NASIG 2011: Using Assessment to Make Collection Development Decisions

Speaker: Mary Ann Trail & Kerry Chang FitzGibbon

It is not in the interest of faculty to cut journal titles because it may be perceived as an admission that it is not needed. With relying on faculty input for collection decisions, the collection can become skewed when certain faculty are more vocal than others.

When a new director arrived in 2000, they began to use more data to make decisions. And, the increase in aggregator databases and ejournals changed what was being collected. In addition to electronic publishing, electronic communication has changed the platform and audience for faculty communicating with each other and administrators, which can be both good and bad for library budgets.

In 2005, after some revision of collection methods, cancellations, and reallocation, they went to a periodicals allocation formula. This didn’t work out as well as expected, and was abandoned in 2008.

As a part of their assessment projects in 2008, they looked at the overlap between print and electronic titles to see if they could justify canceling the print in order to address the budget deficit. Most importantly, they wanted to proactively calm the faculty, who were already upset about past cancellations, with assurances that they would not lose access to the titles.

They used their ERMS to generate overlap analysis report, and after some unnecessary and complicated exporting and sorting, she was able to identify overlaps with their print collection. Then she identified the current subscriptions before going to the databases to verify that the access is correct and noted any embargo information. This was then combined with budget line, costs, and three years of usage (both print and electronic for non-aggregator access).

They met their budget target by canceling the print journals, and they used the term “format change” instead of cancel when they communicated with faculty. Faculty showed more support for this approach, and were more willing to advocate for library funds.

Did they consider publications that have color illustrations or other materials that are better in print? Yes, and most of them were retained in print.

Did they look at acquiring other databases to replace additional print cancellations? No, not with their funding situation.

What was the contingency plan for titles removed from the aggregator? Would resubscribe if the faculty asked for it, but funds would likely come from the monograph budget.

VLACRL Spring 2011: Patron-Driven Acquisitions panel

“Selectors are more fussy about the [ebook] platform than the students.” – Nancy Gibbs

Speakers from James Madison University, Duke University, and the College of William & Mary

James Madison University has done two trials of patron-driven acquisitions. The first one was mainly for print books that had been requested through interlibrary loan. If the book is a university press or new (past two years) imprint, they rush order it through an arrangement with the campus bookstore. The book arrives and is cataloged (actually, the book gets cataloged when it’s ordered, saving additional processing time) in about the same time it would take if it was coming through the ILL system, and most of these books ended up circulating frequently with renewals.

Their second trial was for ebooks through their book jobber, Coutts, and their MyiLibrary platform. They used the same parameters as their approval plan and set it up like most PDA ebook programs: drop the records in the catalog and after X number of “substantial uses” (i.e. not the table of contents, cover, etc.) the book is purchased using a deposit account fund. They excluded some publishers from the PDA process because they prefer to purchase the books on the publisher’s platform or have other arrangements (i.e. Gale or Wiley). If your library needs certain fields in the MARC record added, removed, or modified, they recommend that you have the vendor do that for you rather than touching every record locally, particularly given the volume of records involved.

The ebook PDA trial was initiated last calendar year, and they found that 75% of the ebooks purchased were used 5-19 times with an average of 14.77 per title. Surprisingly enough, they did not spend out their modest deposit account and were able to roll it over to this year. Already for 2011, they are seeing a 30% increase in purchases.

Duke University was one of the ARL libraries in the eBrary PDA pilot program. Out of the 90,000 titles offered, they culled the list down to 21,000 books published after 2006 with a $275 price per title limit. Even with that, they blew through the deposit account quickly. But, they found that the titles purchased were within the scope of what they would have collected anyway, so they added more funds to the deposit account. In the end, they purchased about 348 ebooks for $49,000 – mainly English-language titles from publishers like Wiley, Cambridge, and Oxford, and in areas like business and economics.

Other aspects of the Duke trial: They did not match up the 21,000 books with their approval plan, but used other criteria to select them. They negotiated 10 “clicks” to initiate a purchase (whatever the clicks mean). They were send approval slips for many of the titles that were purchased, but for whatever reason the selector did not choose them.

About 183 (over 50%) of the ebooks purchased were already owned in print by the library. One of their regrets is not capturing data about the time of day or day of week that the ebooks were accessed. It’s possible that the duplicates were accessed because the user was unable to access the print book for whatever reason (location, time of day, etc.). Also, two of the books purchased were already owned in electronic format in collections, but had not been cataloged individually.

Duke has also done a PDA program with interlibrary loan. The parameters are similar to JMU’s, and they are pushing OCLC to include preferred format in the ILLiad forms, as they would like to purchase ebooks if the user prefers that format.

They are also looking to do some topic-specific PDAs for new programs.

The College of William & Mary is a YBP customer for their print books, but they decided to go with Coutts’ MyiLibrary for their ebook PDA trial. This was initially the source of a great deal of frustration with de-duping records and preventing duplicate purchases. After several months and a duplication rate as much as 23%, they eventually determined that it was a time gap between when Coutts identified new titles for the PDA and when W&M sent them updates with what they had purchased in print or electronic from other sources.

In the end, they spent the $30,000 private Dean’s fund on 415 titles fairly evenly across the disciplines. About 45 titles had greater than 100 uses, and one title was used 1647 times (they think that was for a class). Despite that, they have not had to purchase a multi-user license for any title (neither has JMU), so either MyiLibrary is letting in multiple simultaneous users and not charging them, or it has not been an issue for a single user to access the titles at a time.

One thing to consider if you are looking to do patron-driven acquisitions with ebooks is the pricing. Ebooks are priced at the same rate as hardcover books, and multiple user licenses are usually 50% more. Plan to get less for the same money if you have been purchasing paperbacks.

There are pros and cons to publicizing the PDA trial during the process. In most cases, you want it to be seamless for the user, so there really isn’t much reason to tell them that they are initiating library purchases when they access the ebooks or request an interlibrary loan book. However, afterwards, it may be a good marketing tool to show how the library is working to remain relevant and spend funds on the specific needs of students/faculty.

COUNTER book reports are helpful for collection assessment, but they don’t quite match up with print use browse/circulation counts, so be careful when comparing them. Book Report 2 gives the number of successful section requests for each book, which can give you an idea of how much of the book was used, with a section being a chapter or other subdivision of a reference work.

Final thoughts: as we shift towards purchasing ebooks over print, we should be looking at revising and refining our workflow processes from selection to acquisition to assessment.

“Selectors are more fussy about the [ebook] platform than the students.” – Nancy Gibbs

VLACRL Spring 2011: Clay Shirky, Fantasy Football, and the Future of Library Collections

As we shift to a demand-driven collection development approach, we will better be able to provide content at the point of need.

Speaker: Greg Raschke

Raschke started off with several assumptions about the future of library collections. These should not be a surprise to anyone who’s been paying attention: The economics of our collections is not sustainable – the cost and spend has gone up over the years, but there is a ceiling to funding, so we need to lower the costs of the entire system. We’re at a tipping point where just in case no longer delivers at the point of need. We must change the way we collect, and it will be hard, but not impossible.

The old system of supply-side collection development assumes that we’re working with limited resources (i.e. print materials), so we have to buy everything just in case someone needs it 10 years down the road when the book/journal/whatever is out of print. As a result, we judge the quality of a collection by its size, rather than by its relevance to the users. All of this contributes to an inelastic demand for journals and speculative buying.

The new system of demand-driven collections views them as drivers of research and teaching. It’s not really a new concept so much as a new workflow. There’s less tolerance in investing in a low-use collection, so there is an increase in the importance of use data and modifying what we collect based on that use data. The risks of not evolving and failing to innovate can be seen in the fate of the newspapers, many of whom held onto the old systems for too long and are dying or becoming irrelevant as a result.

Demand-driven collection development can create a tension between the philosophy of librarians as custodians of scholarship and librarians as enablers of a digital environment for scholars. Some think that this type of collection development may result in lower unit costs, but the reality is that unless the traditions of tenure and promotion change, the costs of publishing scholarly works will not go down. One of the challenging/difficult aspects of demand-driven collection development is that we won’t be getting new funds to do it – we must free funds from other areas in order to invest in these new methods (i.e. local digital production and patron-driven acquisitions).

The rewards of adapting are well worth it. The more our constituencies use the library and its resources, the more vital we become. Look at your data, and then bet on the numbers. Put resources into enabling a digital environment for your scholars.

Demand-driven collection development is not just patron-driven acquisitions! It’s about becoming an advanced analyst and increasing the precision in collection development. For NCSU‘s journal review, they look at downloads, impact factors, publications by NCSU authors, publications that cite NCSU authors, and gather feedback from the community. These bibliometrics are processed through a variety of formulas to standardize them for comparison and to identify outliers.

For print resources, they pulled circulation and bibliographic information out of their ILS and dropped it into SAS to assess the use of these materials over time. It was eye-opening to see what subject areas saw circulation greater than one over 10 years from the year they were added to the collection and those that saw no circulations. As a result, they were able to identify funds that could go towards supporting other areas of the collection, and they modified the scopes of their approval profiles. [A stacked graph showing the use of their collection, such as print circulation, ejournals/books downloads, reserves, and ILL has been one of their most popular promotional tools.]

As we shift to a demand-driven collection development approach, we will better be able to provide content at the point of need. This includes incorporating more than just our local collections (i.e. adding HathiTrust and other free resources to our catalog). Look to fund patron-driven acquisitions that occur both in the ebook purchasing models and through ILL requests. Integrate electronic profiling with your approval plans so that you are not just looking at purchasing print. Consider ebook packages to lower the unit costs, and use short-term loans for ebooks as an alternative to ILL. Get content to users in the mode they want to consume it. Do less speculative buying, and move money into new areas. It is imperative that libraries/librarians collaborate with each other in digital curation, digital collections, and collective bargaining for purchases.

There are challenges, of course. You will encounter the CAVE people. Data-driven and user-driven approaches can punish niche areas, disciplinary variation, and resources without data. The applications and devices we use to interact with digital content are highly personalized, which is a challenge for standardizing access.

I asked Raschke to explain how he evaluates resources that don’t have use data, and he says he’s more likely to stop buying them. For some resources, he can look at proxy logs and whether they are being cited by authors at his institution, but otherwise there isn’t enough data beyond user feedback.

NASIG 2010: Integrating Usage Statistics into Collection Development Decisions

Presenters: Dani Roach, University of St. Thomas and Linda Hulbert, University of St. Thomas

As with most libraries, they are faced with needing to downsize their purchases in order to fit within reduced budgets, so good tools must be employed to determine which stuff to remove or acquire.

The statistics for impact factor means little to librarians, since the “best” journals may not be appropriate for the programs the library supports. Quantitative data like cost per use, historical trends, and ILL data are more useful for libraries. Combine these with reviews, availability, features, user feedback, and the dust layer on the materials, and then you have some useful information for making decisions.

Usage statistics are just one component that we can use to analyze the value of resources. There are other variables than cost and other methods than cost per use, but these are what we most often apply.

Other variables can include funds/subjects, format, and identifiers like ISSN. Cost needs to be defined locally, as libraries manage them differently for annual subscriptions, multiple payments/funds, one-time archive fees, hosting fees, and single title databases or ebooks. Use is also tricky. A PDF download in a JR1 report is different from a session count in a DB1 report is different from a reshelve count for a bound journal. Local consistency with documentation is best practice for sorting this out.

Library-wide SharePoint service allows them to drop documents with subscription and analysis information into one location for liaisons to use. [We have a shared network folder that I do some of this with — I wonder if SharePoint would be better at managing all of the files?]

For print statistics, they track separately bound volume use versus new issue use, scanning barcodes into their ILS to keep a count. [I’m impressed that they have enough print journal use to do that rather than hash marks on a sheet of paper. We had 350 reshelved in last year, including ILL use, if I remember correctly.]

Once they have the data, they use what they call a “fairness factor” formula to normalize the various subject areas to determine if materials budgets are fairly allocated across all disciplines and programs. Applying this sort of thing now would likely shock budgets, so they decided to apply new money using the fairness factor, and gradually underfunded areas are being brought into balance without penalizing overfunded areas.

They have stopped trying to achieve a balance between books and periodicals. They’ve left that up to the liaisons to determine what is best for their disciplines and programs.

They don’t hide their cancellation list, and if any of the user community wants to keep something, they’ve been willing to retain it. However, they get few requests to retain content, and they think it is in part because the user community can see the cost, use, and other factors that indicate the value of the resource for the local community.

They have determined that it costs them around $52 a title to manage a print subscription, and over $200 a title to manage an online subscription, mainly because of the level of expertise involved. So, there really are no “free” subscriptions, and if you want to get into the cost of binding/reshelving, you need to factor in the managerial costs of electronic titles, as well.

Future trends and issues: more granularity, more integration of print and online usage, interoperability and migration options for data and systems, continued standards development, and continued development of tools and systems.

Anything worth doing is worth overdoing. You can gather Ulrich’s reports, Eigen factors, relative price indexes, and so much more, but at some point, you have to decide if the return is worth the investment of time and resources.