Working in Beta: Library Web Labs Let Users Shape Service

One of my favorite programs at the recent LITA National Forum was a presentation on library web labs given by Jason Battles of the University of Alabama and Jody Combs of Vanderbilt University. At its heart, the concept is an adaptation of Google Labs. Library web labs list new and experimental services offered by the library—usually though not necessarily web-based—and offer ways for patrons to try them and, more importantly, offer feedback. The library can then use that feedback to improve the service before its “official” release. The idea intrigued me both on its own merits and because it so nicely encapsulated one of the purposes of this blog, which after much struggle was finally getting started right then. Much of what we post here (including this post and Leonard Kniffel’s open letter to President-Elect Obama yesterday) will eventually be adapted to the print magazine, and some will be experimental to see if a topic has enough interest for broader consumption. So please, don’t be shy about sharing your comments. But to the labs themselves. Ken Varnum, the web systems manager at the University of Michigan, maintains that school’s MLibrary Labs site. “Something libraries have not been great at historically is experimenting in public,” he told me, noting the urge to make services “perfect” before release. The problem with that is that the endless tinkering in pursuit of perfection needlessly delays what is probably a functional and useful service. Or as he put it: “The great gets in the way of the good.” Combs noted that “In a test environment, if it doesn’t pan out, we haven’t invested too much,” and pointed out that the lab reduces pressure on the library to force innovations to fit the academic calendar, which would normally discourage new releases other than at the beginning of a semester. Battles observed that the lab helps the library to serve distance learners who wouldn’t be in the physical library. “This kind of platform gives them a place to see what we’re doing and give feedback,” he said. Building the Lab MLibrary Labs is a static index page that lists new services, so creating it took little effort. There’s also a blog associated with the site, which generates an RSS feed that can notify interested parties when a new service is available. New York Public Library’s NYPL Labs is similarly simple, built as a WordPress blog with plug-ins. Battles and Combs took a more involved approach, creating with a team of five the Test Pilot site when both were working at Vanderbilt in June 2006. That site includes a mySQL database to keep and organize comments, PHP code to interact with the database, and Javascript snippets from script.aculo.us, for the user interface to provide feedback. That code is portable, however; when Battles moved to the University of Alabama in July 2007, he used it and had a functioning site, Web Laboratory, installed in a single day, although he has since adapted it using Ajax to further improve the comment interface. They are willing to share the code with other librarians, and since presenting at LITA, Battles said he has had three requests for it. Test Pilot’s greatest success thus far has been when Vanderbilt was a development partner for Ex Libris’s Primo discovery tool. Branded as DiscoverLibrary, it was added to Test Pilot early in its beta release and received more than 150 user comments. As a result of user feedback, the library moved faceted browsing links that were originally on the right side of the screen to a more prominent location on the left. “That made all the difference in the world,” Combs said. “It really is a great way to get that kind of early feedback.” Alabama, meanwhile, got more than 100 comments during a website redesign, which pointed out some confusing search inconsistencies that developers had overlooked: Most searches gave item-level results, but database searches only gave title-level results. Once brought to their attention, changes could be implemented. NYPL Labs also received significant feedback on a year-long redesign of its digital gallery of more than 650,000 images. “It led to ultimately a much more finely honed site,” said Josh Greenberg, NYPL’s director of digital strategy and scholarship. Not all projects get this much response. Combs said that Test Pilot projects have received a total of about 280 comments, while Varnum said that MLibrary Labs gets about 750 visits per month. “It’s not the highest traffic thing on our side, but we wouldn’t expect it to be,” Varnum added. Lab promotion at all four sites generally takes place through links from the library websites, although Varnum noted that Michigan librarians also direct users to the lab when a tool there is appropriate to their needs.  Of course, library web labs are still a relatively new concept. “The lab itself is kind of an experiment as well as a space for other experiments,” Greenberg said, and the transparency of the lab requires a new relationship between the library and the public. But he added: “Because of how we’re funded, we can do that even better than companies like Google that have competing pressures.” See also: Ken Varnum’s Directory of Experimental Library Tools at his RSS4Lib blog.