convenient for heavy ILL users. Volpe Library also offers the Get It Now service through the Copyright Clearance Center. This service allows faculty and graduate students to purchase articles on demand. This service is not possible without a discovery tool, and now users can click a "request" button, which leads directly to requesting the article, again not having to type in any additional citation information.
Citation Management Tools: Volpe Library subscribes to RefWorks, an online citation management tool. Users can always go to the RefWorks website directly and manually type in citations. However, the discovery tool allows users to export multiple sources at once to their RefWorks account with all fields auto-populated. This saves a lot of time and effort when putting citations into RefWorks. RefWorks stores only the citation, not a copy of full text. To help users find the article again, RefWorks connects back to the library collection through the discovery tool and link resolver. That means there is a button in RefWorks for each citation that links the user to the item within the library. This button appears for all citations including those originally typed into RefWorks and those exported from the discovery tool. This is much more convenient than forcing users to remember the original search and perform it again.
Other Search Entities: Some users may opt to use other search entities and not the discovery tool to perform their research. Those other entities can often be configured to connect to the discovery tool. The two most common examples are individual databases and Google Scholar. When a library subscribes to an individual database, there are often branding options within that database and settings related to user access the information. There may also be options within the database setup that are similar to the linking option from within RefWorks. In other words, there might be a way to include a button that uses the link resolver and discovery tool to find those resources that are not available in the database based on the subscription/embargo. The ability to set this up will vary by database, but is important to check because even if the discovery tool does not find the full text, it can lead users to interlibrary loan and article purchase options with the ability to auto-populate the fields. Google Scholar can also be connected to various libraries within its settings. As part of this connection, Google Scholar uses the link resolver and the discovery tool to find items that are not already identified as full-text. Again, if the library does not have access to the item, users are easily connected to interlibrary loan and article purchase options with auto-populated fields.
Future Technologies: Although it is impossible for libraries to predict their future technology upgrades and purchases, it is beneficial to consider potential purchases and how they might coordinate with the discovery tool. Items and services to consider for compatibility when changing and/or upgrading may include the library catalog, digital library/repository, citation management tools, library website, interlibrary loan, document delivery services, article purchase on demand, Google Scholar, and other databases. Even if the current technologies are compatible, it is important to consider how changes and updates will affect the discovery tool. It is also vital to plan ahead for how changes to the discovery tool will affect these other services and technologies since the connection often goes both ways.
The ideal situation is to consider all these activities and issues prior to releasing a discovery tool to the public. However, the reality is that many of these entities are not revealed until after implementation. Libraries should do their best to complete as much as possible before going live with the discovery tool. They should also realize, however, that even after implementation there is still much work to be done to continually improve users’ search experiences, especially considering the rate of change in libraries.
Sharon Holderman is Coordinator of Public Services at the Volpe Library at Tennessee Technological University. She can be reached at firstname.lastname@example.org.
April Crockett is Web Manager at Volpe Library at Tennessee Technological University. She can be reached at email@example.com.