On chat, you can only type things and you just know someone is on the other side, also sitting there and staring and waiting for an answer, which not only should be completely correct, it should be there right now! You can only type so fast, and in many instances, you also actually have to find an answer. Unlike the telephone, you can't say, "I'm doing this and looking there" while you are actually doing it. Also, unlike when the person is standing there in front of you, they can't observe you diligently trying various keywords in the catalog. Maybe more importantly, you can't see them. You can't tell if they're angry and impatient. You can't tell if they're desperate. Like D. Scott Brandt said in his column [November/December 2000 CIL, p. 66], these types of communications "are more demanding than e-mail. They tend to put pressure on you to respond right now...." That is the whole point of integrating real-time communication into a Web site: Users can conveniently communicate with you at their time of need.

That is precisely why some form of an easy-to-use, online, real-time communication method will inevitably be integrated fully into reference services. It is why we decided to give it a try at the main reference desk here at the Jerome Library at Bowling Green State University (BGSU) in Ohio. BGSU is a residential university offering bachelor's, master's, and doctoral degrees, with about 16,000 full-time students. Last spring, we began experimenting with using online chat for remote reference. We certainly have found it exciting, and we think it offers huge potential for serving our users. It also brought up some surprising and interesting issues for us to struggle with as librarians and managers.
 

Save the Remote Users!
It only makes sense that since we spend hundreds of thousands of dollars making resources accessible remotely, we now need to serve the people who use them. Providing users with access to remotely available electronic resources without providing assistance and instruction on how to use them is like telling them which airport they are scheduled to depart from, but not giving them a flight number, airline, or gate number. They might be able to figure it out eventually, if they know their destination and approximate departure time, but only after considerable effort and frustration. Our patrons are increasingly using our remote resources. Statistics for searches and downloads continue to grow as our in-person reference contacts decline.

As reference librarians, one of our struggles has been trying to make the difficult and complex process of conducting research as easy as possible. Sure, interfaces have improved over the past few years. At BGSU, we even have the advantage of being in OhioLINK, a consortium that provides a single interface for nearly 40 of our most popular databases, all with easy one-click access from the bibliographic citations to local paper holdings information and electronic full text! But the truth is that research is complex. There's no getting around it. It's hard. It can be frustrating, even for us. Imagine what it must be like for people who don't do it for a living. On top of all that, not only are we offering users greater access, but we're also offering them more variety, too. Here at BGSU, we now have over 120 different databases from which to choose. Throw in the information available on the Internet and the subtleties of using Web search engines proficiently, and there is no doubt about the need to assist users remotely.

Finally, consider this: If we don't do it now, someone else will. Now is the time to take advantage of this opportunity to prove our worth. There are plenty of non-library-oriented "ask-a" services out there on the Internet, and not all of them do things nearly as well as we could. New products such as Questia and XanEdu, which market research content and assistance directly to students and faculty members, are also something to keep your eye on. They claim to be assisting classroom faculty and students by adding preselected, relevant, scholarly content to online courses or offering easy, one-stop access to research materials for a fee. They give the false impression that research can be made easy. Additionally, most, if not all, of the content has already been purchased by the libraries that serve the very same students and faculty! Certainly, for OhioLINK institutions, the content is not only already available; it is much better packaged and serviced, albeit in a more complex package. But it's a truthful package. So we need to be doing what we can to make that complex package work for users. Answering their questions online, in real time, only seems logical.
 

How Can We Help?
Here at the Jerome Library, we started out just wanting to see if our suspicions were right—that users, especially students, needed and would use a service that would enable them to interact with a librarian in real time over the Internet. For that reason, we needed to choose something with a minimum of expense, both in purchasing dollars and labor. One of our systems staff members volunteered to help us with software and technical issues, and all the librarians were willing to try, a few even enthusiastically. No one had much time to spend on it.

First, we considered instant messenger applications, but they presented one large obstacle. In our reference area, we have a bank of iMac computers that allow downloads onto the hard disk that remain there until the computer is re-booted. During high-traffic times, you can find at least three different instant messenger applications on these computers: AOL, Netscape, and MSN seem to be the most popular. If we chose just one of these applications and the user didn't use that one or any one at all, he or she would have to download it before asking a question. That barrier seemed to defeat the whole purpose.

We found free software on the Web from a company called HumanClick. It had lots of other advantages, in addition to its price. It required nothing but a Java-enabled browser for the user. It required very little on our end, too. We loaded a small piece of software on all of the workstations we wanted use to answer queries, and we put a short Java script into our Web page. It ran off the company's servers and only when we turned it on at our end. If we had it "off-line," it allowed the users to send us an e-mail instead.

In May 2000, we loaded the HumanClick application and tested it with a couple of librarians and staff. Then we had a very short training session on its use. It took less than 20 minutes. We then asked librarians and staff in other departments and branches to give it a try. About a week later, we added a link to the library's home page and a link from our e-mail reference page. We decided to try it while we were at the reference desk over the summer and see what happened. We kept statistics using hash marks on a clipboard, and eventually we started to cut and paste the entire transactions and print them out, noting a few items like the length of the interaction, and the time and day.

"Despite the strangeness of communicating via chat, it didn't take long to get accustomed to it."

A Few Points on 'Chat'
Observing my teenager chat online, I finally began to understand a few things about communicating this way. She was multitasking. She was listening to music, chatting in two or three different conversations, and surfing the Web. The time lag didn't bother her at all. She was doing other things in between thoughts in one particular conversation, and if she had another thought before she received her friend's reply, she just went ahead and sent it. Our colleague, Carol Singer, called this "curiously disembodied." There's no better way to explain the feeling. It certainly is not natural to us, but many of the students don't seem to even notice.

Despite the strangeness of communicating via chat, it didn't take long to get accustomed to it. We discovered that users tended to send many short messages rather than one long paragraph. The transcripts don't make complete linear sense, but while you're in the conversation, it's understandable. It certainly eases the anxiety of empty waiting time. The users didn't seemed to be as bothered by the length of time it took (to send, have the other person read, and then reply and send back) as we were. Eventually, we got used to doing something else too (like directing someone to the pencil sharpener, looking up items in the catalog, etc.).

However, we encountered a major technological problem in that we could not make the system reliably work for our users on Macintosh computers. One of the main disadvantages of using free software is that you have no technological support. We of course made the company aware of the problem and received assurances from them that they were working on a solution. About 50 percent of our campus desktops are Macs, so this became a rather large concern for us. Unfortunately, as we investigated other free products that provided chat, we couldn't reliably make them work for our Mac users either.

Late in the summer, HumanClick added two features that really changed how we communicated and how we felt about the potential of this communication method in reference services. First, we could now "can," or store, messages. After a short discussion, we added about a dozen preset messages that we hoped would help ease the time lag and save us from carpal tunnel syndrome. These were accessible to the librarian from a simple pull-down menu.

The second and more important feature that HumanClick added was the ability to "push" Web pages. By simply selecting a pull-down menu option and pasting in a URL, we could send any Web page we wanted. This is an enormous advantage. It is extremely difficult to teach someone by using only the written word. Nothing to point to. No facial expression to discern. No way to observe if they really could do it on their own. In academic libraries, we already struggle with this in using e-mail for reference. For example, someone e-mails us about how to find articles for a paper about antioxidants in wine and heart disease. It's really not all that great to send an e-mail back saying, "Try the keyword search '(antioxidants or wine) and (heart or cardiovascular) disease' in the research databases Biological Abstracts,CINAHL, or Periodical Abstracts," with a statement about how to get more assistance. You never really know if that's enough. But now at least we could send them Web pages.

Of course, this technological advance begged a new question: Which page do you send? With this cut-and-paste-the-URL method, it certainly isn't practical to send the users every page that they would see if we were helping them in person—the library home page, the research database's home page (and why), the subject or alphabetical list of research databases, the main menu of the selected database (and why we chose it) ... We haven't even gotten to the search yet, and that's already four Web pages and a lot of explanation. It doesn't take too long to say, hear, and see, but it sure takes a lot of time to send and read. The other option is to do the searches ourselves and send them the results list with a short explanation of where to check holdings and look for full text. In an academic library, our mission is to teach, to make it so that they can do at least some of it on their own next time. This sure doesn't help much at all with our goals in helping our users become more information literate and self-sufficient.
 

Time to Try a Better System
When HumanClick added the canned messages and pushing features, the company also announced that these features would not always be available for free. By this time we knew that we had a valuable service, so we began looking for funding opportunities. We also decided to stick with the best we could offer at the moment, hoping that we wouldn't have to go back to a service with fewer features if we failed. We also decided to go for the gold—to try to find the best possible product for our situation, knowing we may have to settle for less.

Our statistics showed that we were already receiving more reference questions on chat than we were on our e-mail reference account, and we didn't think it would be too difficult to argue that an improved product would greater enhance the service and improve our teaching capabilities. Additionally, our university was coming close to completing a multimillion-dollar project that included an entire new network. To that end, dollars, in the form of competitive grants, were being awarded for projects that had the potential to take advantage of the new network. Three of us co-authored a grant proposal asking for $10,000 to partially fund the purchase of a commercial product that would allow us to offer an improved online reference service. This improved service would not only include the features we had become accustomed to with HumanClick, but many more, as well as promise the potential for future enhancements.

We looked at a variety of fee-based customer service products. The product we chose is called Virtual Reference Desk (VRD), and it is available from Library Systems & Services, LLC. The VRD product stood out, despite its rather large price tag ($8,000 for setup and training, $500 per month for maintenance). First and foremost, it is a product aimed at libraries, not commercial enterprises. A few other libraries were in the process of making it available to their patrons and reported that the company worked diligently to make it work with their proprietary databases. The company's product developer is Steve Coffman [see his feature article on p. 20], who is a librarian. Finally, we would be able to make it available almost immediately, without major hardware upgrades and with very little work needed by our already overloaded systems staff.

The VRD product is powered by eGain, a Web-based customer call center application. It offers many features over and above simple chat. Some of these are transcripts that include URLs sent to both the user and the librarian via e-mail after each interaction, statistical reports for the library, queuing and customization features, and 2 full days of in-person training by company representatives. It also offers co-browsing and other collaboration features that will allow us to escort users around the Web and improve our ability to teach online. Additionally, as the university switches over to the new network and we and our users upgrade our desktops, VRD also offers voice and video features that we hope to integrate into our online reference services.

Our interim dean of the libraries agreed to make up the $4,000 cost difference if we were to receive the $10,000 grant from the university. We received notice in early December 2000 that we would be getting our request, and we began the process of purchasing the VRD product.

Since then, our training process has gone well, and our experience with HumanClick really paid off because as we're writing, it's February and we're ready to go live with our new product at the end of this month. We're busy preparing canned messages, slide shows, and completing our customization. Also, two of us are beginning to work with a committee in our consortium, OhioLINK, to determine how OhioLINK can help its member libraries begin delivering remote reference services. It promises to be an exciting and challenging end to our academic term.
 

We Can Make It Work
One of our major concerns is the impact this is going to have on our workload. Luckily, we are learning this new product during a semester where we are not only fully staffed, but we also have a returning faculty member on supplemental retirement and 16 hours a week from a graduate assistant. We'll be more concerned with staffing the service as it grows in popularity. Funding for continuing and expanding the service will also continue to take effort and concentration.

Many libraries have been trying to effectively accomplish real-time online communication for quite a few years in a variety of ways using multi-user domains, Net-conferencing, instant messengers, and other applications. For a variety of reasons,I don't think users or librarians really found any of these communications methods to be ideal. With the dawn of Web-based customer call center software, there may be a solution at hand.