1. Jacob Nadal Executive Director, ReCAP
Discovery to Delivery
OCLC Regional Print Archives Symposium 28
MAR
2014
Dublin, OH
Hello. My Name is Jacob Nadal, and I am the Exec Director of ReCAP, the Research Collections and Preservation Consortium. Today, I am going to talk briefly about the development of ReCAP with an eye to giving a case study in how preservation repositories offer different opportunities at different stages in their development, as the scale and maturity of their operations develops, and as the prevailing challenges in our field shift over time.
As I do this, I am conscious of standing on some substantial shoulders. A great deal of credit is due to Eileen Henthorne, ReCAP’s founding executive director, and the to staff at the partner institutions who have developed ReCAP over the last 10 years and worked on the Discovery to Delivery project I’ll discus in a few minutes. Especially, I want to acknowledge my board members, Jim Neal, Stephen Rittenberg, Karin Trainer, Paul LaMarche, Ann Thornton, and Jeffrey Roth. ReCAP is a tangible sign of their work and commitment to aligning their institutional needs with a set of professional concerns that rise above the boundaries of any particular campus or collection.
Jacob Nadal
Executive Director, ReCAP

2. The New York Public Library
History of ReCAP
The Research Collections and Preservation Consortium was founded by three partner libraries:
Columbia University
Princeton University
The New York Public Library
ReCAP is a high-bay high-density preservation repository that:
Holds over 11 million items at a TWPI of 200+
Circulates an average of 15,000 items per month
Adds 50-60,000 items per month
Never misplaces anything
ReCAP was founded at the turn of the century by three partner libraries, Columbia University, Princeton University, and The New York Public Library. At the time, these libraries were largely facing a problem of storage space, and as an adjunct to that, needed to make the preservation environment and service costs associated with the less used parts of the collection more affordable.
At present, ReCAP holds over 11 million items – from individual monographs and films to volumes of serial publications and a fast growing contingent of archival containers. Each month we circulate about 15,000 items per month, about 1-2% of total holdings, and add around 50-60,000 items.

3. Our new entryway, since the addition of MODS 8-9
To put this into terms native to Dublin, OH, ReCAP is both a shadow side and an elegant expression of Lorcan Dempsey’s observations about the information economy problems that libraries address.
The first of these is scarcity of resources, the second is scarcity of the attention to make use of those resources. ReCAP was created to deal with the partners success in addressing the scarcity of resources problem – they collected more than their facilities could hold. As it has grown, ReCAP is now retrofitting its existing inventory functions to become a tool for sharing its assets more broadly, exposing them to the attention of a larger community of researchers.
Recap in winter
Our new entryway, since the addition of MODS 8-9

4. The obligatory glamour shot
By bringing lots of things together in one place, it’s easier to find and get them, and the per-item costs of “keeping” can be highly optimized. Those low item-level costs constitute major capital costs, however.
MODULE 2: Aisle 12
The obligatory glamour shot

5. A showoff picture, showing the scale of the new mods
The ReCAP modules are built on a massive scale (click: perspective example) and the associated costs force certain issues: sharing fixed costs up front, fairly allocating the costs for partner-specific needs, and planning for a return on that investment over an extended period.
Module 9: Aisle 74
A showoff picture, showing the scale of the new mods

6. ReCAP at Present Partnership shares costs of building and operating
Effective core operations and financial plans
With tremendous credit due to Eileen Henthorne,
ReCAP’s founding Executive Director
Three collections in a shared building, supported by shared services
Independent policies for selection, acquisition, circulation, and management of collections
Capital Projects done until approx. 2020
“Return on investment over an extended period,” is pretty close to my own working definition of preservation: sustainable activities, optimized over time, that maintain or improve the usability of information resources.
In this context, I think “extended” probably means something like “more than two strategic planning cycles” or “beyond the legal mandate for retaining financial records.”
That’s only 10 years or so. With that in mind, it’s instructive to think of ReCAP as just entering its second “extended” phase of development. At present, ReCAP is effectively three collections in a shared building, supported by shared services, with no major capital projects on deck until about 2020.
The partners have independent policies for selection, acquisition, circulation, and management of collections, however.

7. Development Cycles
: ReCAP launches and develops its core services
: ReCAP explores new directions
2014 on: ReCAP implements new services and refines core operations
I’m sure that some of you know this model – “Cynefin,” or “habituation”, something developed by Dave Snowden and Cynthia Kurtz – it’s a framework for thinking about organizational development. It’s especially useful in a place like ReCAP that serves capital-P Preservation, a big mission and vision role, through very specific and measurable operations.
Early on, ReCAP had to go from a bright idea to a fully functioning building (click), and had to bring a known process – the Harvard Depository model (click) – to life with new staff and within a different organization context.

8. Development Cycles
: ReCAP launches and develops its core services
: ReCAP explores new directions
2014 on: ReCAP implements new services and refines core operations
As time went on, that process matured and worked very well. Lots of collections were in ReCAP, and the usage began to climb. This is a function of quantity and serendipity to some extent – the more millions of items you have, the more likely any randomly selected research process is to need some of them. Other factors drive increased use as well, such as improved bibliographic control, or the eventual transfer of low- or moderate-use materials, since ReCAP grows more quickly than the relatively fixed (or declining) amount of on campus shelving space.
During this period ReCAP changed from being a low-use, last resort repository, to the largest single-site accumulation of materials within each partner’s operations. This called for an examination of new service areas. (click)

9. Development Cycles
: ReCAP launches and develops its core services
: ReCAP explores new directions
2014 on: ReCAP implements new services and refines core operations
Interlibrary Loan and Electronic Document Delivery were the first of these to be implemented, and can (click) be understood as the first steps towards a larger vision of a shared collection and expanding access service options.
The rest of this talk will explore those planned developments, but I also want to remind you that this is progressive work. Our topic today is “Selecting for Sustainability” in the context of “Right-scaling” of solutions. One of the things that’s implicit in our new directions is that we continue to refine and improve existing operations, and the Discovery to Delivery report I’ll discuss next provides and important framework in which to do that.
So, just as it took ReCAP about 10 years to reach this point of exploring new directions, it's taken me about 10 minutes to get around to actually to talking about them. This is a deliberate rhetorical ploy. I want to make sure you do expect to see results from what I'm about to discuss, but that you expect those results over the next couple years, not the next couple months.

10. New Directions
With support from Andrew W. Mellon Foundation, ReCAP investigated three opportunities:
Shared collections among the partners
Makes more resources available to each partner, more quickly
Cooperative collection development
Lowers cost to each partner, for access to more comprehensive collections
Long-term preservation commitments
Ensure that the costs sunk into collections and the facility will create long-run value for the partners
And the tools (mostly software) and services required or desirable to realize those opportunities
“Discovery to Delivery” was supported by the Andrew W. Mellon foundation, and resulted in an extensive set of reports and recommendations that are published on our website. You’ll get that URL at the end of the slides. The project identified three major thematic areas for our next stage of development – shared collections, cooperative collection development, and long-term preservation commitments – and included substantial work to define the technology required to bring those thematic areas into real practice.

11. Shared Collections ReCAP holds over 11 million items in trust
Equitable distribution of space:
24% PUL, 38% CUL, 38% NYPL
And annual accessions:
246,369 PUL, 203,041 CUL, 182,102 NYPL*
Majority of items are unique within ReCAP
Approx 10% are duplicates; 1% triplicates
Approx 10,000,000 unique items
Each partner gains access to over 4 million items
* Five-year average
ReCAP holds over 11 million items in trust, and those holdings are already distributed amongst the partners in an equitable fashion. NYPL and Columbia have more items in storage currently, and occupy more square footage, but Princeton actually adds to the collection at a slightly higher rate and as simple function of proximity, gets requests to campus faster.
Approximately 10% of the items in the facility are duplicates; 1% triplicates, so ReCAP has well over 10M unique items, and with a shared collection, this means that each partner stands to gain access to some 3 or 4 million additional items for their users, available on the same terms as everything else in their collections.

12. Obstacles to Sharing
Three ILS installations, using two different systems
Adding 4 million records to any ILS involves prohibitive costs in labor, project fees, and license fees
No common, pre-existing resource-sharing network
Different circulation policies
Different patrons
To achieve that, we have to overcome some obstacles. There is no common ILS or resource sharing system among the partners, naturally they have different circulation policies, and to some extent, this is because they have several different patron groups, between two distinctive private universities and a public research library.

13. POLICY Solutions Reframe existing ReCAP collections:
Shared: Available to all partners and subject to a retention commitment and other agreed policies
Open: Available to all partners but not subject to a retention commitment or other Shared Collection policies
Restricted: Available only or primarily to the owning library (often special collections material)
New accessions to be Shared by default. Open or Restricted status must be specifically set.
Overcoming those barriers requires some policy decisions and some technology development. On the policy side, we are beginning to frame the existing and forthcoming holdins of ReCAP in three ways:
Shared: Available to all partners and subject to a retention commitment and other agreed policies
Open: Available to all partners but not subject to a retention commitment or other Shared Collection policies
Restricted: Available only or primarily to the owning library (often special collections material)
New accessions to be Shared by default. Open or Restricted status must be specifically set.

14. Technology Solutions
Middleware system high-level architecture, developed by HTC Global, Inc.
Middleware pulls metadata from ILS and LAS, normalizes, and exposes to discovery environments
Allows ReCAP Shared to be searched alongside the library’s other collections and licensed resources
Patron initiates an order using discovery environment’s handlers
Middleware translates this to a retrieval request in LAS and creates on-the-fly item records to establish a hold in the ILS
Alongside this change in policy, there has to be a change in technology to facilitate sharing.
One solution would be to just dump a few million new records into everyone’s ILS – but take a moment to contemplate the complexity of that initial records load, the increase in ILS maintenance costs, and think of the records maintenance to follow, and then please check to be sure that any catalogers seated near to you are still breathing.
Instead of that, we are gravitating towards a piece of middleware that will harvest bibliographic data from the partners and expose it to each institution's federated discovery environment, and then translate any circulation activities back to the appropriate ILS.

15. Middleware Architecture 1.0
To give you a sense of where we’re at in development I’ll show a few excerpts from HTC architecture proposal. (click) This shows the connections between the actual systems involved and the three major layers of the system – a presentation layer for staff and patrons, a set of enterprise services to connect these systems and interfaces, and a database and SOLR to enable discovery and circulation transactions. (click) This can also be show in a systems view, with the browser-based OPAC and staff interfaces connecting to load balancers and web services that speak to the ReCAP database, our LAS inventory manager, and the partners ILS systems.

16. WHICH MEANS Batch Data Exchange ILL between partners Now Then
(Known) periods of mismatch between item state and reported status
Hard cutoff times for transactions
ILL between partners
Highly expedited, but still adds 1-2 days of administrative action
Real-time Data Availability
(On query) State and status are the same
Operations optimized to match patterns in transaction data
Partners have immediate access to shared collection
Fulfillment schedule is the same as the partners owned materials
Taken together, this shifts us from an operation that works on batch transactions to one that works on real-time data. This means that we close gaps between actual and reported status, and reduce wait times for materials. This also means that we can experiment with more efficient workflows. At present, certain jobs occur at a particular time because the data are delivered at a certain time. With a real-time data management system, we can take some steps towards lean operations, doing work on-demand or just-in-time, when the data reach a certain threshold.

17. COOPERATIVE COLLECTION DEVELOPMENT
Selectors at all three partners have hand-shake agreements to avoid prospective duplication
Substantial interest in a deeper collaboration, but this is in principle, not in practice, because:
Middleware is needed to create interfaces for cross-collection evaluation
Bibliographic data across ReCAP holdings is needed for discussions about assigning cataloging and collecting responsibilities, and developing shared approval plans
Bib data in three ILS; ReCAP holdings in LAS; Middleware treats LAS as a connector between stores of Bib data.
Beyond the on-site work and patron services, the middleware system will also support shared collection management. Right now, selectors have some agreements to avoid duplication, but these are based on people having an understanding of one another’s existing collections and acquisition goals. This can be good enough in the instances where you have good people working together – and the partners have lots of good people – but it leaves inevitable gaps and begs for some coordination and overarching assessment. The middleware systems facilitates this kind of large-scale cross collections comparison and planning.

18. Retention Commitment Forever is not a useful planning horizon
x$ / ∞yrs is not a meaningful way to do ROI, but does reflect the commitment being made.
Other retention periods make functional sense, but don’t reflect partner intention
x$ / 25yrs is workable math, but begs questions about the commitment in year 26
The other issue that received substantial treatment in the study was our commitment to long-term retention of these collections. One of the questions before us at this symposium is not if we can trust one another – sure; we’re decent, well intentioned folks – but how to trust one another and how to make meaningful assurances of our commitment.
One approach is to make a series of agreements over a set term, which has many practical advantages. It’s easy to implement and the math is meaningful, and those factors can carry the day for good reason, but it is difficult to put so-called “teeth” into these agreements without undoing some of the easiness, and without those legal or financial bonds, we have to fall back on a less definitive form of trust.

19. Retention Commitment
Shared Collection Plan decouples intent and operations
An agreement (made according to ReCAP’s bylaws) records the intent:
To retain Shared Collections for the life of the corporation (ReCAP, Inc., a New-Jersey not-for-profit) which lives on indefinitely
That agreement is in turn supported by layers of law and policy regarding the dissolution or transfer of assets: Partner and ReCAP bylaws and charters, State and Federal law, etc.
ReCAP has been able to answer this issue by separating intent and operations in its retention commitment. ReCAP is a New Jersey not for profit corporation, and like all corporations, has an indefinite life. By making the shared collections agreement “for the life of the corporation”, we give some legal weight to our perpetual commitment. There is a lot of law and policy around the maintenance and dissolution of corporations, so even if there were an end to our retention, it would progress in a way that was clearly anticipated, predictable, and in accord the public interest, at least insofar as that has been expressed in the law.

20. Retention Commitment
Shared Collection Plan decouples intent and operations
The corporation continues capital and operating efforts using the best available means
Operating and capital decisions are more accountable:
do they measurably compromise or advance intent, then
what are the costs, benefits, and timelines involved, and
are those competitive
Supported by over a decade of local operating and capital data, and a nation-wide knowledgebase of other repositories and allied professions
With intent settled in a meaningful way, the questions of how to operate ReCAP in a cost effective manner can be settled on the most appropriate terms. For example, we have a 10-year running major maintenance schedule, updated annually, covering every replaceable part of the facility, so that the partnership can forecast its needs and capitalize itself appropriately.

21. Development Cycles
: ReCAP launches and develops its core services
: ReCAP explores new directions
2014 on: ReCAP implements new services and refines core operations
This separation also lets us have a freer hand in optimizing our work processes to see if they hold up or, (click) if they break, to see if a better plan emerges as we fix things, and all of that with having to disentangle agreed-upon processes in a retention agreement.
Having freedom to refine process is at the heart of preservation – sustainable activities optimized over time– activities need to be tested over and over, (click) across the whole operation, so that nothing goes stale, and so that over time, everything set of activities moves towards its simplest, more reliable and predictable form.
If that all seems a little abstruse or high-falutin’ for you, then let me put it in the context of some classic preservation maxims: avoid basements and skylights, and watch out for people. Have a good pair of boots handy, don’t fly too close to the sun and keep procedures simple enough that it’s possible to minimize, detect, and correct for inevitable human errors.

22. All this and More Discovery to Delivery Report:
This project was funded by the Andrew W. Mellon Foundation, and the report is issued as CC-BY 4.0
Jacob Nadal, Executive Director, ReCAP
With that, let me thank you for your time and attention, and give you the links to the “Discovery to Delivery” report.