Wednesday, July 27, 2011

Let Nothing Be Lost

Photo by Philippe Artero
We all have our passions. Good songwriting is one of mine. My pantheon includes grizzled veterans like Dylan and Steve Earle, relative newcomers like Conor Oberst, minor rocking word-mongers like Ian Hunter, Paul Westerberg, and Graham Parker, and choice imports like Paul Kelly. It's a long list of semi-obscure names, from Todd Snider and Peter Case to Butch Hancock and Bill Morrissey.

Bill Morrissey died a traveling musician's death this past weekend, alone in a hotel room in Georgia, on his way back to New Hampshire from a series of gigs in the South. Whether his failing health or the hellhound on his trail caught up to him really doesn't matter. He's gone, and the world is down one quirky voice, one wicked sense of humor, one consummate lyrical craftsman -- and, oddly enough for folk music--one subtle arranger of horns.

Bill's work remains, though: 12 recordings, a novel published in 1997 by Knopf, and his now-spectral presence on YouTube. Morrissey somehow grafted the jaunty blues of Mississippi John Hurt onto a small-town northern New England world. You could say he originated a modest new genre: the north-country blues. Try "Night Shift." But his songs also showed wit and verve, like "Big Leg Ida." He was a small-scale musician with a unique sound and sensibility, the sort of artist who is an acquired taste, and whose works might easily disappear.

As with most small-scale artists, there are inconsistencies in his catalog, and it is easy to miss important distinctions. His first album "Bill Morrissey" was released on LP in 1984 by Reckless Records, a small label based in Cambridge, MA. Its 12 tracks exude that first-album mix of energy and pent-up ambition. Reckless 1917 (its catalog #) never made the transition to CD. Instead, in 1991, Morrissey re-recorded the same songs (plus a few others) directly to DAT (remember that format?) for a new label. These new versions became the CD version, which was released as Philo 1105.

Both recordings are called "Bill Morrissey." Both contain the same songs in the same order. But Reckless 1917 and Philo1105 are very different works. Seven years and two other albums separate them. The singer's take on the songs has softened, phrasings have become more subtle. There's more experience, but less bite. The guitar is lower in the mix. All these years later, it is fascinating to compare the two, and to witness the changes that time and continuing work have wrought on the artist. This is only possible because both versions still exist.

But that shared existence is tenuous. I suppose that's why I dwell on Bill's legacy in the context of deselection. This is content I care about and know a lot about. I believe it matters that his full range of work be preserved and remain accessible to anyone who may be interested. I happen to own that original LP, because I bought it back in 1984. It has languished on my unweeded and largely untouched shelves of vinyl for many years. But yesterday I needed it, because the 1991 version of "Barstow" on my MP3 player didn't seem quite right. I needed something more like the first time I heard it from the tiny stage at The Stone Church, before it was ever recorded. I needed something very specific.

This is an experience common to any expert in any field. This is why we have to be so careful about the scholarly and cultural record. Minute variations can matter. In comparing these two versions of "Bill Morrissey" without sufficient care, it would be easy to conclude, erroneously, that they were the same--and to discard one of them. As we begin to draw down our redundant library collections, we need to be exacting in our determination of what is redundant and what is unique. We need to be certain that we're not discarding some irreplaceable piece of the record, even if it's only of interest to a handful of experts or zealots. While ultimately we may not need many copies of anything, it is essential that we keep at least one copy of everything--even if that work has not been used for decades.

In my world, it's bad enough that Bill Morrissey the man has gone to join Django and Robert Johnson. I'm glad not to have compounded the loss by tossing an LP I hadn't touched for years. Bill's record is complete, if only in my basement. As librarians, we need to recognize that this fear of irretrievable loss exists for every corner of our collections. The risk of loss is real. We need to recognize this and take commensurate care as we manage our collections down. We all have our passions.

Monday, July 18, 2011

The Sweet Spot

"A sweet spot is a place, often numerical as opposed to physical, where a combination of factors results in a maximum response for a given amount of effort." --Wikipedia

The term is most often used in relation to baseball, tennis, or (one hears but does not comprehend) cricket. But virtually every activity, including the humble task of responsible deselection, boasts an area of maximum effect for effort applied. In other words, it is easy to identify a substantial number of withdrawal candidates with a handful of conservative criteria. These data points are not only straightforward to decide upon and generate, they also demonstrate that even a highly conservative approach can yield appreciable results.

Consider these recent projects run by Sustainable Collection Services. In each case, the library chose a set of attributes to apply against its circulating monograph collections. All of these criteria were later refined or adapted by subject or location, or in response to the initial results. But the initial results themselves are quite instructive. (Please note that these are US libraries; the chosen factors reflect that perspective.)

Library A chose these attributes for its first-pass deselection candidates:
  • No circulations since1998
  • Publication date: 1999 or earlier
  • More than 100 holdings in US libraries (WorldCat data)
  • Not listed in Resources for College Libaries
  • Not reviewed in CHOICE
Library A Result: 17% of circulating monographs, yielding approximately 47,000 withdrawal candidates.

Library B chose a slightly different set of attributes:
  • No circulations
  • Publication date: 1999 or earlier
  • Last use: 1999 or earlier
  • More than 100 holdings in US libraries (WorldCat data)
  • At least one other holding in consortium
  • Not listed in Resources for College Libraries
  • Not reviewed in CHOICE

Library B Result: 31% of circulating monographs, yielding more than 60,000 candidates.


Library C chose an even simpler variation:
  • No circulations
  • Publication Date: 1999 or earlier
  • More than 100 holdings in US libraries (WorldCat data)
  • 2 or more other copies in consortium

Library C Result: 10% of circulating monographs, yielding more than 70,000 candidates. 


Although the same criteria are not universally applied, they are similar and consistently conservative. Candidates are indisputably little-used. They are indisputably widely available, and definitely available locally. There is no reliance or even reference to digital versions of this content. This combination of factors involves virtually no risk, but nonetheless creates a "maximum response." Although the percentage of the total collection varies from 10% to 31%, tens of thousands of books are identified as candidates in each case. The result: a zone of immediate action. A sweet spot.

Some caveats apply. (After all, baseball bats can also break only inches from the sweet spot.) For one thing, all disciplines are not equal, either in book use or in the importance of monographs to their users. Deselection criteria in Art will need to vary from those in Chemistry. On the other hand, project managers may want to know what happens if titles that have circulated once since 2000 are included, or how many more titles may be eligible if the number of US holdings is set at 50 rather than 100. The ability to interact with the data to model different scenarios can help refine the withdrawal/storage criteria.

It is also critical to understand the context before acting on data, even if it is this compelling. It is important to understand the commitment of consortial partners and other libraries to these same titles. It is equally important to share that responsibility by retaining some of them. On a regional level, last-copy commitments are being worked out now in many places. On a national level, the movement to use the MARC 583 Preservation Action Note to disclose such commitments is growing. The eventual inclusion of separate holdings symbols for shared print storage facilities will make the context clearer still.

But we shouldn't wait too long. Some action is possible now. Look again at the criteria outlined above. We're talking about books that have remained untouched for more than ten years. All of them were published more than 12 years ago. With more than 100 copies (often many more) in the US. With at least 1-2 other copies in the region. Tens of thousands of books that fit this very conservative description, in library after library. Yes, we need to be careful. We need to base decisions on data. But if this isn't a sweet spot, then... I'm an expert on cricket.

Wednesday, July 13, 2011

Solving One Problem

"You see this? [Holding up a bullet]. This is this. This ain't something else. This is this." -- Michael (The Deer Hunter)

Anyone who is paying even cursory attention to academic libraries knows the story. We have entered an era of consolidation: aggregation of content, metadata and user demand, with library services migrating to the network level. End-to-end solutions are the order of the day. Webscale management systems promise to take library data and transactions to the cloud. Vendors of every stripe are vying to build the one discovery or management system that answers all needs. Systemic change is underway, and there is real value to be created and reaped as this transition progresses. As a new vendor, Sustainable Collection Services plans to create and consume some of that value ourselves. 

But at times the scale and complexity of it all threaten to overload the circuits. It's hard to wrap one's mind around all the permutations. Every task or interaction seems to depend on some other task or operation. Simple conversations devolve into acronym-ridden gibberish. What's most frightening is that sometimes we actually seem to understand one another! In twelve years as a workflow and organizational design consultant, I've become accustomed to swimming in complexities and dependencies, striving to master the intricacies of process and technology and to align them with changing organizational priorities and fiscal realities. Confusion and conflict are constant companions, even when the results are good.
Well, where's the fun in that? It can be ridiculously hard. In response to both of these realities, my partners and I at Sustainable Collection Services have set a more modest goal: to solve a single problem to the best of our ability. One problem: This is this. This ain't something else. This is this! This is ours.

There are plenty of problems ripe for the picking. We choose this one: managing down print monograph collections. Many libraries suffer from stacks crowded with books that are little-used. Many copies of these same books are held elsewhere in the region or country. Space is needed for other purposes. Keeping unused print books on the shelf costs money. Many libraries could benefit significantly from reducing their onsite print collections. But this needs to be done carefully and responsibly, to assure that no content is lost to users. This is a situation we understand. This is a problem we can deal with well.

Deselection decisions involve many factors that are best considered simultaneously. How often has this book been used? Are there other copies in my region? How many copies are in the US collective collection? Have other libraries made a retention commitment for this title? Does it appear in Hathi Trust, and if so, is it public domain or in-copyright? Does it appear on authoritative title lists? Are commercial eBook versions available? Does it need to be added to my regional print archive? Was it written by a faculty member? 

Informed deselection decisions require assimilation and presentation of this diverse deselection metadata for each title. Efficient decision-making requires the ability to build rules based on this metadata, i.e., to move away from title-by-title scrutiny. Rational decision-making requires customization of deselection rules by discipline or location. Prudent decision-making requires the ability to iterate and refine rules and to view provisional results. Locally-aware decision-making requires the ability identify both withdrawal candidates and preservation candidates that suit a library's mission, priorities, and acceptance level.

We founded SCS solely to create and deploy a decision-support tool with these characteristics. For our first few customers, SCS has produced Collection Summary reports and Withdrawal Candidate Lists, often in several iterations. By doing this, we are learning which criteria and which data are useful to most libraries. We are eager to work with more libraries on this basis. But we have also begun to design a web-based application that enables libraries to interact with their own deselection metadata, and to predict the impact of various scenarios. On behalf of my partners Ruth Fischer, Andy Breeding, and Eric Redman, I invite you to take a closer look at our reports and services.

Wordle Tag Cloud of SCS Content

It is immensely satisfying to delve so deeply into this one issue, and to focus on solutions to a single problem. But it is also very effective. We are working daily to become experts in responsible monographs deselection. We can read everything on the topic. We can write about it in detail. We can confer with individual libraries. We can experiment. We can adapt. Best of all, the single-threaded approach gives SCS great clarity in our priorities and imposes a healthy discipline on our activities. We are not seeking to solve all the problems a library may face. We are not even seeking to solve all the problems related to deselection; journals and Government Documents are not on our agenda. SCS is all about monographs. This is this.

Tuesday, July 5, 2011

Disturbing Dust and Data

There must be something in the Australian autumn air.

On March 8, 2011, the Sydney Morning Herald reported that the University of New South Wales Library "is throwing away thousands of books and scholarly journals as part of a policy that critics say is turning its library into a Starbucks." The initiative, which aimed to remove 50,000 volumes per year from the stacks, outraged faculty, students, and librarians.  

On one side: "They're getting rid of books to make space for students to sit around, have lunch, and plug their laptops in."  

On the other: "The library has an ongoing program to remove print journals where online archival access is provided. Our academic community prefers to use the online versions and they use them very heavily."

For the record, UNSW's deselection policy prohibits discard of the last Australian copy of any book.
At about the same time, the University of Sydney announced plans to remove 500,000 low-use books and journals as part of a major renovation of its Fisher Library. University Librarian John Shipp noted that "volumes not borrowed in the past five years will be removed [to remote storage]." He referred to a "dust test" that indicated that books not borrowed are also not read. No books will be discarded; they will be moved to one of two offsite locations.
Picture by Melvyn Knipe
In response, University of Sydney students organized a protest on Facebook. They planned a "mass-borrowing" action to save the books and "disturb the dust", with each of the expected 50 protestors checking out the maximum allowable number of books. In the event, more than 500 students turned up. It is unclear how many books were checked out as part of the protest. An average of 10 per borrower seems like a reasonable estimate, although one woman reportedly arrived with a book trolley and the intent to save an entire section. Central News Magazine described the protest as follows:
Students and staff who oppose the move said Wednesday’s action was an attempt to prevent their removal.
“Our strategy is to borrow as many ‘old’ books as possible at lunchtime on Wednesday, to highlight the ridiculous notion that books have ... an expiry date,” the [protest] Facebook page states.
It is certainly laudable that undergraduates in particular were roused to defend their print collection. This speaks to the value that the idea of an academic library holds in users' imaginations. But between idea and reality falls...the dust. No matter how much anyone wants it to be otherwise, the fact remains that these 500,000 books have not been used for at least five years--and in many cases much longer. According to Paul Courant's estimate, it costs $4.26 per volume per year to retain these low-use titles in central stacks. Removal to a high-density storage facility reduces that cost to $.86 per volume per year. These books will remain available to users, and no content is being lost or even put at risk.

There are also opportunity costs that argue against the status quo. Clearly, like most academic libraries, the Fisher Library needs more space for students to study and collaborate. On upper floors, the stacks are reportedly too close together to allow adequate access for disabled users. Wider aisles require fewer shelves. The library is legally bound to comply with this mandate. And, of course, many of those students "having lunch and plugging in their laptops" are in fact accessing the library's electronic resources. All in all, University Librarian John Shipp has made a good case for a sensible proposal--one that balances responsiveness to users and collection integrity.

The students' response, led by history majors, appears both heartfelt and media-savvy. The organizers have clearly recognized that use (or rather, lack of use) determines how many books and which books will move offsite. Checking out thousands of older titles is an inspired strategy. It beefs up the circulation statistics, and may exempt thousands of titles from being moved to storage. It makes for compelling photographs and good news stories. Young people are rising up to protect our cultural and scholarly record!

Picture by Melvyn Knipe
But while the protest makes a good point with good theater, it also seems disingenuous. The "use" generated by this mass borrowing action is entirely artificial. It's hard to imagine that the "protest books" will be read, consulted, or cited in the same manner as those that actually relate to an assignment or research project. They are being used in a different way, to make a rhetorical point. (Of course, one might argue that this is the highest use of a book!) They are being used as a symbol or a category; the actual titles don't matter, so long as the volumes are dusty. In the end, this is not an argument about content at all. It's an argument about the idea of a library versus the reality of a library.

This is an important debate to have within our communities. But the discussion, however spirited, needs to occur honestly. One consequence (unintended?) of the mass borrowing: distorted circulation data. Thousands of titles now appear to have circulated that, if we are honest about it, would not otherwise have done. We now have a somewhat false -- and essentially romantic -- picture of collection use. The picture has been shaded toward what University of Sydney students and faculty would like it to be. The ballot box has been stuffed, a thumb placed on the scale. And while the effect in this case is not statistically significant (perhaps 1%-2% of the proposed withdrawal candidates have been affected), we need to be aware that the data, along with the dust, has been disturbed.
One checkout..ah-ah-ah...
As we grapple with the future of print collections, it is vital to retain the distinction between what we wish were happening and what is actually happening. Circulation and in-house use data, coupled with data on lifecycle costs, offer the most reliable picture of what is actually happening. That's the best place to start, whatever we ultimately decide to least according to our recently-hired Director of Statistics, pictured at right.