Thursday, January 16, 2014

Shared Print Monographs: The Question of Scale

Over the past three years, my partners and I at Sustainable Collection Services (SCS) have worked on a number of shared print projects for books:

  • MI-SPI: Michigan Shared Print Initiative (7 libraries)
  • California State University (6 Los Angeles-area libraries)
  • Connect New York (12 libraries)
  • Maine Shared Collection Strategy (9 libraries)
  • TUG: Tri-University Group (3 libraries)
  • CI-CCI: Central Iowa Collaborative Collections Initiative (5 libraries)
  • WRLC: Washington Research Library Consortium (9 libraries)
  • VIVA: Virtual Library of Virginia (12 pilot libraries)
  • ALI: Academic Libraries of Indiana (34 libraries)

Some of these projects are complete, some in progress, some just beginning. But we have enough experience to begin to identify patterns of success as well as some of the challenges that arise as projects increase in size. Complexity and data variability grow with the number of records (often from different systems and subject to different data management practices) under analysis. Communication and decision-making require more time and effort as the number of participants increases. These factors raise a pressing question: Are there practical limits to the size of shared print monograph projects?

In particular, I am interested in comparing the SCS experience of working locally or regionally at relatively modest scale with the "mega-regions" approach so well described by Brian Lavoie, Constance Malpas, and JD Shipengrover in Print Management at "Mega-Scale": A Regional Perspective on Print Book Collections in North America. [PDF]

In some respects, SCS is working the shared print problem from the bottom up, seeking to address immediate needs of specific institutions. SCS data sets have ranged from 2 million to 7 million records, shared across 5-34 libraries. In contrast, the OCLC Research folks are looking at the issue systemically, exploring an over-arching strategy that posits sharing on a much larger scale. A mega-regions picture of California, for instance, potentially involves 22 million records and more than 1,500 libraries--a very different order of project. There is a great deal to be learned, both practically and conceptually, from work at both ends of this spectrum.

Some key topics to be considered in future posts include:
  • Relationship of a mega-region to existing 'trust networks' 
  • Discovery and delivery infrastructure: can we serve users effectively at the mega-regions level?
  • Challenges of scaling decision-making and communication
  • Shared access and/or ownership across state lines
  • Data work may scale up well, but complexity and errors can grow proportionately
  • Limits of data: is a holding actually on the shelf? Differences between local holdings data and WorldCat?
  • Harmonizing data from various systems, practices, date ranges
But for today, I'm interested in advancing this proposition for discussion: responsibility for archiving of print monographs is best shared at the mega-regional or national level, while responsibility servicing low-use print monographs is best shared at the micro-regional or local level. This hypothesis rests to some degree on a distinction I have previously made between "archive copies" and "service copies": Collection Security and Surplus Copies; Library Logistics: Archiving and Servicing Shared Print Monographs.

I remain convinced that we need to think about the archiving and servicing functions differently. They have very different operational requirements, and it's unclear whether a copy can serve both purposes, or whether we need to designate some archive copies and some service copies. Somehow we need both to assure the integrity of the print archival record, and simultaneously free ourselves to experiment with new approaches to access and delivery. 

The management of the "service" function may change dramatically as our profession integrates the "active service vision" that Emily Stambaugh of the California Digital Library outlined so compellingly last fall, in a presentation on Curating Collective Collections: Reinventing Shared Print. Over time, the need for service copies may decline, as users become more comfortable with book-length content in electronic form, as chapter-level delivery and print-on-demand become more viable, and as expectations change. But for the forseeable future, proximity and rapid delivery remain important, even for little-used materials.

Archive copies, on the other hand, should rarely, if ever, need to be moved. The point is to have a sufficient number of print copies to back up Hathi Trust and other digital repositories, and to assure that the titles remain available for the long term. Responsibility for the archival function could be shared across much larger regions and much larger groups of institutions.

Our current shared print projects tend to emphasize access and service. In SCS shared print projects to date, nearly every group has committed to retain a minimum of two holdings of every title currently held within the group (assuming there are two to begin with, of course). This has occurred even in the smallest groups, as librarians seek to assure that a print copy, if no longer in their own stacks, resides nearby, preferably within an existing resource-sharing network with proven partners, and deliverable within 24-48 hours. And while collection managers appreciate the context provided by total US or Canadian holdings or knowledge that a title has been securely archived by Hathi Trust, so far these factors have exerted minimal influence on the group's retention thresholds. The shared print preferences we've seen to date look like this, reading from the bottom up:

Shared Print Priorities: proximity, existing trust networks, rapid delivery

But these groups are also hedging their bets on the archiving front. The preference for two holdings, even among the smallest groups, reflects the recognition that one holding may be missing or damaged--and the commitment to their users that shared print will not result in the loss of a single title from the group's existing collection. They are in effect archiving the local group's shared collection, and in many cases making explicit retention commitments for those titles. As those MARC 583 fields are updated in WorldCat, we see the genesis of the "intentionally retained" collective collection; i.e., a local service-oriented solution morphing into a contribution to a broader archival solution.

These are early days in rethinking the role of local print, and at present there are many surplus copies of many titles in the collective collection. It's possible to retain two holdings of every title within a small group of libraries and still find significant opportunity for withdrawal of unused material. It's possible to confine shared print to existing trust networks and still make progress. And this micro-regional approach may continue to make good operational sense as long as print books need to be put in the hands of readers. The two smallest circles in this diagram may remain the realm of service copies.

But as explicit retention commitments are made in these smaller groups, it's also possible to glimpse how the collective print collection might evolve. As the need to service low-use print declines in frequency (because of new ways of delivering that content), the eleven mega-regions may be exactly the construct we need to organize and distribute the archival function across the country. The emerging Hathi Trust print monographs archive could utilize this framework, and retention commitments expressed in WorldCat could help collection managers recognize where archive copies have been secured and where they are still needed. As the overall number of copies declines, stewardship would gravitate toward larger areas and groups.

For now, progress can be made safely at both ends of the spectrum without fully resolving the long-term questions. But the question of viable project size--in terms of participation, geographic reach, and data analysis-- remains interesting, and in my next post I will offer some ideas about scaling of that kind.