We were digitizing a historical collection, so we were more concerned about technology than relationships. But then the barrage of e-mails proved that we'd become a personal part of users' lives, and that we'd always be married to our users--for better or for worse.
The Historic Pittsburgh Web site is the maiden project of the Digital Research Library (DRL) of the University of Pittsburgh's University Library System. This multi-component site contains rich material that documents the history of Pittsburgh and Western Pennsylvania from the times of exploration to industrialization and modernization. This noble idea from 1998 resulted in the formation of the DRL, a library department devoted to creating digital resources. It also involved a multi-institutional partnership with the university's Archives Service Center, Special Collections department, and the Historical Society of Western Pennsylvania to identify and supply content for this project, which now contains a full-text collection, a map collection, finding aids, a chronology, and census data (http://digital.library.pitt.edu/pittsburgh).
We have completed the majority of work needed to meet our original concept of Historic Pittsburgh--automated scripts have been written; vendors selected; software tested and chosen; items procured, scanned, indexed, and made searchable; interface design completed. In this first project we achieved our initial goals of mounting digital library content and attracting a significant user base. So outside of ongoing systems maintenance, migration of files and software, and the occasional server upgrade, it would seem as if it's time to sit back and reap the rewards of our job well-done. However, while many aspects of digital collection development go away once the product is online, one very important element will remain as long as the collection is accessible: the users, who will inevitably have questions and comments. Who answers these questions? How are responses coordinated among the partner institutions? How much time should we spend on these questions as we juggle the production work for new digital projects ? The way we chose to manage user correspondence has helped the DRL evolve into a wiser matron of the digital library field.
Until Death Do Us Part?
"Are any of the church records for Trinity Episcopal or the first Presbyterian church in Pittsburgh online? Any suggestions on how to get access to baptismal records from around 1819-1820?"
"I am searching for more information on Patrick Best ... Patrick was supposed to have killed a bear with a knife and his bare hands. There was supposed to be a statue for this momentous occasion. What information I have found so far, that could be a possible lead, has been on your web-page. Would you have anything more on this man?"
"Do you have any information regarding Adam Jacobs packet boats? I am attempting to preserve a model built by the original packet boat builder and I need 'all the help I can get."'
These questions are examples of what we have received from users of the Historic Pittsburgh Web site. They illustrate that questions about digital resources can be straightforward, complex, fun, or just plain unpredictable. Above all else, these e-mails require time and effort to respond to. We are not the only digital library faced with such an ongoing task, yet the notion of user e-mail remains an unspoken fact of life. In a way, it is like a marriage between user and digital library, where reality sets in after all the energy has been spent on the glitz and glamour of planning the wedding. In other words, managing user correspondence is a persistent but often-unidentified responsibility associated with digital collection maintenance.
The Honeymoon's Over, or, What Have We Gotten Into?
In the fall of 1999, Historic Pittsburgh was up and running and bringing in a steady stream of user correspondence. We quickly realized that we had a loyal audience, but not an audience entirely made up of the academic patrons to whom we were officially betrothed as a library department serving the University of Pittsburgh. While the University Library System limits licensed e-resources and Web-based services to the Pitt academic community, the DRL is committed to a much broader audience that is able to freely access our collection from anywhere in the world. Surveys and usage statistics showed us that faculty and student populations did indeed use our collection, but through user correspondence we learned that the most active and vocal user group was a national community of genealogists and family historians. For our first digital initiative, this audience was nothing less than a real blessing. This demographic was enthusiastic about incorporating digital resources into their research, and it is thanks to th em that we have had such a successful maiden digital project.
Perhaps there is something inherent about genealogists that makes them fearless about e-mailing questions and comments to the nameless and faceless hp firstname.lastname@example.org. This group of genealogists and family historians regularly fills our in boxes. We have been able to ascertain a few general characteristics about them that Web statistics and surveys could not reveal. First of all, many of them have traversed a generational digital divide. We are not making unsubstantiated assumptions about age; in fact, many of our most vocal users find subtle ways to let us know how old they are and how delighted they are that they have mastered genealogical research online. Second of all, they are not trained librarians or academic researchers. This may seem obvious, but often we digital library professionals get caught up in providing highly efficient retrieval features without acknowledging that the benefits of full-text proximity and Boolean searches remain a mystery to most people. Last but not least, they are not satisfied being an anonymous and distant audience. They value personal interaction, even if it is mediated and delayed through e-mail. Users often respond with messages of appreciation containing anecdotes, family stories, or Pittsburgh memories. These e-mails are reminders that we are providing a digital product with very personal meaning for many of our users.
A Comfortable Compatibility
User questions fall into three general areas: administrative, technical, and contextual. For administrative and technical questions, we have compiled a cache of stock responses to such common inquires as these: "May I use an image from the digital collection for a presentation or published piece?" "How do I get a page from a digital text to print?" "Why does a page or image show up as a broken link?"
The majority of user interaction, though, involves questions related to the content of Historic Pittsburgh. Because they are unique by nature, content questions require more time to answer, may involve follow-up e-mails, and may necessitate coordinating responses from our partner institutions. These questions fall into a few major categories as well:
1. Most of our content questions come from users who need guidance in finding specific information within our site. Either they are not utilizing the search features at all, or they do not know how to use them to their fullest capacity. In these cases we provide instruction and search strategies that will ensure results.
2. Other popular inquires require resources we do not maintain, such as deeds, birth and death records, cemetery records, or material from other geographic regions. Enough of these questions have beenposted that we have been able to maintain bookmarked links to the appropriate repositories. Although we are passing the buck, so to speak, we do have the responsibility of gauging these available resources for legitimacy and relevancy. In many of these cases, our historical society and archive content partners can handle the questions directly.
3. The other common messages carry suggestions for additional content and requests to purchase material that they've viewed there.
Initially we had two goals for answering user questions: 1) Respond to every non-spain e-mail, even those asking us to appraise old red sofas purchased in Pittsburgh a century ago, and 2) deliver the response within in a day or two. Answering user e-mail was an activity that evolved and eventually demanded its own work flow. We built a team that includes the DRL staff, a representative from the university library preservation department, and representatives from each of our content partners to monitor user e-mail. Since there were nine possible people who could answer any given question, it was necessary for one person to coordinate their responses. This responsibility fell to the DRL's digital production librarian. She had overseen selection of Historic Pittsburgh content and had direct knowledge of the resources that were available in each component of the site as well as general knowledge of the physical resources maintained in the partner institutions. Also, she was in a good position to parcel out respon sibility, for she was either in proximity to or could act as liaison to the other user e-mail recipients.
The feedback coordinator receives, categorizes, and prioritizes user e-mails. If clarification is needed, she contacts the sender for more information. If the question can be answered using Historic Pittsburgh components and their search functions, the coordinator fields the question herself. If the question falls into technical or administrative territory or requires specialized knowledge from a content partner, she passes it on to the appropriate librarian or archivist. If the question requires external resources, the coordinator performs a quick review of available sites and collections from other academic, cultural, or government institutions. After she posts responses to the user and to the rest of the team, she awaits follow-up questions. The DRL staff archives the questions and responses for future reference and yearly statistics.
What had started as a routine byproduct of our daily production work flow ended up mimicking a sophisticated design for virtual reference management. In our own way, we were working through the steps of question negotiation and following a digital reference model, and for part of the day some of us were wearing the hat of a reference librarian. For a small department of three faculty librarians and two staffers who are juggling numerous projects at varying stages of development, this is a time-consuming use of human resources. In the last 3 years, we have recorded over 500 user inquiries and comments. Although this may be a drop in the bucket to some library departments, user correspondence is a significant and predictable presence in the DRL's weekly production activities. Based on our experience with Historic Pittsburgh, we realize that to maintain a balance between our public service and production roles for future projects, we will need to pass on more of this responsibility to our project partners, parti cipating faculty, and the library's public service department.
The Marriage Is Tested
The most valuable lessons about our public service role came about in the fall of 2000 when we unveiled the census component to our Web site. A University of Pittsburgh history professor and his student assistants had transcribed 19th-century census data from microfilm to electronic format more than a decade earlier. His data were passed along to us for use in Historic Pittsburgh. Very rarely does a project develop with such ease. With a tweak here and there we had a new and seemingly complete component to add to our site, which we thought would please everyone.
To our surprise, our enthusiastically vocal users were now nothing less than rabid, and for good reason. It quickly became clear to us how little we knew about our audience of genealogists and what they expected from an online resource. None of us were genealogists. We knew little about what makes them tick or how they conduct their research. Unknowingly, we had not supplied them with all of the pieces they needed for proper census research. Census volume page numbers, soundex codes, and microfilm reel numbers were important to them, but were very difficult for us to translate from the original electronic files into our census database. Furthermore, we had no way of knowing whether names were transcribed correctly or if personal information about such important things as marital status or insanity was correct. Genealogical research is passionate and inherently about identity. We received numerous user e-mails pointing out the deficiencies of our data, misspellings, and embarrassing blemishes on family names. We received an equal number of e-mails asking us to fix these problems.
What we learned from user response to our U.S. Census schedules was nothing less than enlightening. First of all, don't misspell great-great-grandmother's name, and above all else, do not tarnish the good name that it most surely was! If you do, make sure you have a clearly stated policy to back you up. Because we lacked the means to validate these requests, we made a decision not to fix reported misspellings and unfortunate transcription mistakes. We documented this decision and point to it often when responding to this type of e-mail. Second of all, we learned to be more accountable for our database projects. We feel a lot more secure when we have some control over data input. The mistakes in the census database stemmed from the poor legibility of the original census schedules and were not unreasonable, but because we had no responsibility for the data collection we were less prepared to establish policies for quality control and user dissatisfaction.
Most importantly, we learned how valuable it is to have intimate knowledge about your users, just as it is to intimately understand your marriage partner. Exploring user needs and research patterns can benefit anyone who builds digital collections. Attending a lecture in genealogical research or consulting with the local genealogical society or historical society might have helped us to make simple modifications to search and display problems before this component of the site went live.
Ready for Lasting Fidelity
Managing user correspondence for Historic Pittsburgh takes time, patience, and commitment. In fact, it takes more time, patience, and commitment than we had initially planned. Granted, we could have immediately passed the primary responsibility for managing user response to the library's public service department, or we could have placed it entirely in the hands of our content partners. But we're happy we didn't. In the end, maintaining control over user questions and feedback for our first project worked in our favor. We learned a lot from our users' correspondence. While we were still in the active stages of selecting content, users' suggestions helped us to gauge areas of weakness in our collection. Their suggestions were also the basis for the development of other digital projects. Consistent requests for visual images of Pittsburgh eventually led us to a 2002 IMLS National Leadership grant to collaboratively select and digitize over 7,000 photographs.
We are still compiling questions about searching the Historic Pittsburgh Web site that will lead to the creation of more intuitive Help pages and guidelines. Comments and questions also help us to determine how users learn about our site, whether any of our publicity has brought users to our site, and how we can better market our digital collections in the future. Repeated requests from our users to buy reprints of our material has encouraged us to provide a service where facsimile reprints of many of our online maps and books are for sale.
There are psychological benefits as well. Maintaining contact with our users simply makes us feel good. We know that real people with real information needs are using our site. Otherwise, it would be much too easy to lose touch with the human factor that led us to produce these digital collections in the first place.
RELATED ARTICLE: What's in Historic Pittsburgh
* Almost 500 monographs searchable by full text and/or bibliographic information
* Over 1,000 plat maps searchable by street and/or building name
* Nearly 300 finding aids encoded in EAD representing collections held by the University of Pittsburgh's Archives Service Center and the Historical Society of Western Pennsylvania
* Census schedules for Pittsburgh (1850-1880) and Allegheny City (1850-1870)
* Chronology of Pittsburgh comprising over 3,100 entries (1717-2000)
* Catalog to the collections held by the Historical Society of Western Pennsylvania
* Photographs of Pittsburgh: Coming soon!
Anna Maria Mihalega is the digital production librarian in the Digital Research Library at the University of Pittsburgh's University Library System. She holds an M.L.I.S. from the University of Pittsburgh. One of her duties is coordinating responses to e-mail questions from users of the Historic Pittsburgh Web site. She can be reached at email@example.com. Edward A. Galloway is the coordinator of the Digital Research Library at the University of Pittsburgh's University Library System. He holds an M.L.I.S. from The University of Texas-Austin. As project manager of Historic Pittsburgh, he replies to administrative questions including reproduction requests and permission rights, and also tracks and reports usage of the site. His e-mail address is firstname.lastname@example.org.…