Making the Case for Weeding

Collection managers at work last week

Last week, our small world of print monographs management approached its own instance of 'harmonic convergence'. A number of unrelated events coincided to illuminate the changing role of print collections in the academic community. Heat, light, and even the threat of fire emanated from several angles, and the humble task of library weeding reached the edges of the mainstream media. Overall, despite their sometimes unexpected origins, these developments advanced a timely and necessary discussion. Consider the forces aligned on this topic:

• Hathi Trust Constitutional Convention:An excellent summary from Feral Librarian includes the good news that "HathiTrust will establish a distributed print monograph archiving program among HathiTrust member libraries." This supplements Hathi's pioneering archive, that now includes secure digital versions of "over half of the collective holdings of ARL libraries."  John Wilkin's opening remarks to the Convention [pdf] emphasized the power of large-scale collaboration toward "an increasingly comprehensive shared collection." The addition of a print archive will assure that no content is lost. Check the Twitter hashtag #htcc for more.

Cultural critics?

• 6 Reasons We're In Another Book-Burning Period in History: Meanwhile, incomprehensibly, Cracked.com, the web version of a Mad magazine imitator, served up a provocative post by S. Peter Davis. The piece is a strange mix of inflammatory and informed, and occasionally even humorous. It taps into the outrage that can be engendered by discarding and destroying books, no matter how strong the rationale. As of today, the article has been viewed 665,668 times--far beyond the audience reached by librarians wrestling with these issues. Davis sees that weeding may be necessary; he objects to the destruction of withdrawn volumes.

• 'Hard Choices: Do Libraries Really Destroy Books?" An NPR blog called Monkey See brought the "Cracked" entry to a still wider audience but with a more balanced perspective and some actual reporting. The author spoke with Betsy Simpson, President of Association for Library Collections and Technical Services (ALCTS). Ms. Simpson articulated the academic library perspective fully and well, noting the library's "mission to preserve the cultural record' in the face of simultaneous need for [user] "space to interact and collaborate" and the existence of many unused copies of the same titles.

• ARL Membership meeting: The Chronicle's 'Wired Campus' blog reported that the agenda for the Association of Research Libraries' Directors annual meeting in Washington, DC included presentations on the Digital Public Library of America, an update on HathiTrust, and on the changing nature of research. In a session on 'Rebalancing the Investment in Collections', H. Thomas Hickerson, the Vice Provost and University Librarian at University of Calgary, noted that 'the comprehensive and well-crafted collection is no longer an end in itself.' Ed van Gemert, deputy director of libraries at UW-Madison, noted: "We simply can't afford to do work separately that could be done collectively."

Not weeding has consequences, too

All in all, not a bad week for the concepts of weeding and collaboration: a healthy mix of attention, controversy, innovation, misunderstanding, good sense, and passion. (More like dissonant convergence, I guess.) Perhaps these events are most usefully seen as a challenge to all of us engaged in deselection, shared print management, and digital archiving. The challenge is to continue to clarify our own thinking, refine our message, and get better at making our case not only to ourselves but to the wider community. Weeding is necessary. Weeding is responsible. Weeding can be made safe. Collection security will assured  collaboratively.

This is the main reason that we at Sustainable Collection Services have adopted an approach to monographs deselection that relies on data. Taken together, the number of copies held globally or within a specific region, circulation and in-house use over time, and the existence of a secure digital copy can create a safe and suitable list of withdrawal candidates. Disposition options are numerous, and each involves different trade-offs and costs. But, as Betsy Simpson and even Cracked.com indicated, the underlying problem is real. We need to explain why some books may need to be removed, and what is to be done with them. We need to make the case.

Here are my suggestions for talking points on the rationale for weeding. I'll add supporting data points in another post. First, let's see if this high-level message works. Thoughts?

• Electronic resources are the dominant information format
• Print use is low and declining
• Library stacks and storage facilities are crowded
• Library space is wanted for other purposes
• Keeping print books on the shelves is expensive
• Many copies of the same titles exist in many libraries
• Secure digital versions exist for millions of titles
• The infrastructure to support sharing exists and is growing
• Savings from adopting shared print can support other library services

Library Logistics

Recently, in my thinking about servicing shared print monograph collections, the concept of logistics has loomed large. In order to satisfy user expectations for delivery of books from regional repositories, these centers will have to operate as fast and efficiently as Amazon, UPS, or FedEx. In addition to assuring the persistence of the scholarly record, the core competencies of regional print repositories will become inventory control and speed of fulfillment. Digitization and print-on-demand will follow in time. In this deeply unromantic view of library services, shared print service centers will become a key link in the long-tail supply chain.

In a technical definition of logistics drawn from Wikipedia, service centers must achieve 'high due-date reliability, short delivery times, low inventory level and high capacity utilization.'  To my mind, this frames the management challenges around shared print collections very accurately. From Paul Courant and others, we are learning just how expensive it is to manage low-use print material. We need to lower those costs while assuring continued and convenient access. To do so, libraries need to embrace logistics.

Always ahead of the curve, Lorcan Dempsey introduced the phrase 'library logistics' in November 2004, and amplified its meaning in subsequent posts. To excerpt from a couple of his attempts at definition:

Logistics is about moving information, materials, and services through a network cost-effectively. Resource sharing is supported by a library logistics apparatus. [...] Increasingly, as libraries look at shared solutions for off-site storage, [...], digitization and archiving, they run into logistics and supply-chain management questions.

There are a number of interesting permutations to this idea, especially in relation to shared print monograph collections. Here are some steps I believe we need to take to begin to benefit from a very different approach to managing tangible collections.

Separate 'archive copies' from 'service copies.' This distinction is underdeveloped in discussions of shared print. I suspect this is largely because those efforts are at present focused on journals. With journals, a single copy can often support both archiving and distribution, because article scanning and document delivery are well-developed systems, and because articles are shorter than books. This allows a print volume to be cared for, but also for its contents to be disseminated. In August, the WEST group, which has wrestled with these issues thoroughly, codified its recommended practices in its "Access Guidelines for WEST Archives [pdf]. For delivery methods, it recommends, in order of preference:

1. Provide electronic document delivery (including color scans when appropriate)
2. Provide photocopies
3. Loan the physical issue or volume to the borrowing library for building use only.

This make excellent sense for journals. For now, though, this approach is not well suited to books. The length of texts and copyright issues make scanning, photocopying, and building-only use much less practical.

Monographs will require a different service model, at least for the foreseeable future. It will most often require the delivery of a copy to a user. This model introduces both risk (of loss or damage) and significant logistical challenges. The risk is easily mitigated: designate a shared copy or copies 'archival' and prohibit their use, except for subsequent non-destructive digitization. For books, one or more dark archive copies, supplemented (where possible) with a secure digitally-archived version in HathiTrust, would satisfy preservation and security needs. This step is foundational; no service layer can be built until the archiving imperative has been met.

The logistical challenges can then be met with an active, well-managed inventory of  'service copies.' Since most titles in offsite storage have seen little or no user demand, the number of service copies needed in a region may be quite small. This would depend on the size of the user base and the number of libraries relying on the regional facility.. This is where inventory management techniques developed in other contexts could begin to contribute to a radically different service environment.

Raise the bar for regional library service centers. Distribution and supply-chain management are highly evolved in other industries and sectors. Servicing of shared print collections could benefit enormously from the expertise of large-scale book distributors like Ingram, Follett, or Baker & Taylor. The library world in general could learn from logistics experts at UPS, Amazon, or other companies. Service from regional library centers should be built to include 24-hour delivery direct-to-user, email order confirmation and tracking capability, real-time display of availability, and perhaps even the option to purchase via partner relationships.

One important question is how many such centers are needed. Amazon services the entire country from a handful of warehouses. UPS positions its distribution centers near airports and highway interchanges to enable rapid delivery. In short, once we separate servicing shared monograph collections from archiving them, we can manage the service copy inventory according to best practices drawn from other industries. A very few such centers, optimized for 24-hour delivery and long-tail inventory management, might serve the whole continent. Over time, these would be obvious nodes for print-on-demand production as well.While service copy distribution centers would cost money to establish and run, much of that cost could be offset by space and time savings in participating libraries. Cost would also suggest that a very few, high-volume centers would make the most sense.

Some library systems operate along these lines already, using a combination of storage facilities and courier services. As a small-scale example, the Tri-Universities Group (TUG) in Ontario provides delivery 'by midafternoon the following weekday' to its three member libraries from a shared storage facility in Guelph. The OrbisCascade Alliance provides 48-hour delivery within its two-state membership of 36 libraries. University of Missouri System Libraries Depository (UMLD) processes up to 100 book requests per day, with requested items leaving the facility within 24 hours of placement. Many similar initiatives exist and are constantly being improved. But the benefits relating to greater scale, automation, and shared costs remain intriguing.

Provide 24-hour delivery directly to any authenticated user in the region. Users will be much more inclined to accept that large portions of the print collection are stored elsewhere if it does not affect their personal workflow. For service copies stored in shared print repositories, 24-hour direct to user delivery would mostly eliminate this concern. While the shipping cost may be higher than at present, this would make offsite storage more palatable, and potentially almost invisible. Delivery direct-to-user would eliminate circulation desk mediation, holds, and staging for pickup by the user. A higher volume of transactions would lower the per-transaction cost, so this again suggests consolidation of supply in fewer, larger regional centers. Delivery and return could be standardized, and bid to local couriers, UPS, DHL, or FedEx. 

Manage inventory like a book distributor. Booksellers are experts at determining which titles are moving and which are not. Although these are likely to be long-tail operations (many titles, very few copies of each), it remains true that faster-moving titles require more copies on hand, and slower-moving titles fewer copies. Because archiving is handled separately, there is much more latitude to manage service copies based on their activity level. Titles that do not circulate might become candidates for print-on-demand or digital delivery only. Or, they might simply be retained in the warehouse, where the $.86 per year estimated cost of ownership is amortized across all the libraries in the group. 

In the end, to adapt another phrase from Lorcan, broad adoption of library logistics could confer the benefits of the 'network effect' on low-use monographs, driving costs out of the system while improving service to users.